Independent Assessment of the Florida Statewide Medicaid Managed Care Long-term Care Program

Transcription

1 Independent Assessment of the Florida Statewide Medicaid Managed Care Long-term Care Program Final Interim Report (FY ) Deliverable 6 Prepared for Florida Medicaid MED 186 June 30, 2017 College of Medicine College of Social Work Florida State University

2 Florida Medicaid Long-term Care Program Final Interim Report (FY ) Executive Summary The Florida State University evaluation team examined measures of access to care, quality of care, and costeffectiveness of care for the Florida Statewide Medicaid Managed Care (SMMC) Long-term Care (LTC) program year , the first full state fiscal year of the LTC program implementation. Seven managed care plans were contracted by the Agency for Health Care Administration (Agency) to provide LTC services to nearly all Medicaid recipients who met financial qualifications and level of care requirements. Each of the state s eleven regions had at least two managed care plans offering services to Medicaid recipients. The Florida Legislature directed the Agency to adjust managed care plan capitated rates annually to provide an incentive to shift services from nursing facilities to home and community-based settings. The Agency, in collaboration with the Department of Elder Affairs (DOEA), has successfully undertaken the task of moving approximately 83,000 LTC recipients from the previous delivery model directed by the state, as well as approximately 41,000 new LTC recipients (those who became eligible during the study year) to a managed care model using capitated payments to private managed care plans. Access to Care The Independent Assessment of the Florida Statewide Medicaid Managed Care Long-term Care Program examined the impact on enrollees access to care during state fiscal year (SFY) The evaluation team and the Agency identified key issues of importance to policy makers and LTC stakeholders. They developed four research questions (RQs) to guide this evaluation, which uses quantitative analytical methods to support findings. The evaluation team draws comparisons between access to care under the LTC program versus under legacy waiver programs in the Pre-LTC period (SFY through SFY ). Appendix 1 of this report provides further information on the methodology and data sources. The four RQs listed below were used to guide the evaluation of access to care. 1. Have there been changes in the accessibility of services for enrollees compared to the previous LTC programs? 1 This report summarizes head-to-head descriptive comparisons, so the analytic periods for the Access to Care evaluation for this assessment are SFY1415 for the LTC period and SFY1213 for the Pre-LTC period. All five years of the Pre-LTC period will be used in future evaluations that involve trend analysis methods. 2 Final Report (Fiscal Year ) June 30, 2017

3 2. How has the population being served in the LTC program shifted (characteristics of the population and service utilization) between nursing facilities and home and community-based services (HCBS) over time? What LTC plan strategies are impacting these shifts? 3. Do plans offer additional (expanded) benefits and ways to access services, including a Participant Directed Option (PDO), and to what extent do enrollees use these services? 4. Are there disparities by racial and ethnic groups in enrollees placements in certain settings and utilization of services? Adequate access to care requires a network of willing LTC providers contracted by the plans located in reasonable proximity to enrollees residences. On a statewide basis, the network at face value appears to be robust. However, there are geographical differences in facility availability by county. These differences may be even greater within certain plans, depending on the robustness of individual plans networks. This possibility requires additional inquiry in future reports before reaching a universal (i.e., statewide) conclusion about network adequacy. In the aggregate, Florida LTC plans are realizing the Agency s stated policy goal of reducing the Florida Medicaid LTC enrollee nursing facility (NF) population by transitioning enrollees to home and community-based settings. The NF residency rate slowly grew over the Pre-LTC period, with about a one percentage point increase in the nursing facility population from the start to the end of SFY1213, but this rate steadily decreased over the LTC period, with about a three percentage point reduction in the NF population from the start to the end of SFY1415. The success of enrollee transitions contributed to this decline, as enrollees were more than twice as likely to successfully transition into HCBS settings and remain there for the duration of the evaluation year in the LTC period (4.0%) than in the Pre-LTC period (1.7%). It is reasonable to conclude that HCBS services can keep a substantial number of frail seniors and individuals with disabilities in the community even if a caregiver is in crisis. The percentage of enrollees experiencing multiple location shifts, i.e., moving in/out of nursing facility and community-based care settings two or more times in a year, increased between the Pre-LTC and LTC period. This may be a positive finding if the shifts are appropriate, however, these transitions may be a time of heightened stress for both enrollee and caregiver(s). The plans should strive to minimize the occurrence of multiple shifts. Accordingly, the evaluation team wants to understand more about how the plans identify enrollees in their original nursing facility residencies for eventual transition into home and community-based settings. Data permitting, the evaluation team would like to address this issue in future reports. Enrollment in the Participant Directed Option during the LTC period was modest, though PDO enrollment slowly grew throughout the evaluation year. At the end of SFY1415 about 5.4% of eligible enrollees who receive services in home-based settings had signed up to receive at least one service via the PDO. The evaluation team identified substantial differences in location of care by race/ethnicity. Hispanic enrollees were far more likely than black enrollees and white enrollees to reside in home and community-based settings. There are further differences by race/ethnicity within the home and community-based population; specifically, differences in rates of assisted living facility and assistive care service receipt. White HCBS enrollees were more than 1.5 times as likely as Hispanic HCBS enrollees and almost 2.5 times as likely as black HCBS enrollees to 3 Final Report (Fiscal Year ) June 30, 2017

4 receive care in assisted living facilities. Moreover, white enrollees were more likely to receive assistive care services than both Hispanic and black enrollees. Some of this imbalance may be explained by sociocultural differences and possibly differences in the geographic availability of assisted living facilities and assistive care services. Quality of Care Seven quality of care research questions were examined based on established indicators and expert opinion to measure quality of care under a long-term care program. It uses both quantitative and qualitative methods. Key research questions (RQs) identified differences in quality of care between those eligible Medicaid recipients who were in nursing facilities (NFs) or participated in the waiver programs before the LTC program (Pre-LTC) was implemented and those LTC program enrollees who currently participate in the LTC program. Appendix 2 of this report provides further information on specific methodologies employed and data sources used. The seven RQs listed below were used to guide the evaluation of quality of care. 1. Has the quality of services that enrollees receive changed compared to the previous LTC programs and over time? 2. What have managed care plans done to improve quality of care? 3. How has timeliness of service delivery changed compared to the previous LTC programs and over time? 4. Are enrollees rights being addressed in accordance with the waiver and contract? 5. How effective are the LTC plans complaint, grievance and appeals processes? 6. What are the rates of complaints, grievances and appeals in the LTC program and have these changed over time? 7. Is there sufficient monitoring and oversight of the program by the state? What types of monitoring are done, and how are they conducted? This report for Year largely reaffirms some of the findings from the transition year ( ). For example, the findings indicate that enrollee satisfaction with their overall health and quality of life were greater in the LTC group than pre-ltc among residents living in other residential settings (ORS) 2 and among residents living at home. Self-reported overall health improved from pre-ltc to LTC among residents living in other residential settings and among residents living at home. Specifically, significantly more LTC residents in ORS or home rated their health as excellent/very good/good and fewer reported their health as fair or poor. Additionally, residents living in ORS settings were significantly more likely to indicate that they are satisfied or very satisfied with their overall quality of life compared to ORS residents Pre-LTC. Considerably more LTC 2 For the purposes of these analyses, ORS includes care in assisted living facilities (ALFs) and adult family care homes (AFCHs). ORS provide many of the same services as those required by enrollees living in their own homes. They are home-like environments that are shared with others and are considered social rather than medical settings. 4 Final Report (Fiscal Year ) June 30, 2017

5 program home enrollees and more than half of the ORS enrollees indicate that they have someone who can help them if needed, someone other than their primary caregiver. Also, the majority of home and ORS residents report that their quality of life has improved since enrolling in their LTC plan. Furthermore, the level of unmet need declined from Pre-LTC to LTC among ORS and home residents. While self-reported depression increased from pre-ltc to LTC among residents in ORS and home, it decreased among NF residents. A review of select NF Quality Measures provided mixed results with the presence of unhealed pressure ulcers increasing in the LTC group; the percent of those experiencing one or more falls decreasing; and those experiencing moderate to severe pain decreasing. Most of the quality measures for residents of ORS and home improved from Pre-LTC to LTC, however, among home residents there was a small increase in those eating alone most of the time, which may suggest isolation. Quality of care for ORS and home residents is affected by the health (mental, emotional, and financial) of their caregivers. The percent of caregivers in crisis rose from Pre- LTC to LTC among both residents of ORS and home. Cost-effectiveness of Care The Independent Assessment of the Florida Statewide Medicaid Managed Care Long-term Care Program provides the cost analysis of the program implemented on August 1, 2013 using quantitative methods. A detailed report of specific methodologies employed and data sources used for the evaluation of cost-effectiveness of care are provided in Appendix 3 of this report. The following research questions were used to guide the assessment evaluation of cost-effectiveness. 1a. How has the LTC program affected the growth of Medicaid costs for LTC program enrollees? 1b. How has the shift between HCBS and nursing facilities under the LTC program affected enrollees' Medicaid costs? The cost analysis takes into account the LTC program s programmatic change. Prior to the LTC program, the Agency paid Medicaid providers of LTC services primarily on a fee-for-service basis, with the exception of Nursing Home Diversion (NHD) and the Frail Elder Option. The LTC program implements a capitated payment system. For the first full fiscal year of the LTC program implementation, the results presented in these cost analyses show that the cost of the LTC program has been cost neutral as measured by the comparison of the enrollee s average per individual per month costs for long-term care services in the pre-program period to those the enrollee incurred after the program s full implementation. The average monthly nominal 3 LTC program cost was estimated at $3, versus $3, in the Pre LTC period. This represents an increase of $ After adjusting for inflation, the average monthly LTC program cost decrease or savings comes down to $0.03 per individual per month. Additionally, the evaluation team calculated a Medical Cost Ratio defined as the average monthly inflated costs for LTC services provided Pre-LTC program divided by the average monthly costs for LTC services provided 3 Price or cost adjusted for inflation 5 Final Report (Fiscal Year ) June 30, 2017

6 under the LTC program. (Please note: This is not a Medical Loss Ratio measurement.) Given the average encounter costs of $3, per member per month, the Medical Cost Ratio (MCR) 4 is calculated at (or approximately 99.0%). The encounter claims costs seem to decline at a rate of approximately percent annually (SFY ), where the LTC program claims seem to indicate cost neutrality (both nominal based). This means that overall the MCR is under downward pressure. The continuation of this downward MCR trend will depend upon continued rebalancing from more expensive nursing facilities and hospice to care in HCBS settings. Additional years of data will be required to determine whether these short-term findings are more representative for the LTC program. 4 The Medical Cost Ratio (MCR) used in this report is much narrower in as it is calculated by the average per-individual per-month encounter costs relative to per-individual per-month Medicaid claims (x 100%). Hence it is duly noted that the MCR definition used here is not the same as the more generally used Medical Loss Ratio, as Plan revenues are not part of the equation. 6 Final Report (Fiscal Year ) June 30, 2017

7 Report Prepared By Co- Principal Investigators Project 1 Access to Care Project 2 Quality of Care Project 3 Cost of Care Data Warehouse Henry J. Carretta, PhD Jean C. Munn, PhD Martijn R. Niekus, Drs Glenn Mitchell, PhD Co-investigators Gail R. Bellamy, PhD Paul Katz, MD George Rust, MD Antonio Terracciano, PhD Data Analysts Karen Geletko, MPH Samantha Goldfarb, PhD, MPH Katelyn Graves, PhD Amelia Jones, MS GIS Jeffrey Harman, PhD Margaret M. Holland, MSW Kelsey Houser, MA Heidi Kinsell, PHD Yuxia Wang, MPH Project Manager Michael P. Smith, MA, MPA Principal Investigator Leslie M. Beitsch, MD, JD 7 Final Report (Fiscal Year ) June 30, 2017

8 Table of Contents Executive Summary...2 Access to Care...2 Quality of Care...4 Cost-effectiveness of Care...5 Report Prepared By...7 Table of Contents...8 List of Tables List of Figures Introduction Background Core Demographics I: Access to Care Purpose Findings RQ1: Have there been changes in the accessibility of services for enrollees compared to the previous LTC programs? RQ2: How has the population being served in the LTC program shifted (characteristics of the population and service utilization) between nursing facilities and HCBS over time? What LTC plan strategies are impacting these shifts? RQ3: Do plans offer additional (expanded) benefits and ways to access services, including a Participant Directed Option (PDO), and to what extent do enrollees use these services? RQ4: Are there disparities by racial and ethnic groups in enrollees placements in certain settings and utilization of services? Conclusions Recommendations II: Quality of Care Purpose Quantitative Findings RQ1: Has the quality of services that enrollees receive changed compared to the previous LTC program(s) and over time? RQ2: What have managed care plans done to improve quality of care? RQ3: How has timeliness of service delivery changed compared to the previous LTC programs and over time? RQ4: Are enrollees rights being addressed in accordance with the waiver and contract? RQ5: How effective are the LTC plans complaints, grievances and appeals processes? RQ6: What are the rates of complaints, grievances and appeals in the LTC program and have these changed over time? Final Report (Fiscal Year ) June 30, 2017

9 RQ7: Is there sufficient monitoring and oversight of the program by the state? What types of monitoring are done, and how are they conducted? Qualitative Findings Conclusions Recommendations III: Cost-effectiveness of Care Purpose Findings RQ: 1a. How has the LTC program affected the growth of LTC Medicaid costs for LTC program enrollees? RQ: 1b. How has the shift between HCBS and nursing facilities under the LTC program affected enrollees Medicaid costs? Conclusions Evaluation Limitations Appendix Access to Care Methodology Study Design Data Sources Analytic Methods for the Quantitative Evaluation Preliminary Risk Adjustment Categories GIS Methods Appendix Quality of Care Methodology Study Design for the Quantitative Evaluation Data Sources for the Quantitative Evaluation Analytic Methods for the Quantitative Evaluation Study Design for the Qualitative Evaluation Data Sources for the Qualitative Evaluation Analytic Methods for the Qualitative Evaluation Interview Guide Appendix Cost-effectiveness of Care Methodology Study Design and Data Methods Methodology Inflation Factors and Difference Analyses Final Report (Fiscal Year ) June 30, 2017

10 List of Tables Table 1. List of Acronyms Table 2. Implementation Schedule for LTC Program LTC Rollout Schedule Table 3. Long-term Care Managed Care Plan Providers in Florida Table 4. LTC Enrollees by Age, August 1, December 31, Table 5. LTC Enrollees by Gender, August 1, December 31, Table 6. LTC Enrollees by Race and Ethnic Categories, August 1, December 31, Table 7. LTC Enrollees by Region, August 1, December 31, Table 8. LTC Enrollees by County, August 1, December 31, Table 9. LTC Enrollees by Age at Death, August 1, December 31, Table 10. Nursing Facility Residency Rates by County, August 1, December 31, Table 11. Nursing Facility Residency Rates by Race, August 1, December 31, Table 12. Logistic Regression, HCBS Enrollee = Yes, August 1, December 31, Table I.1. Number of Providers Offering Each LTC Service, LTC period Table I.2. Count of In-Network Facilities by County, LTC Period Table I.3. Mean Monthly Percentage of Enrollees in HCBS Settings Receiving Each LTC Service, Pre-LTC and LTC Period Table I.4. Mean Number of Days until First Service Delivery, Pre-LTC and LTC Period Table I.5. Total Member Months by Location of Care, Pre-LTC and LTC Period Table I.6. Enrollee Location of Care Rates Over Time, 6 Pre-LTC and LTC Period Table I.7. Transition Rates, Pre-LTC and LTC Period Table I.8. Predicted Probabilities of Nursing Facility Residency in the Pre-LTC and LTC Period Table I.9. Mean CDPS Risk Scores for Enrollees in the First Month of the LTC Period Table I.10. Contracted Bed Counts by Plan and Region in the LTC Period Table I.11. List of Expanded Benefits Offered by the Seven Plans in the LTC Period Table I.12. Overall Home-Based Member Months with PDO Enrollment, LTC Period Table I.13. Home-Based Member Months with PDO Enrollment by Plan, LTC Period Table I.14. PDO Enrollment by Month, LTC Period Table I.15. PDO Enrollment by Plan, LTC Period Table I.16. PDO Enrollment Span by Plan, LTC Period Table I.17. Services Selected Through PDO and Received by PDO Enrollees in the LTC Period Table I.18. Total Member Months by Location of Care Period Stratified by Race/Ethnicity, Pre-LTC and LTC Period Table I.19. Mean Monthly Percentage of Enrollees in HCBS Settings Receiving Each LTC Service by Race/Ethnicity, LTC period only Table I.20. Mean Number of Days until First Service Delivery by Race/Ethnicity, LTC period Table I.21. Mean Distance to Nearest Nursing Facility for Community-based Enrollees by Race/Ethnicity, LTC period Table II.1. Quality of Life and Self-Reported Health for Enrollees in Other Residential Settings (ORS) Pre-LTC and LTC Final Report (Fiscal Year ) June 30, 2017

11 Table II.2. Quality of Life and Self-Reported Health for Enrollees Living at Home with HCBS Pre-LTC and LTC. 68 Table II.3. NF Enrollee Participation in Assessment Process Pre-LTC and LTC Table II.4. Proportion of Enrollees with Unmet Need in Other Residential Settings (ORS) and Home Pre-LTC and LTC Table II.5. Depression as Measured by PHQ-9 Scores 1 > 10 in All Sites of Care Table II.6. Depression as Measured by PHQ-9 Scores 1 Distribution of Scores (Pre-LTC) Table II.7. Depression as Measured by PHQ-9 Scores 1 Distribution of Scores (LTC) Table II.8. Quality Measures Selected from CMS QMs for NFs Table II.9. Healed Pressure Ulcers by Period (Pre-LTC, LTC) Table II.10. Nursing Facility (NF) Enrollees with Pain Table II.11. Quality Measures for Enrollees Residing at Home Table II.12. Quality Measures Specific to Enrollees Living in Other Residential Settings Table II.13. Number and percent of caregivers in crisis based on assessor observation (Other Residential Settings) Table II.14. Number and percent of caregivers in crisis based on assessor observation (Home) Table II.15. Performance Improvement Projects by Plan Table III.1. Pre-LTC Program Per Member Per Month (PMPM) Claims SFY , SFY , and LTC Program Per Member Per Month (PMPM) Claims SFY Table III.2. Average Expected Per Member Per Month (PMPM) Values Based on Interrupted Time Series Table III.3. LTC Program Claims SFY Per Member Per Month Bucket Profile Table III.4. Average Service Utilization by Location of Service for SFY (in SFY $) Table III.5. Average Service Utilization by Location of Service for SFY (in SFY $) Appendix 1 Table 1. Access Research Questions, Associated Measures, and Data Sources Appendix 1 Table 2. Enrollee Residence Location Determination Methodology Appendix 1 Table 3. Frequency of New Program Entrants in HCBS Settings Who Meet First Service Delivery Exclusion Criteria (when Case Management Services Excluded) Appendix 2 Table 1. Quality Research Questions, Associated Measures and Data Sources Final Report (Fiscal Year ) June 30, 2017

12 List of Figures Figure 1. Statewide Medicaid Managed Care Region Map Figure 2. Model of the Dynamic Flow of Enrollees in Florida s Long-Term Care Program Figure 3. Nursing Facility Residency Rates by Region, August 1, December 31, Figure 4. Death, Enrollees in Nursing Facilities, August 1, December 31, Figure 5. Death, Enrollees in HCBS Settings, August 1, December 31, Figure 6. Age Categories, Nursing Facility v. HCBS Enrollees, August 1, December 31, Figure 7. Available Caregiver, Nursing Facility v. HCBS Enrollees, August 1, December 31, Figure 8. Caregiver in Crisis, Nursing Facility v. HCBS Enrollees, August 1, December 31, Figure 9. ADL Bathing, Nursing Facility v. HCBS Enrollees, August 1, December 31, Figure 10. ADL Dressing, Nursing Facility v. HCBS Enrollees, August 1, December 31, Figure 11. ADL Eating, Nursing Facility v. HCBS Enrollees, August 1, December 31, Figure 12. ADL Toileting, Nursing Facility v. HCBS Enrollees, August 1, December 31, Figure 13. ADL Transferring, Nursing Facility v. HCBS Enrollees, August 1, December 31, Figure 14. ADL Walking, Nursing Facility v. HCBS Enrollees, August 1, December 31, Figure I.1. Proportion of Enrollees Residing in a Nursing Facility or and HCBS Setting Over Time, Pre-LTC and LTC Period Figure I.2. Change in the Predicted Probabilities of Nursing Facility Residency in the LTC Period versus the Pre- LTC Period Figure I.3. Proportion of Member Months in a NF or HCBS Setting by Race/Ethnicity, Pre-LTC and LTC Period 54 Figure I.4. Proportion of Enrollees Residing in a Nursing Facility Over Time by Race/Ethnicity, LTC Period Figure I.5. Proportion of Enrollees Residing in an HCBS Setting Over Time by Race/Ethnicity, LTC Period Figure I.6. Map of LTC Enrollees Living in the Community Figure I.7. Map of Medicaid Certified Nursing Facilities in the State of Florida Figure I.8. Map of Medicaid Certified Nursing Facilities Overlaid with LTC Enrollees Living in the Community Stratified by Race/Ethnicity Figure II.1. Community Dwelling 1 LTC Enrollees Change in Quality of Life Figure II.2. Unmet Need in Nursing Facilities Figure II.3. Proportion of Nursing Facility Residents Receiving Specific Medications Figure II.4. Community Dwelling LTC Enrollees Ratings of Long-term Care Program Services Figure II.5. Percentage of Enrollees Reporting Services on Time Figure II.6. Percentage of Enrollee Contacts Completed within Five Business Days of Enrollment for Community- Based Enrollees Figure II.7. Percentage of Enrollee Contacts Completed within Seven Business Days of Enrollment for Nursing Facility Enrollees Figure II.8. Percentage of Case Manager Follow-up Calls Completed within Seven Days of Initial Assessment.. 93 Figure II.9. Percentage of Calls Documented in the Case Notes Figure II.10. Enrollee Rights Figure II.11. Personal Emergency Plan Figure II.12. Primary Care Physician Documentation Final Report (Fiscal Year ) June 30, 2017

13 Figure II.13. Care Plan Summary Figure II B Comprehensive Assessment Figure II.15. Screening for High Risk Figure II.16. Average Days to Resolution, Beneficiaries Issues Figure II.17. Complaints/Issues Reported by Plan from July 1, 2014 to June Figure II.18. Numbers of Complaints/Issues per Quarter by Plan Figure II.19. Qualitative Data Collection Figure III.1. Interrupted Time Series Results on Nominal Per Member Per Month (PMPM) Claim Values, Pre-LTC Program SFY , SFY , and LTC Program Per Member Per Month (PMPM) Claims SFY Figure III.2. Per Member Per Month Nominal Encounter Values, SFY (in $) Figure III.3. Absolute Distribution of LTC Program Enrollees by Location, SFY Figure III.4. Relative Distribution of LTC Program Enrollees by Location, SFY Appendix 1 Figure 1. Proportion of Enrollees with HCBS Residency Assigned to Each Risk Category in the Pre- LTC Period Appendix 1 Figure 2. Proportion of Enrollees with NF Residency Assigned to Each Risk Category in the Pre-LTC Period Appendix 1 Figure 3. Proportion of Enrollees with HCBS Residency Assigned to Each Risk Category in the LTC Period Appendix 1 Figure 4. Proportion of Enrollees with NF Residency Assigned to Each Risk Category in the LTC Period Final Report (Fiscal Year ) June 30, 2017

14 A/DA ADL AEC AFCH AHCA AHRQ ALF AMG CARES CIRTS (DOEA) CIRTS (AHCA) CMS COV DOEA EQRO FSU GAO HCBS HUM LTC MDS MOL NF ORS OTC PERS PHQ-9 PIPs POC Pre-LTC QM RAP RQ SMMC SUN URA Term Case Managers or Case Management Dually Eligible Enrollee(s) Frail Elder Option Individual(s) Legacy Legacy Waivers or HCBS Legacy Waivers Participant(s) Plans or LTC plans Recipient(s) Table 1. List of Acronyms List of Acronyms Aged/Disabled Adult Activities of Daily Living American Eldercare, Inc. Adult Family Care Homes Agency for Health Care Administration (Florida) Agency for Healthcare Research and Quality (United States) Assisted Living Facility Amerigroup Florida, Inc., d/b/a Amerigroup Community Care Comprehensive Assessment and Review for Long-term Care Services Client Information and Registration Tracking System Complaint Issues Reporting and Tracking System Centers for Medicare and Medicaid Services (United States) Coventry Health Care of Florida Department of Elder Affairs (Florida) External Quality Review Organization Florida State University United States Government Accountability Office Home and Community-Based Services Humana Medical Plan, Inc. Long-term Care/The period following the implementation of SMMC LTC Minimum Data Set Molina Health Care of Florida, Inc. Nursing Facility Other Residential Settings Over-the-Counter Personal Emergency Response System Patient Health Questionnaire (9 items) Performance Improvement Projects Plan of Care The period prior to the implementation of SMMC LTC Quality Measure Resident Assessment Protocol Research Question Statewide Medicaid Managed Care Sunshine State Health Plan, Inc. UnitedHealthcare of Florida, Inc. Glossary Meaning LTC plan employees who deliver case management services to LTC enrollees. LTC recipients or enrollees eligible for Medicare and Medicaid. Recipients in managed care plans A Pre-LTC program for qualifying members of the LTC population. Persons not yet eligible for Medicaid Programs and waivers in place during the Pre-LTC period. HCBS waivers operated by DOEA during the Pre-LTC period: (Aged/Disabled Adult, Assisted Living, Channeling for the Frail Elderly, and Nursing Home Diversion). Managed Care enrollees who elect the Participant Direction Option Managed care plans contracted under the Statewide Medicaid Managed Care (SMMC) plan Individuals eligible for Medicaid regardless of whether or not they are receiving services 14 Final Report (Fiscal Year ) June 30, 2017

15 Introduction Background The Florida Agency for Health Care Administration (AHCA), hereafter referred to as the Agency, submitted a 1915(b) and a 1915(c) waiver application to the Centers for Medicare and Medicaid Services (CMS) to implement the Florida Long-term Care Managed Care Program mandated by the 2011 Florida Legislature. House Bill 7107 created Section 409 of Florida Statutes to establish a statewide long-term care managed care program for Medicaid recipients who are (a) 65 years of age or older, or age 18 or older and eligible for Medicaid by reason of a disability; and (b) determined to require nursing facility (NF) level of care. The Agency received approval for the 1915(b) and 1915(c) waivers from CMS on February 1, 2013 and began administering the Florida Long-term Care (LTC) Managed Care Program in partnership with the Department of Elder Affairs (DOEA). The authorities requested in the CMS waiver applications allow the State to require eligible Medicaid recipients to receive their NF, hospice, and home and community-based services (HCBS) through managed care plans selected by the State through a competitive procurement process. NF level of care is determined by the existing statewide Comprehensive Assessment and Review for Long-term Care Services (CARES) Program. Medicaid recipients eligible for the Florida Long-term Care Managed Care Program received a choice of plans available in each of the eleven regions. With the implementation of Statewide Medicaid Managed Care (SMMC), Florida s legacy HCBS waivers (Aged/Disabled Adult [A/DA] Waiver, Consumer Directed Care Plus for individuals in the A/DA Waiver, Assisted Living Waiver, Nursing Home Diversion Waiver, Frail Elder Option, and Channeling Services Waiver) were phased out. Eligible recipients aged 65 and older, and individuals with disabilities aged who met nursing facility level of care, were required to enroll in the SMMC LTC program. Hospice is a state plan service covered under the 1915(b) waiver, but Medicaid enrollees may receive hospice services through the Managed Medical Assistance (MMA) program without being required to enroll in the LTC program. Hospice care for dual eligible LTC program enrollees is provided by their Medicare coverage. The Medicaid payer of last resort rule requires dual eligible LTC program enrollees hospice coverage to be paid by Medicare. Some home and communitybased waiver program recipients (e.g. Developmental Disabilities Waiver Program, Traumatic Brain & Spinal Cord Injury Waiver, Project AIDS Care Waiver, Adult Cystic Fibrosis Waiver, Program of All Inclusive Care for the Elderly, Familial Dysautonomia Waiver, and Model Waiver) were not mandated to enroll in the LTC program but could do so if they chose and meet the eligibility requirements of the program. 5 The vast majority of LTC program enrollees are dually eligible for Medicare and Medicaid. 5 Florida Agency for Health Care Administration. Quarterly Statewide Medicaid Managed Care Report (Winter 2014). In: Business Intelligence Unit Medicaid Program Analysis, ed. Tallahassee, Florida: Florida Agency for Health Care Administration; Final Report (Fiscal Year ) June 30, 2017

16 The Agency s goals of the LTC program include: Enhancing fiscal predictability and financial management by converting the purchase of Florida Medicaid services to capitated, risk-adjusted, payment systems. Transitioning LTC individuals who wish to go home from nursing facility care to assisted living or their own homes. Maximizing opportunities for LTC enrollees to remain in their homes/communities Improving patient centered care, personal responsibility, and active patient participation. Providing recipients with a choice of plans and benefit packages. Improving the health of recipients, not just pay claims when people are sick. Improving quality of life Avoiding more costly care The Agency is responsible for making payments to the managed care plan, adjusting applicable capitation rates to reflect budgetary changes in the Medicaid program, and reconciling payments for NFs and hospices. In addition, the Agency is responsible for verifying the managed care plans achieved savings rebate (ASR). Florida s LTC program employs a shared-savings model, defined as the ASR, under which the Agency will calculate a managed care plan s pre-tax income as a percentage of revenues. Managed care plans will share their income with the Agency based on three revenue tiers, and ultimately may retain up to 7.5 percent of pretax income as a percentage of revenue. In addition, a managed care plan that exceeds the Agency s quality measures to achieve better health outcomes for enrollees may retain an additional one percent of that revenue. Most importantly, as noted in Section (5), Florida Statutes, the Agency is required to make an incentive adjustment in payment rates to encourage the increased utilization of HCBS and a corresponding reduction in institutional care. Incentive adjustments must continue until no more than 35 percent of each plan s enrollees reside in institutional settings. Furthermore, the managed care plans will assess their enrollees who reside in NFs for appropriateness of transition to receiving HCBS. Managed care plans must have centralized executive administration and adequate staffing and information systems capable of ensuring that the plan can appropriately manage financial transactions, record keeping, data collection, and other administrative functions, including the ability to submit any financial, programmatic, encounter data, or other necessary information. In addition, the managed care plan reports financial information to the Agency, including quarterly and annual financial statements, and establishes functions and activities governing program integrity in order to reduce the incidence of fraud and abuse. 16 Final Report (Fiscal Year ) June 30, 2017

17 In early 2013, the Agency selected and contracted with six Health Maintenance Organizations (HMOs) paid on a capitated basis, and one provider service network (PSN) using a fee-for-service payment mechanism. The 26 required LTC core benefits cover LTC services only and do not cover medications, doctor s visits, or other medically necessary services. 6 The PSN was subsequently converted to a capitated plan (American Eldercare, Inc.) on September 1, 2014, which then merged with Humana Medical Plan, Inc. on July 1, The seven managed care plans examined in this report are: American Eldercare, Inc. (AEC) Amerigroup Florida, Inc. (AMG) Coventry Health Plan (COV) Humana Medical Plan, Inc.(HUM) Molina Healthcare of Florida, Inc. (MOL) Sunshine State Health Plan (SUN) United Healthcare of Florida, Inc. (URA) In August 2013, enrollment for the LTC program began in Region 7 and continued through March 2014 throughout the remaining regions. Table 2 shows the timetable for rollout, from August 2013 through March Figure 1 depicts the different LTC program regions, while Table 3 lists the LTC Plans and the respective regions in which they operate. Table 2. Implementation Schedule for LTC Program LTC Rollout Schedule 7 LTC Rollout Schedule Plans Aug Sep Nov Dec Feb Mar Regions American Eldercare, Inc. x x x x x x x x x x x Amerigroup Community Care x x Coventry Healthcare of Florida, Inc. x x x x Humana Comfort Choice x x x Molina Healthcare of Florida, Inc. x x x Sunshine Health x x x x x x x x x x United Healthcare of Florida, Inc. x x x x x x x x x 6 Florida Agency for Health Care Administration. Statewide Medicaid Managed Care (SMMC) - Long-term Care Program. Available at: Florida Agency for Health Care Administration. Quarterly Statewide Medicaid Managed Care Report (Winter 2014). In: Business Intelligence Unit Medicaid Program Analysis, ed. Tallahassee, Florida: Florida Agency for Health Care Administration; Final Report (Fiscal Year ) June 30, 2017

18 Figure 1. Statewide Medicaid Managed Care Region Map Table 3. Long-term Care Managed Care Plan Providers in Florida LTC Plans Plan Name Regions of Operation American Elder Care, Inc.* 1 through 11 Amerigroup Community Care Florida 10 and 11 Coventry Healthcare of Florida, Inc. 6, 7, 9, and 11 Humana Medical Plan Comfort Choice, Inc. 4, 10, and 11 Molina Complete Care Healthcare of Florida, Inc. 5, 6, and 11 Sunshine Health 1, 3, 4, 5, 6, 7, 8, 9, 10, and 11 United Healthcare of Florida, Inc. 2, 3, 4, 5, 6, 7, 8, 9, and 11 * Humana Inc, purchased American Eldercare Inc. in early September 2013 and merged the entities on July 1, On October 6, 2016, the Agency contracted with Florida State University (FSU) to conduct an independent assessment of the LTC program. Under the terms of its contract, for this report, the FSU evaluation team evaluated Program Year (July 1, 2014-June 30, 2015) of the LTC program in terms of access to care, quality of care and cost-effectiveness of care. 18 Final Report (Fiscal Year ) June 30, 2017

19 Core Demographics Persons with a Medicaid claim record between August 1, 2013 and December 31, 2015, for LTC Administrative Fees or for LTC Prepaid Capitated Payments, and who received LTC services were merged with Medicaid demographics to derive the core LTC population. The tables below present the distribution of the LTC population by core demographic characteristics. Table 4. LTC Enrollees by Age, August 1, December 31, 2015 Age LTC Enrollees Percent < % , % , % , % , % , % Total 124,276 Source: Florida Medicaid FMMIS and Managed Care Plan Roster files. Table 5. LTC Enrollees by Gender, August 1, December 31, 2015 Gender LTC Enrollees Percent Female 84, % Male 40, % Unknown % Total 124,276 Source: Florida Medicaid FMMIS and Managed Care Plan Roster files. Table 6. LTC Enrollees by Race and Ethnic Categories, August 1, December 31, 2015 Race LTC Enrollees Percent White 70, % African American/Black 19, % Asian % Hispanic 22, % Native American % Not Determined 5, % Other 5, % Total 124,276 Source: Florida Medicaid FMMIS and Managed Care Plan Roster files. 19 Final Report (Fiscal Year ) June 30, 2017

20 Table 7. LTC Enrollees by Region, August 1, December 31, 2015 Region LTC Enrollees Percent 1 4, % 2 5, % 3 9, % 4 11, % 5 13, % 6 12, % 7 13, % 8 8, % 9 12, % 10 8, % 11 24, % Unknown % Total 124,276 Source: Florida Medicaid FMMIS and Managed Care Plan Roster files. Table 8. LTC Enrollees by County, August 1, December 31, 2015 County LTC Enrollees Percent Alachua 1, % Baker % Bay 1, % Bradford % Brevard 3, % Broward 8, % Calhoun % Charlotte 1, % Citrus 1, % Clay 1, % Collier % Columbia % Dade 24, % DeSoto % Dixie % Duval 5, % Escambia 2, % Flagler % Franklin % Gadsden % Gilchrest % 20 Final Report (Fiscal Year ) June 30, 2017

21 County LTC Enrollees Percent Glades % Gulf % Hamilton % Hardee % Hendry % Hernando % Highlands % Hillsborough 6, % Holmes % Indian River % Jackson % Jefferson % Lafayette % Lake 1, % Lee 2, % Leon 1, % Levy % Liberty % Madison % Manatee 1, % Marion 1, % Martin % Monroe % Nassau % Okaloosa % Okeechobee % Orange 6, % Osceola 1, % Palm Beach 8, % Pasco 2, % Pinellas 10, % Polk 3, % Putnam % St. Johns % St. Lucie 1, % Santa Rosa % Sarasota 3, % Seminole 1, % Sumter % 21 Final Report (Fiscal Year ) June 30, 2017

22 County LTC Enrollees Percent Suwanee % Taylor % Union % Volusia 4, % Wakulla % Walton % Washington % Out of State % Unknown % Total 124,276 Source: Florida Medicaid FMMIS and Managed Care Plan Roster files. Between August 1, 2013 and December 31, 2015, 34.72% of the LTC enrollees (43,147) died. This is not unusual given the advanced age of the majority of the population and their frailty. Table 9 shows the breakdown of deaths by age group. Table 9. LTC Enrollees by Age at Death, August 1, December 31, 2015 Age at Death LTC Enrollees Percent of All Deaths Percent of Age Cohort < % 13.89% , % 15.78% , % 21.33% , % 25.90% , % 32.82% , % 45.17% Total 43,147 Source: Florida Medicaid FMMIS and Managed Care Plan Roster files. The Florida LTC program is a dynamic system with enrollees moving between home and community and facilitybased settings of care. Inpatient hospitalizations result in temporary service disruptions, after which plan representatives reassess enrollees to ensure these persons continue to receive the proper level of care. They may shift from home and community to facility-based care settings, if the assessment determines more intensive services are needed after a hospital stay. People exit the system permanently when they move out of state or die, but there is an increasing influx of recipients into the system being driven by aging of the baby boom generation. There are also a large number of frail elders living at home with the support of HCBS offered by the Medicaid program. Initially, the HCBS waiver was approved to serve 35,852 individuals at any given time. The nursing facility services component has no cap. When demand for HCBS exceeds the supply, persons are placed on the waitlist. Figure 2 shows the flow of persons between the different care categories as well as flows in and out of the LTC program. 22 Final Report (Fiscal Year ) June 30, 2017

23 Figure 2. Model of the Dynamic Flow of Enrollees in Florida s Long-Term Care Program Aged and Adults with Disabilities Requests In Home Services, ALF, or AFCH Screened by ADRC Enters Medicaid Nursing Facility Services Waitlist Placement Discharged from Nursing Facility Services <60 days Determination of sufficient funding by DOEA & AHCA for release of individuals on waitlist Clinical or Financial Eligibility not met >120 days Highest prioritized are assessed for Clinical and Financial Eligibility Loss of Eligibility <120 days Enter Long-Term Care Program: Nursing Facilities and HCBS Exit Enrollee Moves Out of State Enrollee Dies Note: This diagram represents a high-level overview of the LTC program population dynamics. It is not exhaustive of all LTC program entrance or exit processes. The evaluation team provided it to assist with readers understanding. 23 Final Report (Fiscal Year ) June 30, 2017

24 The ratio of nursing facility to HCBS enrollees varies across the state. South Florida (particularly Miami Dade and Broward counties) show considerably more HCBS utilization relative to nursing facility utilization. Statewide, 52.25% of Medicaid enrollees were resident in a nursing facility during their most recent month in the program (N = 64,928). Figure 3. Nursing Facility Residency Rates by Region, August 1, December 31, % 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% N = 124,276 Source: Florida Medicaid FMMIS and Managed Care Plan Roster files. Table 10. Nursing Facility Residency Rates by County, August 1, December 31, 2015 County Percent Alachua 62.28% Baker 78.43% Bay 69.66% Bradford 80.45% Brevard 55.06% Broward 38.54% Calhoun 71.67% Charlotte 65.55% Citrus 65.19% Clay 69.58% Collier 59.83% Columbia 57.36% Dade 29.57% DeSoto 49.47% 24 Final Report (Fiscal Year ) June 30, 2017

25 County Percent Dixie 55.17% Duval 61.51% Escambia 61.09% Flagler 42.60% Franklin 59.63% Gadsden 52.76% Gilchrest 83.71% Glades 27.27% Gulf 69.51% Hamilton 43.95% Hardee 55.45% Hendry 72.10% Hernando 52.94% Highlands 54.50% Hillsborough 44.98% Holmes 57.37% Indian River 58.52% Jackson 71.09% Jefferson 57.21% Lafayette 68.13% Lake 63.94% Lee 56.02% Leon 60.67% Levy 61.11% Liberty 17.65% Madison 65.42% Manatee 64.87% Marion 59.72% Martin 75.20% Monroe 63.64% Nassau 64.33% Okaloosa 71.46% Okeechobee 62.96% Orange 65.26% Osceola 52.60% Palm Beach 53.94% Pasco 54.01% Pinellas 62.78% Polk 64.44% 25 Final Report (Fiscal Year ) June 30, 2017

26 County Percent Putnam 65.64% St. Johns 61.80% St. Lucie 54.77% Santa Rosa 54.17% Sarasota 69.98% Seminole 51.69% Sumter 60.87% Suwanee 72.21% Taylor 61.54% Union 4.76% Volusia 61.61% Wakulla 63.69% Walton 59.04% Washington 61.82% N = 124,276 Source: Florida Medicaid FMMIS and Managed Care Plan Roster files. Differences are also apparent in the use of nursing facility relative to HCBS when race is considered. Whites and African Americans are more likely than other races to reside in a nursing facility under the LTC program. Table 11. Nursing Facility Residency Rates by Race, August 1, December 31, 2015 Race Percent White 59.75% African American/Black 57.05% Asian 47.52% Hispanic 48.66% Native American 44.53% Not Determined 14.49% Other 59.37% N = 124,276 Source: Florida Medicaid FMMIS and Managed Care Plan Roster files. There is a pronounced difference in deaths between nursing facility residents and HCBS enrollees. 44% of the LTC enrollees in a nursing facility died during the study period. This contrasts with 25% of the HCBS enrollees. 26 Final Report (Fiscal Year ) June 30, 2017

27 Figure 4. Death, Enrollees in Nursing Facilities, August 1, December 31, % 56% No Yes N = 124,276 Source: Florida Medicaid FMMIS and Managed Care Plan Roster files. Figure 5. Death, Enrollees in HCBS Settings, August 1, December 31, % 75% No Yes N = 54,498 Source: Florida Medicaid FMMIS and Managed Care Plan Roster files. Figure 6 indicates that this difference in death rates owes in part to the age differences between the nursing facility and HCBS enrollees, particularly those enrollees aged 85+. It also very likely reflects the health status of the two subpopulations. Generally, those enrollees in nursing facilities are in much more frail condition. 27 Final Report (Fiscal Year ) June 30, 2017

28 Figure 6. Age Categories, Nursing Facility v. HCBS Enrollees, August 1, December 31, 2015 Age Category < Nursing Facility HCBS N = 124,276 Source: Florida Medicaid FMMIS and Managed Care Plan Roster files. One important characteristic that distinguishes nursing facility enrollees from HCBS enrollees is the availability of a caregiver to assist with daily care. The Client Information and Registration Tracking System (CIRTS) survey instrument from the Florida Department of Elder Affairs asks respondents whether a primary caregiver is available % of NF residents surveyed indicated they had no available caregiver. This compares with 68.01% for HCBS enrollees surveyed. This chart demonstrates that HCBS services can keep a substantial number of individuals in the community who lack an available caregiver. Figure 7. Available Caregiver, Nursing Facility v. HCBS Enrollees, August 1, December 31, 2015 Available Caregiver 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Nursing Facility HCBS Yes No N = 103,846 Source: Florida Department of Elder Affairs CIRTS and Managed Care Plan Roster files. 28 Final Report (Fiscal Year ) June 30, 2017

29 Another important characteristic that distinguishes Nursing Facility enrollees from HCBS enrollees is caregiver crisis. The table below shows the assessors determinations regarding whether the surveyed caregiver was in crisis. That crisis could be economic, emotional, or physical. The source is the Florida Department of Elder Affairs CIRTS assessment and is typically conducted when the enrollee is still in an HCBS setting. Assessors determined that 67.04% of surveyed caregivers who had someone residing in a nursing facility were in crisis. This compares with 53.54% with someone enrolled in HCBS services. It is reasonable to conclude that HCBS services can keep a substantial number of frail seniors in the community even though a caregiver is in crisis. Figure 8. Caregiver in Crisis, Nursing Facility v. HCBS Enrollees, August 1, December 31, 2015 Assessor Determined, Caregiver in Crisis 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Nursing Facility HCBS Yes No N = 35,345 (9,389 in Nursing Facility, 25,946 in HCBS) Source: Florida Department of Elder Affairs CIRTS and Managed Care Plan Roster files. Activities of Daily Living (ADLs) are important indicators of frailty. They measure the self-reported ability of survey respondents to perform common tasks, such as dressing, eating, toileting, transferring into and out of bed, and walking with varying degrees of assistance required. These range from total independence to complete dependence. Figures 9-14 below, derived from the Florida Department of Elder Affairs CIRTS assessment, compare the nursing facility resident population with individuals in the community receiving one or more HCBS services. As expected, the nursing facility resident population is more impaired than its HCBS counterpart. However, the HCBS population is also a very frail population. 29 Final Report (Fiscal Year ) June 30, 2017

30 Figure 9. ADL Bathing, Nursing Facility v. HCBS Enrollees, August 1, December 31, 2015 ADL Bathing 70% 60% 50% 40% 30% 20% 10% 0% Independent Assistive Device Supervision or Coaching Some Assistance Required Total Dependence Nursing Facility HCBS N = 88,166 Source: Florida Department of Elder Affairs CIRTS and Managed Care Plan Roster files. Figure 10. ADL Dressing, Nursing Facility v. HCBS Enrollees, August 1, December 31, 2015 ADL Dressing 70% 60% 50% 40% 30% 20% 10% 0% Independent Assistive Device Supervision or Coaching Some Assistance Required Total Dependence Nursing Facility HCBS N = 88,166 Source: Florida Department of Elder Affairs CIRTS and Managed Care Plan Roster files. 30 Final Report (Fiscal Year ) June 30, 2017

31 Figure 11. ADL Eating, Nursing Facility v. HCBS Enrollees, August 1, December 31, 2015 ADL Eating 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% Independent Assistive Device Supervision or Coaching Some Assistance Required Total Dependence Nursing Facility HCBS N = 88,166 Source: Florida Department of Elder Affairs CIRTS and Managed Care Plan Roster files. Figure 12. ADL Toileting, Nursing Facility v. HCBS Enrollees, August 1, December 31, 2015 ADL Toileting 60% 50% 40% 30% 20% 10% 0% Independent Assistive Device Supervision or Coaching Some Assistance Required Total Dependence Nursing Facility HCBS N = 88,166 Source: Florida Department of Elder Affairs CIRTS and Managed Care Plan Roster files. 31 Final Report (Fiscal Year ) June 30, 2017

32 Figure 13. ADL Transferring, Nursing Facility v. HCBS Enrollees, August 1, December 31, 2015 ADL Transferring 60% 50% 40% 30% 20% 10% 0% Independent Assistive Device Supervision or Coaching Some Assistance Required Total Dependence Nursing Facility HCBS N = 88,166 Source: Florida Department of Elder Affairs CIRTS and Managed Care Plan Roster files. Figure 14. ADL Walking, Nursing Facility v. HCBS Enrollees, August 1, December 31, 2015 ADL Walking 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% Independent Assistive Device Supervision or Coaching Some Assistance Required Total Dependence Nursing Facility HCBS N = 88,166 Source: Florida Department of Elder Affairs CIRTS and Managed Care Plan Roster files. The figures show comparisons between LTC enrollees in NF or HCBS. When it comes to determining a likely nursing facility candidate from someone likely to remain in the community, a number of significant variables emerge. A logistic regression analysis was run using several of the variables in this section. The dependent variable was HCBS location of care. The benefit of a multivariate approach like logistic regression is that it allows us to measure the independent effects for a number of variables; e.g., Available Caregiver, Caregiver in Crisis. 32 Final Report (Fiscal Year ) June 30, 2017

33 Table 12. Logistic Regression, HCBS Enrollee = Yes, August 1, December 31, 2015 Parameter Estimate Std. Error Odds Ratio Intercept Has Caregiver In South Florida Is Non-White Needs Help Dressing Needs Help Transferring Caregiver In Crisis Needs Help Eating Is Age 85 Plus Needs Help Bathing Needs Help Walking Needs Help Toileting N = 31,313 Source: Florida Department of Elder Affairs CIRTS and Managed Care Plan Roster files. The ADLs from the tables above were reclassified into dichotomous (Yes/No) variables. When originally coded as requiring some assistance or total dependence, they were coded as NeedsHelp.. = Yes. In South Florida was coded Yes for enrollees located in Regions 10 and 11, predominantly Broward and Miami-Dade Counties. The most interesting column in the table above is the Odds Ratio column. It provides the likelihood that someone who has that attribute (i.e., Parameter ) is receiving HCBS services when compared against someone who does not have that attribute. When the Odds Ratio value is approximately 1.0, it means that the variable has little or no effect on predicting who is enrolled in HCBS services. A number much greater than zero means that the variable is effective at discriminating who is likely to use HCBS services relative to nursing facility services. A number significantly less than one means the variable is more predictive of nursing facility residents than HCBS service recipients. The most telling odds ratio is Has Caregiver. At 7.176, this says that all other factors considered, if one has an available caregiver, the odds of receiving care in HCBS settings is 7 times higher than the odds of receiving care in a nursing facility. Residing in South Florida also increases the odds of home/community-based care by times. This may reflect a more widespread availability of HCBS support services in South Florida or cultural preferences for HCBS in ethnic groups that are prevalent in South Florida. Which factors are most predictive of nursing facility residency? Requiring assistance with toileting is the strongest predictor among these variables. Requiring toileting assistance reduces the odds of home/community-based care by more than 50%. Dependence on physical assistance with bathing and walking have similar independent effects on nursing facility residency according to the table above. Requiring assistance with dressing or transferring has little or no independent effect on the likelihood that an enrollee receives assistance through HCBS services or in a nursing facility setting. The pseudo r 2 for this model 8 is For an individual-level model, this is impressive. There are many individual pressures and considerations in the decision of care in a 8 A pseudo r 2 is a measure of the goodness of fit for a logistic regression model. It is used to assess how well the variables in the model predict the value of the dependent variable. A perfect prediction for every observation would have a pseudo r 2 of 1.0. No relationship at all would result in a pseudo r 2 of Final Report (Fiscal Year ) June 30, 2017

34 nursing facility versus a community setting. The pseudo r 2 indicates this model explains approximately 22% of the variation in the nursing facility versus HCBS care decision for individual enrollees. I: Access to Care Purpose The Independent Assessment of the Florida Statewide Medicaid Managed Care Long-term Care Program first examined the impact on enrollees access to care during state fiscal year The evaluation team and the Agency identified key issues of importance to policy makers and LTC stakeholders. They developed four research questions (RQs) to guide this evaluation, which uses quantitative analytical methods to support findings. Where the evaluation team draws comparisons between access to care under the LTC program versus under legacy waiver programs, the Pre-LTC analytic period comprises state fiscal year Appendix 1 of this report provides further information on the methodology and data sources. The four RQs listed below were used to guide the evaluation of access to care. 1. Have there been changes in the accessibility of services for enrollees compared to the previous LTC programs? 2. How has the population being served in the LTC program shifted (characteristics of the population and service utilization) between nursing facilities and HCBS over time? What LTC plan strategies are impacting these shifts? 3. Do plans offer additional (expanded) benefits and ways to access services, including a Participant Directed Option (PDO), and to what extent do enrollees use these services? 4. Are there disparities by racial and ethnic groups in enrollees placements in certain settings and utilization of services? Findings RQ1: Have there been changes in the accessibility of services for enrollees compared to the previous LTC programs? Given the primary intent of LTC program policy to shift enrollees and, therefore, service delivery to less restrictive home and community-based settings, it is important to demonstrate that access to services is at least maintained, if not improved, compared with access under legacy HCBS waiver programs. The evaluation team assesses access to service delivery by means of several measures: provider network size, service utilization rates, and a time to first service delivery measure. I. Provider Network Size This section addresses RQ1 by using each LTC plan s Provider Network Verification (PNV) files as source data. The PNV files categorize providers as solo practices, group practices and facilities, and hospitals. This report excludes hospitals, as these services are not covered by the LTC program, and summarizes the data for individual providers and group practice providers and facilities. Each LTC provider record indicates provision of one or more of the 26 LTC services defined in the LTC contracts, which are designated by a Plan Benefit Code. 34 Final Report (Fiscal Year ) June 30, 2017

35 Table I.1 summarizes the count of providers who offer each LTC service, grouped separately by individual providers and group provider or facility locations. Many providers offer more than one LTC service. Respite care services are the most commonly offered service (combined count of 7,093), followed by assisted living facility services (combined count of 4,758), behavior management services (combined count of 3,184), and adult day care services (combined count of 3,100). Personal emergency response system (PERS) and hospice services are the least commonly offered services (combined count of 125 and 141, respectively). Table I.1. Number of Providers Offering Each LTC Service, LTC period LTC Service Category Count of Individual Providers Offering the LTC Service Count of Provider Groups and Facilities Offering the LTC Service Combined Counts Adult Companion 659 1,512 2,171 Adult Day Care (Adult Day Health Care) 933 2,167 3,100 Assisted Living Facility Services 1,447 3,311 4,758 Assistive Care Services Attendant Care ,455 Behavior Management 1,009 2,175 3,184 Caregiver Training ,065 Case Management Home Accessibility Adaptation Home Delivered Meals Homemaker 671 1,522 2,193 Hospice Intermittent and Skilled Nursing ,390 Medication Administration ,240 Medication Management ,261 Medical Equipment & Supplies 1, ,008 Nutritional Assessment and Risk Reduction Nursing Facility Care ,431 Personal Care 892 1,225 2,117 Personal Emergency Response System (PERS) Respite Care 2,538 4,555 7,093 Transportation Occupational Therapy 384 1,124 1,508 Physical Therapy 387 1,159 1,546 Respiratory Therapy Speech Therapy 383 1,065 1,448 Note: Most providers offer multiple services. Provider counts do not necessarily correlate with provider size or capacity to provide services. Source: LTC PNV files as of April 2015, limited to provider service location record types 4 and 5 Table I.2 summarizes facility counts by county. The Agency considers facility-based services as those that enrollees receive from the residential facility in which they live; these include nursing facility, assisted living, adult family care home (AFCH), and assistive care services. 9 Ideally, each county will contain at least two facilities per category in order to meet facility-based service provider requirements. 10 However, the Agency may give written 9 Note while these are defined as facility-based services, HCBS services are also provided to enrollees in ALFs and AFCHs. 10 Florida Statewide Medicaid Managed Care (SMMC) Long-term Care (LTC) Program Attachment II Core Contract Provisions, Exhibit II-B Effective Date: November 1, Final Report (Fiscal Year ) June 30, 2017

36 permission to count facilities outside of a given county toward its required quota when a plan demonstrates that it cannot reasonably meet this requirement. Note that AFCHs fall under the larger category of assistive care services (ACS), and most counties only contain this type of ACS facility. Other ACS facilities include residential treatment facilities and ACS facilities with extended congregate care licenses. Readers familiar with Florida s counties will note that more rural counties contain fewer facilities overall. The only counties that do not have at least one reported nursing facility include Glades, Liberty, Union, and Wakulla County; additionally, Union and Wakulla County contain zero reported facility-based providers. These four counties are some of the least populated counties in the state. Miami-Dade and Broward County demonstrate disproportionately high counts of assisted living facilities, which exceed expectations in spite of their large population sizes, perhaps, due to lasting effects of legacy waiver programs that emphasized transition into home and community-based residential settings. 11 County Table I.2. Count of In-Network Facilities by County, LTC Period Count of Nursing Facilities Count of Assisted Living Facilities Count of Adult Family Care Homes Count of "other" Assistive Care Facilities Alachua Baker Bay Bradford Brevard Broward Calhoun Charlotte Citrus Clay Collier Columbia Dade (Miami-Dade) 71 1, DeSoto Dixie Duval Escambia Flagler Franklin Gadsden Gilchrist Glades Gulf Hamilton Hardee Hendry Hernando Highlands Hillsborough Holmes Indian River Jackson Florida Department of Elder Affair s 2009 State Profile Tool, July Final Report (Fiscal Year ) June 30, 2017

37 County Count of Nursing Facilities Count of Assisted Living Facilities Count of Adult Family Care Homes Count of "other" Assistive Care Facilities Jefferson Lafayette Lake Lee Leon Levy Liberty Madison Manatee Marion Martin Monroe Nassau Okaloosa Okeechobee Orange Osceola Palm Beach Pasco Pinellas Polk Putnam St. Johns St. Lucie Santa Rosa Sarasota Seminole Sumter Suwannee Taylor Union Volusia Wakulla Walton Washington All Counties 752 3, Source: LTC PNV files as of April 2015, limited to provider service location record type 5 II. Service Utilization (HCBS only) Comparing service utilization metrics pre- and post-ltc program implementation provides insight into enrollees access to HCBS services under the LTC program versus under legacy HCBS waiver programs. Preferably, these metrics would compare types and intensity of service use; however, the evaluation team felt that there are too many unaccountable duplicate records in the LTC encounter data, which impedes a valid fine-grained analysis of service use intensity. Instead, analysts created a measure of the number of unique persons in HCBS settings who received each service in a given month during the Pre-LTC and LTC periods. Nevertheless, the evaluation team would like to note that the encounter data quality has improved compared to the quality of previous rounds of encounter data; nevertheless, there remains room for improvement. 37 Final Report (Fiscal Year ) June 30, 2017

38 Table I.3 reports the average monthly percentage of enrollees in HCBS settings who received a given LTC service. Overall, this high-level tabulation of service use indicates that there was relatively little change between the Pre-LTC and LTC periods. The highest increase in service use during the LTC period occurred for homemaker, home health, and personal care services, with 10.6%, 6.2%, and 5.6% more service recipients per month on average, respectively. The largest decrease in service use during the LTC period occurred for assistive care services and medical equipment & supplies, with 5.8% and 5.3% less service recipients per month on average, respectively. Note that, while case management service rates are shown in the table, they are considered invalid for the LTC period, as there is a clear underreporting of case management services. This underreporting occurs because most plans use salaried case workers who do not bill based on units of services. The Agency is working with the plans to develop a resolution. The evaluation team hopes this resolution will involve a system of reporting for the units of case management provided, regardless of payment structure. Table I.3. Mean Monthly Percentage of Enrollees in HCBS Settings Receiving Each LTC Service, Pre-LTC and LTC Period Service Category Mean Monthly Count Unique Users (HCBS only) Pre-LTC SFY1213 Mean Monthly % of HCBS Enrollees Receiving Service Std. Dev. Mean Monthly Count Unique Users (HCBS only) LTC SFY1415 Mean Monthly % of HCBS Enrollees Receiving Service Std. Dev. Adult Companion 2, % 0.2% 4, % 0.8% Adult Day Health Care 1, % 0.6% 2, % 0.2% Assisted Living 10, % 1.9% 11, % 0.6% Assistive Care Services 3, % 0.3% 1, % 0.1% Attendant Care Services 1, % 0.2% % 0.1% Behavioral Management % 0.1% % 0.0% Caregiver Training 8 0.0% 0.0% 8 0.0% 0.0% Case Management* 27, % 6.2% 5, % 6.6% Home Accessibility % 0.0% % 0.0% Adaptations Home Delivered Meals 7, % 0.9% 9, % 0.7% Home Health 1, % 0.1% 3, % 0.8% Homemaker 10, % 1.6% 16, % 1.6% Intermittent and Skilled Nursing % 0.0% % 0.0% Medical Equipment & 13, % 1.8% 13, % 1.5% Supplies Medication Administration % 0.0% 38 Final Report (Fiscal Year ) June 30, 2017

39 Service Category Mean Monthly Count Unique Users (HCBS only) Pre-LTC SFY1213 Mean Monthly % of HCBS Enrollees Receiving Service Std. Dev. Mean Monthly Count Unique Users (HCBS only) LTC SFY1415 Mean Monthly % of HCBS Enrollees Receiving Service Std. Dev. Medication Management % 0.0% % 0.0% Nutritional Assessment & % 0.1% % 0.0% Risk Occupational Therapy % 0.0% % 0.0% PERS 6, % 0.8% 7, % 1.0% Personal Care 9, % 2.5% 12, % 0.6% Physical Therapy % 0.0% % 0.0% Respiratory Therapy 2 0.0% 0.0% 5 0.0% 0.0% Respite Care 3, % 0.3% 3, % 0.3% Speech Therapy 3 0.0% 0.0% % 0.0% Transportation 2, % 0.3% 1, % 0.1% Note: The mean monthly count of HCBS enrollees in the Pre-LTC period was 36,700; in the LTC period, the mean count was 41,626 HCBS enrollees. *Not an accurate representation of case management services in the LTC period due to administrative challenges with reporting. Sources: AHCA s LTC service category crosswalk, FSU created enrollee residency file (see Appendix 1), Medicaid FFS claims, NHD encounter records, LTC encounter records III. Time to First Service Delivery (HCBS only) The time to first service delivery provides a crude measure of the timeliness of LTC service receipt. This measure is limited to new enrollees who entered into a Pre-LTC waiver program at some point during SFY1213 and new enrollees who entered into the LTC program at some point during SFY1415, who also resided in a HCBS setting upon program entry. Time to first service delivery, hence, is calculated as the number of days from the first day of program entry to first service receipt. If an enrollee receives a service on their entry date, then his/her time to first service is zero days. It is evident from Table I.4 that there has been little change in time to first service delivery in the Pre-LTC and LTC period and that the average time to first service delivery is about 11 days for both periods. Table I.4. Mean Number of Days until First Service Delivery, Pre-LTC and LTC Period Pre-LTC SFY1213 Count new program entrants (only those whose first Pre-LTC month was in an HCBS waiver* program) Mean number of Days until 1st service delivery Std. Dev. LTC SFY1415 Count new program entrants (only those whose first LTC month was in an HCBS setting) Mean number of Days until 1st service delivery Std. Dev. 3, , *The largest waiver program, NHD, is excluded because NHD claims were all reported as occurring on the 1st of each month. Sources: AHCA s LTC service category crosswalk, FSU created enrollee residency file (see Appendix 1), Medicaid FFS claims, LTC encounter records. 39 Final Report (Fiscal Year ) June 30, 2017

40 RQ2: How has the population being served in the LTC program shifted (characteristics of the population and service utilization) between nursing facilities and HCBS over time? What LTC plan strategies are impacting these shifts? Reducing the proportion of LTC service recipients who receive care in a nursing facility and increasing the proportion who receive care in home and community-based settings is an important policy goal of the Agency. Hence, RQ2 examines global enrollee location of care rates in the LTC period compared with these rates in the Pre-LTC period as well as shifts in location over time. Given the frailty and vulnerability of the LTC program population, central to this RQ is an assessment of the success of transitions into HCBS settings for enrollees who once resided in nursing facility settings. Accordingly, an accurate monthly record of enrollee location during both program periods is essential for this RQ. The evaluation team relied on the minimum data set (MDS) records to determine enrollee location in a given month. Readers should refer to Appendix 1 for the detailed methodology the evaluation team implemented in order to determine an enrollee s monthly location. I. Enrollee Location of Care Table I.5 summarizes member month counts by NF and HCBS locations of care for each period. This table provides a gross overview of how the Florida Medicaid population of enrollees who received a determination of need for nursing facility level of care was served in the LTC versus the Pre-LTC period. There has been a clear reduction in the number of member months where enrollees resided in nursing facilities in the LTC period (51.1% of total MM) versus the Pre-LTC period (60.0%) and a corresponding increase in the number of member months where enrollees resided in HCBS settings (48.9% of total member months in the LTC period versus 40.0% in the Pre-LTC period). This simple metric reveals that the plans are fulfilling the Agency s stated policy goal of shifting enrollees into HCBS settings. Further analysis below provides insight into the population shifts over time. Evaluation Period Table I.5. Total Member Months by Location of Care, Pre-LTC and LTC Period Total Member Months Count of NF Months Proportion of NF Months Count of HCBS Months Proportion of HCBS Months Pre-LTC 1,100, , % 440, % LTC 1,021, , % 499, % Source: FSU created enrollee residency file (see Appendix 1) Table I.6 presents enrollee location of care rates over time. While nursing facility residency rates increased by about 1 percentage point throughout the Pre-LTC period, this rate decreased at a fairly constant rate, by about 3 percentage points total, throughout the LTC period. These results indicate that the LTC plans are steadily transitioning enrollees into HCBS settings, yet it remains unclear whether changes in these residency rates are driven by plans enrolling persons newly eligible for the LTC program who already reside in home/communitybased settings or by shifting earlier program entrants who resided in nursing facilities out of these facilities and into home and community-based settings. Table I.7 below addresses that distinction. 40 Final Report (Fiscal Year ) June 30, 2017

41 Evaluation Period Pre-LTC (SFY1213) LTC (SFY1415) Table I.6. Enrollee Location of Care Rates Over Time, 6 Pre-LTC and LTC Period State Fiscal Month* Count Unique Persons Number in NFs Proportion in NFs Number in HCBS settings Proportion in HCBS settings 1 92,187 54, % 37, % 2 92,733 55, % 37, % 3 92,676 55, % 37, % 4 92,974 55, % 37, % 5 92,841 55, % 37, % 6 92,705 55, % 37, % 7 92,726 55, % 37, % 8 91,576 54, % 36, % 9 90,818 54, % 35, % 10 90,259 54, % 35, % 11 89,974 54, % 35, % 12 89,263 54, % 34, % 1 83,438 43, % 39, % 2 83,436 43, % 39, % 3 83,619 43, % 39, % 4 84,293 43, % 40, % 5 84,918 43, % 40, % 6 85,237 43, % 41, % 7 85,407 43, % 41, % 8 85,290 43, % 42, % 9 85,926 43, % 42, % 10 86,285 43, % 42, % 11** 86,804 42, % 43, % 12** 87,168 42, % 44, % *Month 1 of the Pre-LTC period corresponds to July 2012, while month 12 corresponds to June For the LTC period, month 1 corresponds to July 2014, while month 12 corresponds to June This patterns hold throughout any analyses that reference state fiscal months. **Slight overestimation of the HCBS rate and underestimation of the NF rate due to missing inpatient records toward the end of SFY1415. It is likely that the month 11 rates are closer to 50.1% NF/49.9% HCBS and the month 12 rates are closer to 49.9% NF/50.1% HCBS. Source: FSU created enrollee residency file (see Appendix 1). Figure I.1 on the following page presents the information from Table I.6 in graphical form. 41 Final Report (Fiscal Year ) June 30, 2017

42 Figure I.1. Proportion of Enrollees Residing in a Nursing Facility or and HCBS Setting Over Time, Pre-LTC and LTC Period 70% 60% 50% 40% 30% 20% 10% 0% Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Pre-LTC SFY1213 Proportion Placed in NF LTC SFY1415 Proportion Placed in HCBS Source: FSU created enrollee residency file (see Appendix 1) II. Transition Rates and Transition into the Community Success Table I.7 summarizes enrollee rates for transitions into the community from a NF and for transitions into a NF from the community for the Pre-LTC and LTC period. These descriptive statistics provide a glimpse into how successful LTC plans are at shifting the LTC population into less restrictive settings, with the assumption that they should aim to minimize the number of times frail enrollees shift between settings. A transition into the community is considered successful when an enrollee shifts out of a NF and into a home and community-based setting where he/she subsequently resides for at least 60 days. Conversely, an unsuccessful transition into the community is any shift into the community from a NF after which an enrollee returns to a NF within fewer than 60 days. A nursing facility stay is considered an interruption in HCBS services when an enrollee transitions into a NF where he/she subsequently resides for fewer than 60 days. These incidents frequently follow inpatient hospital stays, so they often represent spells of acute care. Transitions are labeled indeterminate when there is insufficient observation time to determine whether or not an enrollee reached the 60-day threshold (due to loss of eligibility, death, or shifting toward the end of the LTC period 12 ). In general, enrollees were more than twice as likely to change settings during the LTC period (20.4% shifted at least once versus 79.6% who experienced no change in their location of care) than during the Pre-LTC period (8.5% shifted at least once versus 91.5% who experienced no change in their location of care). This increase in shifts has both positive and potentially negative implications. On the positive side, enrollees were more than twice as likely to successfully transition into HCBS settings and remain there for the duration of the evaluation year in the LTC period (4.0%) than in the Pre-LTC period (1.7%). Additionally, some of the indeterminate transitions into the community in the LTC period (1.8%) may actually represent successful transitions into the 12 The evaluation team did not have eligibility data beyond June of Final Report (Fiscal Year ) June 30, 2017

43 community that occurred toward the end of SFY1415 but are obscured due to truncating the observation time at the end of June Conversely, some of the other differences in location change rates between the two evaluation periods suggest potentially problematic, negative shifts. Specifically, 1.1% of enrollees who successfully transitioned into HCBS settings for at least 60 days in the LTC period eventually transitioned back into a nursing facility compared with only 0.1% of enrollees who met these criteria in the Pre-LTC period. Moreover, 3.5% of enrollees who transitioned experienced unsuccessful transitions into the community in the LTC period compared with only 0.2% of enrollees who met these criteria in the Pre-LTC period. Finally, 1.5% of enrollees experienced three or more shifts in the Pre-LTC period, but more than twice as many enrollees (3.5%) experienced three plus shifts in the LTC period. It is clear from these results there is a risk/reward balance to transitioning enrollees into HCBS settings. This risk/reward consideration is more noteworthy in the LTC period than in the Pre-LTC period given that there were fewer location shifts before the LTC program began. Table I.7. Transition Rates, Pre-LTC and LTC Period Pre-LTC LTC Moves Start Movement summary Persons % Persons % Started in HCBS Started in NF Started in HCBS Started in NF Started in HCBS Started in NF Started in HCBS Started in NF Resided in the community for the entire observation length 36, % 44, % Resided in a NF for the entire observation length 74, % 42, % Transitioned into a NF 1, % 2, % Indeterminate transition into a NF % 1, % Successful transition into the community 2, % 4, % Indeterminate transition into the community % 2, % NF interruption, then returned to HCBS 2, % 3, % Transitioned into a NF, then successfully transitioned back into the community % % Transitioned into a NF, then an indeterminate transition back into the community % % Successful transition into the community, then transitioned back into a NF % % Successful transition into the community, then an indeterminate transition back into a NF % % Unsuccessful transition into the community, then returned to a NF % 3, % NF interruption, returned to the community, then transitioned into a NF % % NF interruption, returned to the community, then an indeterminate transition into a NF % % Transitioned into a NF, successfully transitioned back into the community, then transitioned back % % into a NF Transitioned into a NF, successfully transitioned back into the community, then an indeterminate % % transition into a NF Transitioned into a NF, unsuccessfully transitioned into the community, then returned to a NF % % Successful transition, NF interruption, then returned to the community % % 43 Final Report (Fiscal Year ) June 30, 2017

44 Pre-LTC LTC Moves Start Movement summary Persons % Persons % Successful transition, transitioned back into a NF, then another successful transition into the % % community Successful transition, transition back into a NF, then an indeterminate transition in the community 7 0.0% % Unsuccessful transition, returned to a NF, then successfully transitioned into the community % % Unsuccessful transition, returned to a NF, then an indeterminate transition into the community % % Started in Started in the community and moved 4+ times % 1, % 4+ HCBS Started in NF Started in a NF and moved 4+ times % 1, % Source: FSU created enrollee residency file (see Appendix 1) III. Nursing Facility Residency Risk The evaluation team conducted an analysis to determine the likelihood of 60 consecutive days of nursing facility residency for a given enrollee in the LTC period versus the Pre-LTC period. The intent of this analysis is to provide a synopsis of which enrollees the LTC plans have targeted for transition to HCBS settings. Due to time and data limitations, the analyst was unable to control for diagnostic criteria, frailty levels, or ADL limitations. Hence, the results presented below are purely descriptive and do not represent a causal analysis of any mechanisms driving location shifts in the LTC program. See Appendix 1 for more information on the model the analyst developed for this metric. Table I.8 and Figure I.2 show that, generally, all enrollees were more likely to shift into HCBS settings in the LTC period, regardless of demographic characteristics. However, it also appears that the LTC plans disproportionately shifted younger enrollees into HCBS settings. It is improbable that age itself drove these shifts; rather, it is probable that age reflects differences in health status that determine who may successfully transition into the community. The younger population is more likely to receive LTC services due to disability, while the older population is more likely to receive LTC services due to the correlates of aging, including frailty and cognitive decline. It may be the case that HCBS sites are better equipped to serve enrollees with disabilities than enrollees with cognitive (or other) impairments. Moreover, younger, disabled enrollees may be more motivated to transition and may have greater access to an available caregiver(s). Additionally, NF personnel may be more motivated to shift younger residents, who are often outliers in their facilities, into home/community-based settings. 44 Final Report (Fiscal Year ) June 30, 2017

45 Table I.8. Predicted Probabilities of Nursing Facility Residency in the Pre-LTC and LTC Period Pre-LTC Period Age Category Female Male Black Hispanic White Other Black Hispanic White Other % 29.6% 43.8% 26.0% 54.8% 37.1% 52.2% 33.0% % 45.5% 60.7% 41.1% 70.6% 53.9% 68.3% 49.4% % 60.2% 73.6% 55.8% 81.3% 67.8% 79.6% 63.8% % 62.7% 75.6% 58.4% 82.9% 70.1% 81.3% 66.2% % 40.4% 55.6% 36.1% 66.0% 48.6% 63.6% 44.2% % 41.6% 56.8% 37.3% 67.1% 49.9% 64.7% 45.4% % 47.4% 62.5% 43.0% 72.1% 55.7% 69.9% 51.3% Age Category LTC Period Female Male Black Hispanic White Other Black Hispanic White Other % 17.5% 27.6% 16.0% 36.1% 21.3% 32.6% 19.5% % 28.1% 41.0% 26.0% 50.6% 33.1% 46.8% 30.8% % 34.1% 47.9% 31.8% 57.5% 39.6% 53.7% 37.1% % 39.5% 53.5% 36.9% 62.9% 45.2% 59.3% 42.6% % 37.1% 51.1% 34.7% 60.6% 42.7% 56.9% 40.2% % 39.5% 53.6% 37.0% 63.0% 45.2% 59.3% 42.6% % 45.9% 60.0% 43.3% 68.8% 51.7% 65.4% 49.1% Sources: FSU created enrollee residency file (see Appendix 1), Medicaid Finder File (for demographic variables) Figure I.2 shows the change in the predicted probabilities of nursing facility residency broken down by enrollee age, sex, and race/ethnicity. The longer bars for younger age groups visually displays the finding that plans are more likely to shift younger enrollees into community-based settings. Figure I.2. Change in the Predicted Probabilities of Nursing Facility Residency in the LTC Period versus the Pre-LTC Period Female Male Percentage Change 0% -5% -10% -15% -20% -25% -30% -35% -40% -45% -50% Black Hispanic White Other Black Hispanic White Other Age Age Age Age Age Age Age 85+ Sources: FSU created enrollee residency file (see Appendix 1), Medicaid Finder File (for demographic variables) 45 Final Report (Fiscal Year ) June 30, 2017

46 IV. CDPS Risk Scores Table I.9 presents the mean Chronic Illness and Disability Payment System (CDPS) risk scores for enrollees who were enrolled in the LTC program in the first month 13 of the evaluation period (July 2014). These risk scores are a predictor of medical service use and cost. For the purposes of this report, they are being used as a proxy for level of frailty. As expected, the population of enrollees who resided in a nursing facility for at least 60 days during the evaluation period has a much higher average risk score (4.56) than the population of enrollees who resided in HCBS settings for the majority of the evaluation period (average of 2.16). Please note the evaluation team presents these results with one caveat; the team remains uncertain as to when a 0 indicates a missing risk score or a valid risk score of 0. Accordingly, this uncertainty should be taken into account when interpreting the scores. It is possible that the mean risk score for both groups is higher than those reported below, especially for enrollees with NF residency Location Table I.9. Mean CDPS Risk Scores for Enrollees in the First Month of the LTC Period N Mean Risk Score Std. Dev. Min. Median Max. NF 43, HCBS 39, Source: AHCA eligibility data V. Contracted Bed Counts How many risk scores are zero? 6,112 (14.0% of enrollees in NFs) 3,031 (7.6% of enrollees in HCBS settings) Table I.10 lists the number of beds each plan has contracted with nursing facilities, assisted living facilities, adult family care homes, and other ACS facilities (those in addition to AFCHs) by region. 14 Note there is significant overlap in bed counts across the plans by region, i.e., it is common for two or more plans to have a contract for all of the Medicaid certified beds in the same facility. Additionally, data validation revealed that the beds in use variable in the PNV files is not a valid measure of bed occupancy rates. Given this data limitation combined with the overlap in bed counts, it was not possible to assess if the LTC plans meet the facility-based service needs of their enrollees at the regional level. Nevertheless, at face value each plan appears to contract with nursing facilities and assisted living facilities for more than enough beds to meet the potential in-facility service needs of their enrollees. Conversely, the plans do not have a robust network of assistive care service facilities, including adult family care homes. 13 The evaluation team chose the first month of the LTC evaluation year because data validation results indicated the CDPS scores in the eligibility data had so many 0 s in months 4-12 that it was impossible the data field contained valid risk adjustment scores after month three. 14 Note that the bed counts for AEC are based on service location type 4 records, as AEC s individual and group practice/facility records are not separated by provider type in the PNV files the evaluation team received. 46 Final Report (Fiscal Year ) June 30, 2017

47 Table I.10. Contracted Bed Counts by Plan and Region in the LTC Period Plan Region* Mean Monthly Count of Enrollees Nursing Facility Contracted Bed Counts Assisted Living Facility Adult Family Care Home Other ACS Facility AEC** AMG COV HUM MOL SUN URA 1 1,427 2, ,294 3, , ,251 8, ,449 9, , ,215 8, ,014 7, ,365 8, , ,260 5,865 1, , , , ,596 4,132 6, ,954 7,282 5, ,069 4,723 1, ,627 3, ,332 4,409 4, ,206 4,704 8, ,260 2,380 1, ,307 1,981 5, ,856 4,602 8, ,518 7,308 5, ,219 5,996 4, ,640 6,467 6, ,549 2, ,731 6,379 2, ,696 8,313 1, ,877 7,603 5, ,546 7,330 6, ,013 8,477 3, ,888 6,656 4, ,296 6,990 4, ,625 3,973 5, ,306 6,626 6, ,387 3, ,924 6,615 2, ,262 6,308 2, ,442 8,284 5, Final Report (Fiscal Year ) June 30, 2017

48 Plan Region* Mean Monthly Count of Enrollees Nursing Facility Contracted Bed Counts Assisted Living Facility Adult Family Care Home Other ACS Facility 6 1,991 7,830 4, ,078 8,061 3, ,284 6,139 3, ,489 4,614 2, ,883 5,874 2, *Official AHCA contracted regions in bold **Service location type 4 used Source: LTC PNV files as of April 2015, limited to provider service location record type 5, except for AEC plan (type 4 used) RQ3: Do plans offer additional (expanded) benefits and ways to access services, including a Participant Directed Option (PDO), and to what extent do enrollees use these services? RQ3 investigates expanded benefit service offerings and Participant Directed Option (PDO) participation and service use in the LTC period. Expanded Benefits Table I.11 presents the expanded benefits offered by each LTC plan. Blank cells indicate an LTC plan did not offer the particular expanded benefit. The total number of expanded benefits offered across all plans was 60 (with a range of 5-13 offered by each plan). All plans offered every expanded benefit indicated in each region for which they were contracted. Dental Services, Over-The-Counter (OTC) Medication/Supplies, and Support to Transition Out of a Nursing Facility were the most commonly offered expanded benefits; in fact, every plan offered these services. Table I.11. List of Expanded Benefits Offered by the Seven Plans in the LTC Period Expanded Benefits that appear in Choice Counseling materials AEC AMG COV HUM MOL SUN URA Total ALF/AFCH Bed Hold X X X X X X 6 Cellular Phone Service X X X X 4 Dental Services X X X X X X X 7 Emergency Financial Assistance X 1 Hearing Evaluation X X X X 4 Mobile Personal Emergency Response System X 1 Non-Medical Transportation X X X 3 OTC Medications/Supplies X X X X X X X 7 Support to Transition Out of a Nursing Facility X X X X X X X 7 Vision Services X X X X X X 6 Wellness Grocery Discount X 1 Expanded Benefits that do not appear in Choice Counseling materials AEC AMG COV HUM MOL SUN URA Total Box Fan X 1 Caregiver Information/Support X X 2 Document Keeper X X 2 48 Final Report (Fiscal Year ) June 30, 2017

49 Emergency Meal Supply X X 2 Household Set-Up Kit X 1 Welcome Home Basket X 1 Nurse Helpline Services X X 2 Pill Organizer X X 2 Total Sources: LTC Plan contracts, A Snapshot of the Florida Medicaid Long-term Care Program PDO Overview The PDO is a service delivery model that empowers Medicaid recipients enrolled in an LTC plan by allowing them to hire, train, supervise, and dismiss direct service worker(s) providing certain long-term care services. The PDO is available to all long-term care enrollees who (a) have any PDO service listed on their authorized care plan and (b) live in their own home or family home. In accordance with state and federal regulations, PDO services must be medically necessary and cost-effective. Participant independence and personal choice is the primary focus of the PDO. Enrollees who select this option must be interested in actively managing their own health care and be willing to take responsibility for hiring and managing their direct service worker(s). A PDO participant may choose a representative to assist with the employer responsibilities of the PDO. The representative cannot be compensated for their services nor be a direct service worker. Participants may hire any individual of their choosing to provide their PDO services, including family members, neighbors, or friends. The enrollee s LTC plan is responsible for ensuring each direct service worker only receives payment for the hours and approved PDO services that are listed in the participant s authorized care plan and on the Participant/Direct Service Worker Agreement. A participant s direct service worker(s) does not have to be in the LTC plan s provider network. Key components of the PDO include: 1. PDO Services These are the services enrollees may self-direct. An enrollee must have at least one of these services on their care plan in order to participate in the PDO: adult companion care, attendant care, homemaker services, intermittent and skilled nursing, or personal care services. 2. Participant This is the Medicaid LTC program enrollee who has chosen to participate in the PDO for one or more services and who acts as the employer. 3. Case Manager In addition to the duties outlined in the LTC plan contract, the case manager is responsible for providing ongoing PDO-related technical assistance to the participant as needed and requested. This responsibility includes providing initial PDO training to the participant upon opting to participate in the PDO. 4. Direct Service Worker This is the employee directly-hired by a participant who provides PDO services as authorized under the care plan. The direct service worker(s) may be any qualified individual chosen by the participant. The direct service worker(s) is paid by the LTC plan based on a set rate. 5. Fiscal/Employer Agent (F/EA) Services Each managed care plan is responsible for providing F/EA services, as described in the LTC plan contract, to the participants who choose the PDO. The F/EA 49 Final Report (Fiscal Year ) June 30, 2017

50 functions include payroll services and processing, filing, and paying all state and federal taxes on behalf of participants and their direct service workers. I. PDO Enrollment Rates Table I.12 summarizes PDO enrollment, limited to home-based residency months, for the entire LTC period. Of the total eligible member months spent in home-based settings, 4.5% were months with PDO enrollment, representing 2,109 total unique enrollees. Table I.12. Overall Home-Based Member Months with PDO Enrollment, LTC Period Count of unique PDO enrollees throughout SFY1415 Total Member Months enrolled in the PDO Total Home-Based Member Months Percentage of Home-Based Member Months with PDO Enrollment 2,109 16, , % *Excludes July 2014 for AEC enrollees because missing plan PDO data for that month. Sources: FSU created enrollee residency file (see Appendix 1), LTC Plan PDO Roster reports Table I.13 summarizes PDO enrollment by plan for the entire LTC period, again, limited to home-based months. Sunshine State Health Plan enrollees experienced the highest rate of PDO enrollment (8.5% of eligible member months), while Molina Health Care enrollees experienced the lowest rate of PDO enrollment (1.1% of homebased member months). Plan Table I.13. Home-Based Member Months with PDO Enrollment by Plan, LTC Period Count of unique PDO enrollees throughout SFY1415 Total Member Months enrolled in the PDO Total Home-Based Member Months Percentage of Home-Based Member Months with PDO Enrollment AEC* , % AMG , % COV , % HUM 182 1,134 23, % MOL , % SUN 1,137 9, , % URA 481 3, , % Missing - - 5,135 - Overall 2,109 16, , % Note: Some overlap between plan-level participant counts for any participants who switched LTC plans in SFY1415. *Excludes July 2014 Sources: FSU created enrollee residency file (see Appendix 1), LTC Plan PDO Roster reports, Medicaid FFS claims (for monthly plan participation). 50 Final Report (Fiscal Year ) June 30, 2017

51 Table I.14 summarizes PDO enrollment over time for the LTC period. This table shows that PDO enrollment for home-based recipients steadily increased throughout SFY1415. Table I.14. PDO Enrollment by Month, LTC Period State Fiscal Month Count of PDO Enrollees Count of Home- Based Enrollees per Month Percentage of Home-Based Enrollees Enrolled in the PDO 1* , % , % 3 1,043 29, % 4 1,129 29, % 5 1,244 29, % 6 1,335 30, % 7 1,397 30, % 8 1,485 30, % 9 1,556 31, % 10 1,615 31, % 11 1,685 32, % 12 1,792 32, % *Excludes AEC Sources: LTC Plan PDO Roster reports, FSU created enrollee residency file (see Appendix 1) Note: Some overlap between plan-level participant counts for any participants who switched LTC plans in SFY1415. II. Length of PDO Enrollment Table I.15 and Table I.16 highlight PDO enrollment lengths by plan for the LTC period. Table I.15. PDO Enrollment by Plan, LTC Period Plan Average number of months enrolled in the PDO* Average proportion of SFY1415 enrollment span spent enrolled in the PDO* AEC %** AMG % COV % HUM % MOL % SUN % URA % Overall % Note: Some overlap between plan-level participant counts for any participants who switched LTC plans in SFY1415. *For those with at least 1 month in PDO enrollment in SFY1415 **Excludes July 2014 Sources: LTC Plan PDO Roster reports, Medicaid FFS claims (for monthly plan participation) 51 Final Report (Fiscal Year ) June 30, 2017

52 Plan Table I.16. PDO Enrollment Span by Plan, LTC Period Count of persons enrolled in the PDO for their entire SFY1415 enrollment span Total Unique PDO enrollees in SFY1415 Proportion of persons enrolled in the PDO for their entire SFY1415 enrollment span AEC* % AMG % COV % HUM % MOL % SUN 506 1, % URA % *Excludes July 2014 Sources: LTC Plan PDO Roster reports, Medicaid FFS claims (for monthly plan participation) III. PDO Service Utilization Table I.17 summarizes PDO enrollees selected service rates and rates of services received on average each month. The most frequent service is Personal Care services, with an average of 89.2% of PDO enrollees receiving this service through PDO each month, followed by Homemaker services (at 73.7% on average each month). Few PDO enrollees receive Attendant Care services (at 1.3%) or Intermittent and Skilled Nursing services (at 0.5%) via the PDO. Enrollees likely receive certain services more/less frequently because some service categories appear more/less frequently in their plans of care. Table I.17. Services Selected Through PDO and Received by PDO Enrollees in the LTC Period Selected services through PDO Mean monthly proportion of PDO recipients selecting and receiving: Adult Companion services Attendant Care services Homemaker services Intermittent and Skilled Nursing services Personal Care services 33.0% 1.3% 73.7% 0.5% 89.2% Received services 30.0% 1.5% 69.4% 1.6% 81.2% *Excludes AEC months 1 & 2 Sources: LTC Plan PDO Roster reports, AHCA s LTC service category crosswalk, LTC encounter records RQ4: Are there disparities by racial and ethnic groups in enrollees placements in certain settings and utilization of services? There is long-standing agreement among health policy experts that sociodemographic factors, especially ethnoracial minority status, underlie systematic differences in LTC service utilization and location. Variations in enrollees residential settings may be motivated by more benign differences in sociocultural preferences for care or more circumstantial differences in the availability of a caregiver. However, some ethnoracial minority groups 52 Final Report (Fiscal Year ) June 30, 2017

53 may experience institutional disparities in access to high-quality LTC services and facilities in the United States. 15 As a result, both the types of services received and the extent of utilization can vary widely across individuals and subpopulations within a given state. Accordingly, RQ4 of this evaluation aims to assess differences and potential disparities in access to care among Florida Medicaid s LTC subpopulations. This report focuses on ethnoracial groups; however, there are other subpopulations of interest, e.g., enrollees who reside in rural areas. I. Enrollee Location of Care Table 1.18 provides an overview of the systematic differences in enrollee locations of care by race/ethnicity in both evaluation periods. Hispanic enrollees were far more likely than black enrollees and white enrollees to receive care in HCBS settings in both evaluation periods, while all enrollees, regardless of race or ethnicity, were more likely to receive care in HCBS settings in the LTC period versus the Pre-LTC period. Black enrollees and white enrollees experienced fairly equivalent location of care rates in both evaluation periods. Without measures that are model-adjusted for levels of frailty or morbidity, it is not possible to determine whether these differences in location directly result from sociocultural differences or indirectly result from differences in levels of frailty or need for care. Table I.18. Total Member Months by Location of Care Period Stratified by Race/Ethnicity, Pre-LTC and LTC Period Evaluation Period Race/ Ethnicity Count of Unique Persons Total Member Months Count of NF Months Proportion of NF Months Count of HCBS Months Proportion of HCBS Months Black 19, , , % 56, % Pre-LTC Hispanic 21, ,809 73, % 124, % White 72, , , % 214, % Other 8,451 75,075 30, % 44, % Black 17, ,914 97, % 69, % LTC Hispanic 20, ,075 59, % 131, % White 61, , , % 232, % Other 11,156 99,277 33, % 65, % Sources: FSU created enrollee residency file (see Appendix 1), Medicaid Finder File (for demographic variables) Figure I.3 presents the information from Table I.18 in graphical form. 15 Smith DB, Feng Z, Fennell ML, Zinn J, Mor V. Racial disparities in access to long-term care: The illusive pursuit of equity. J Health Polit Policy Law. 2008; 33(5): Final Report (Fiscal Year ) June 30, 2017

54 Figure I.3. Proportion of Member Months in a NF or HCBS Setting by Race/Ethnicity, Pre-LTC and LTC Period 80% 70% 60% 50% 40% 30% 20% 10% 0% NF HCBS NF HCBS Pre-LTC LTC Black Hisp Other White Sources: FSU created enrollee residency file (see Appendix 1), Medicaid Finder File (for demographic variables) Figures I.4 and I.5 display changes in location rates by race/ethnicity over time during the LTC period. While Black, Hispanic, and White enrollees all experienced reductions in nursing facility residency throughout the course of the LTC period, there were differences in the degree of reduction among these subpopulations. Hispanic enrollees experienced the greatest decline in nursing facility residency with an 11.4% decrease in their rate of NF residency (and a corresponding increase in their rate of HCBS residency) throughout the LTC period. Black enrollees experienced a 6.6% decrease in their rate of NF residency. White enrollees experienced the smallest decline, with a 3.9% decrease in their rate of NF residency throughout the LTC period. Observed differences in the level of change in enrollee location rates over time may be driven by differences (such as caregiver availability) that are not inherent to the subpopulations themselves. For example, there may be variation in individual plans success rates for transitioning enrollees into the community, the Hispanic subpopulation may happen to be more likely to live in regions in which more successful plans are contracted, and/or the Hispanic subpopulation may be experiencing a temporary increase in location shifts that then levels out over time. 54 Final Report (Fiscal Year ) June 30, 2017

55 Figure I.4. Proportion of Enrollees Residing in a Nursing Facility Over Time by Race/Ethnicity, LTC Period 80% 70% 60% 50% 40% 30% 20% Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun LTC period Black Hispanic Other White Sources: FSU created enrollee residency file (see Appendix 1), Medicaid Finder File (for demographic variables) Figure I.5. Proportion of Enrollees Residing in an HCBS Setting Over Time by Race/Ethnicity, LTC Period 80% 70% 60% 50% 40% 30% 20% Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun LTC period Black Hispanic Other White Sources: FSU created enrollee residency file (see Appendix 1), Medicaid Finder File (for demographic variables) 55 Final Report (Fiscal Year ) June 30, 2017

56 II. Service Utilization (HCBS only) Table I.19 reports the average monthly percentage of enrollees in HCBS settings who received a given LTC service by race/ethnicity for the LTC period. The most noticeable difference among HCBS enrollees is for assisted living services; white enrollees were almost 2.5 times more likely than black enrollees to reside in an assisted living facility and were more than 1.5 times as likely as Hispanic enrollees to reside in an assisted living facility (ALF). White enrollees were also more likely to receive assistive care services than both Hispanic and black enrollees. Because Hispanic and black enrollees were less likely to receive assisted living and assistive care services, they were understandably more likely to receive adult day health care (ADHC) services (on average 13.6% and 6.1%, respectively, of Hispanic and black enrollees received ADHC services each month versus 2.8% for white enrollees). Still, Hispanic enrollees were more than twice as likely as black enrollees to receive this service. Hispanic and black enrollees were also more likely than white enrollees to receive home delivered meal, homemaker, medical equipment and supplies, personal care, and respite care services. While these differences are interesting, there is little to be said about their underlying causes based on the current analysis. Table I.19. Mean Monthly Percentage of Enrollees in HCBS Settings Receiving Each LTC Service by Race/Ethnicity, LTC period only Black Hispanic Other White Mean Monthly Pctage of Mean Monthly Pctage of Mean Monthly Pctage of Mean Monthly Pctage of Service Category HCBS Std. HCBS Std. HCBS Std. HCBS Std. Enrollees Dev. Enrollees Dev. Enrollees Dev. Enrollees Dev. Receiving the Service Receiving the Service Receiving the Service Receiving the Service Adult Companion 11.0% 1.0% 11.6% 0.7% 10.4% 1.0% 8.4% 0.8% Adult Day Health Care 6.1% 0.3% 13.6% 0.2% 10.6% 0.2% 2.8% 0.1% Assisted Living 15.0% 0.2% 20.5% 0.7% 21.0% 0.2% 36.1% 0.6% Assistive Care Services 1.5% 0.1% 1.4% 0.1% 2.3% 0.1% 4.0% 0.2% Attendant Care Services 0.4% 0.1% 0.7% 0.1% 0.7% 0.2% 0.5% 0.1% Behavioral Management 0.3% 0.1% 0.4% 0.1% 0.5% 0.1% 0.3% 0.1% Caregiver Training 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Case Management* 14.7% 7.4% 17.9% 3.2% 14.0% 4.7% 11.4% 8.9% Home Accessibility Adaptations 0.3% 0.0% 0.2% 0.0% 0.2% 0.1% 0.2% 0.0% Home Delivered Meals 32.2% 0.5% 26.9% 0.7% 22.8% 0.5% 19.5% 0.8% Home Health 11.8% 0.8% 8.3% 1.1% 10.2% 1.0% 8.4% 0.6% Homemaker 43.1% 1.5% 49.7% 1.9% 42.4% 1.4% 31.7% 1.2% Intermittent and Skilled Nursing 0.3% 0.0% 0.1% 0.0% 0.1% 0.1% 0.1% 0.0% 56 Final Report (Fiscal Year ) June 30, 2017

57 Service Category Mean Monthly Pctage of HCBS Enrollees Receiving the Service Black Hispanic Other White Std. Dev. Mean Monthly Pctage of HCBS Enrollees Receiving the Service Std. Dev. Mean Monthly Pctage of HCBS Enrollees Receiving the Service Std. Dev. Mean Monthly Pctage of HCBS Enrollees Receiving the Service Medical Equipment & Supplies 40.6% 1.0% 39.3% 2.2% 34.6% 1.5% 24.6% 1.1% Medication Administration 0.1% 0.0% 0.0% 0.0% 0.1% 0.0% 0.1% 0.0% Medication Management 0.0% 0.0% 0.1% 0.0% 0.1% 0.0% 0.1% 0.0% Nutritional Assessment & Risk 0.0% 0.0% 0.2% 0.0% 0.1% 0.0% 0.0% 0.0% Occupational Therapy 0.3% 0.1% 0.1% 0.0% 0.2% 0.1% 0.3% 0.1% PERS 22.0% 1.5% 15.4% 0.8% 15.2% 0.7% 17.4% 1.1% Personal Care 33.2% 0.5% 46.2% 0.9% 36.2% 0.4% 20.5% 0.4% Physical Therapy 0.5% 0.1% 0.1% 0.0% 0.4% 0.1% 0.4% 0.1% Respiratory Therapy 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Respite Care 10.4% 0.2% 11.8% 0.5% 10.3% 0.5% 5.9% 0.2% Speech Therapy 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Transportation 2.5% 0.2% 4.4% 0.1% 3.9% 0.1% 1.5% 0.2% *Not an accurate representation of case management services in the LTC period due to administrative challenges with reporting. Sources: AHCA s LTC service category crosswalk, FSU created enrollee residency file (see Appendix 1), Medicaid Finder File (for demographic variables), Medicaid FFS claims, NHD encounter records, LTC encounter records III. Time to First Service Delivery (HCBS only) Table I.20 presents time to first service delivery results for the LTC period stratified by race/ethnicity. As under RQ1, this metric is limited to new enrollees who entered into the LTC program at some point during SFY1415 who also resided in a HCBS setting upon program entry. On average, Hispanic HCBS enrollees received their first home and community-based LTC service 5 days earlier than white HCBS enrollees and 6 days earlier than black HCBS enrollees. The average number of days is quite variable, as indicated by the size of the standard deviations (Std. Dev.) in relation to each of the groups averages, but the underlying cause of these differences is not discernible without additional inferential analysis. Std. Dev. 57 Final Report (Fiscal Year ) June 30, 2017

58 Table I.20. Mean Number of Days until First Service Delivery by Race/Ethnicity, LTC period Race/Ethnicity Count of new program entrants (only those whose first LTC month was in an HCBS setting) Average number of Days until 1st service delivery Std. Dev. Black 1, Hispanic 3, White 5, Other 2, Total 12, Sources: AHCA s LTC service category crosswalk, FSU created enrollee residency file (see Appendix 1), Medicaid Finder File (for demographic variables), Medicaid FFS claims, LTC encounter records IV. Facility Accessibility Table I.21 summarizes the average distances from community-based enrollees home addresses to their nearest nursing facility stratified by race/ethnicity. The results show that Hispanic enrollees live slightly closer to their nearest nursing facility (mean=1.98 miles) compared with black enrollees (mean=2.31 miles) and white enrollees (mean=2.55 miles). There is also less variation in the distances for Hispanic enrollees (SD=1.56 miles, max=19.94 miles) relative to black enrollees (2.54 miles, max=37.25 miles) and white enrollees (SD=2.80 miles, max=39.75 miles). These outcomes suggest that Hispanic enrollees are more likely to live in urban locales than enrollees of non-hispanic ethnicities. Table I.21. Mean Distance to Nearest Nursing Facility for Community-based Enrollees by Race/Ethnicity, LTC period Race/Ethnicity Count Enrollees Mean Distance Std. Dev. Median Distance Farthest Distance Black 4, Hispanic 7, White 19, Other 12, Note: All distances in miles. Source: The Florida Geographic Data Library Figures I.6 and I.7 display maps of the population of LTC community-based enrollees (1.6) and Medicaid certified nursing facilities (I.7) separately. Figure 1.8 displays a map of the Medicaid certified nursing facilities overlaid with the population of community-based LTC enrollees stratified by race/ethnicity. 58 Final Report (Fiscal Year ) June 30, 2017

59 Figure I.6. Map of LTC Enrollees Living in the Community Source: The Florida Geographic Data Library Data processing and cartography: FSU College of Medicine, Department of Behavioral Sciences and Social Medicine, May Final Report (Fiscal Year ) June 30, 2017

60 Figure I.7. Map of Medicaid Certified Nursing Facilities in the State of Florida Source: The Florida Geographic Data Library Data processing and cartography: FSU College of Medicine, Department of Behavioral Sciences and Social Medicine, May Final Report (Fiscal Year ) June 30, 2017

61 Figure I.8. Map of Medicaid Certified Nursing Facilities Overlaid with LTC Enrollees Living in the Community Stratified by Race/Ethnicity Source: The Florida Geographic Data Library Data processing and cartography: FSU College of Medicine, Department of Behavioral Sciences and Social Medicine, May Final Report (Fiscal Year ) June 30, 2017

62 Conclusions Adequate access to care requires a network of willing LTC providers contracted by the plans located in reasonable proximity to enrollees residences. On a statewide basis, the network at face value appears to be robust. However, there are geographical differences in facility availability by county. These differences may be even greater between certain plans, depending on the robustness of individual plans networks. This possibility requires additional inquiry in future reports before reaching a universal (i.e., statewide) conclusion about network adequacy. The evaluation team s high-level analysis of service utilization in the Pre-LTC and LTC period identified relatively few differences in service receipt for HCBS enrollees between the two evaluation periods. The greatest increase in service receipt occurred for homemaker, home health, and personal care services, whereas the greatest decrease in service receipt occurred for assistive care services and medical equipment & supplies. Proper assessment of the intensity of service utilization requires reliable completion of the units of service components and very few duplicate records in the encounter data. The analysts approach in this report was to treat all services as equal and to measure when a given service category was represented by an encounter record at least once per month for a given recipient. The Participant Directed Option allows some flexibility in the service delivery options available to home-based enrollees eligible for adult companion care, attendant care, homemaker services, intermittent and skilled nursing, or personal care services. All plans are required to make these five services available to their enrollees via the PDO. Participation in the PDO during the LTC period was modest, though PDO enrollment slowly grew throughout SFY1415, with about a 2 percentage point increase in enrollment from the start to the end of the evaluation year for individuals who receive services in home-based settings. There are differences in levels of PDO enrollment across the seven plans. In the aggregate, the LTC plans are realizing the Agency s stated policy goal of drawing down the Florida Medicaid LTC enrollee nursing facility population. The nursing facility residency rate slowly grew over the Pre- LTC period, with about a 1 percentage point increase in the nursing facility population from the start to the end of SFY1213, but this rate steadily decreased over the LTC period, with about a 3 percentage point reduction in the nursing facility population from the start to the end of SFY1415. There is some consistency, in terms of demographic characteristics, that suggests the LTC plans are shifting younger enrollees out of nursing facilities and into HCBS sites of care. It is unlikely that age itself drives these shifts; rather, it is likely that age reflects differences in health status that determine who may successfully transition into the community. It is also likely that younger, disabled enrollees may be more motivated to transition and may have greater access to an available caregiver(s). Additionally, NF personnel may be more motivated to relocate younger residents, who are often outliers in their facilities. The percentage of enrollees experiencing multiple location shifts increased between the Pre-LTC and LTC period. This may be a positive finding if the shifts are appropriate. The evaluation team would like to understand more 62 Final Report (Fiscal Year ) June 30, 2017

63 about how the plans identify enrollees in their original nursing facility residencies for eventual transition into home and community-based settings. The evaluation team identified substantial differences in location of care by race/ethnicity. Hispanic enrollees were far more likely than black enrollees and white enrollees to reside in home and community-based settings. There are further differences by race/ethnicity within the home and community-based population; specifically, differences in rates of assisted living facility and assistive care service receipt. White HCBS enrollees were more than 1.5 times as likely as Hispanic HCBS enrollees and almost 2.5 times as likely as black HCBS enrollees to receive care in assisted living facilities. Moreover, white enrollees were more likely to receive assistive care services than both Hispanic and black enrollees. Some of this imbalance may be explained by differences in the geographic availability of assisted living facilities and assistive care services. Recommendations What follows is a list of recommendations that, if implemented, would improve the evaluation of access to care in the program. Recommendations for Data Quality Improvement 1) The evaluation team recommends that the Agency continue ongoing efforts to work with the LTC plans to improve their encounter record reporting, especially to ensure accurate units of service (UOS) reporting. Additionally, the evaluation team will work with the Agency to understand when 0 UOS may be entered as a legitimate value. A more nuanced evaluation of service utilization should reflect the fact that some services are measured in 15 minute increments and others on a per diem basis. This action would also create added value for the analysis of the PDO by permitting an assessment of PDO service use intensity. 2) Likewise, the evaluation team recommends improvements to the reporting and/or more regular updating of where a plan enrollee is located in a given month. The evaluation team will work with the Agency to identify enrollee residency with greater certainty. 3) Work with the Agency to understand why there are so few case management encounters reported in the utilization data. Previous communication with the Agency (October 11, 2016) indicated that plans should submit monthly case management records, whether that service is provided internally by the LTC plan or externally by a subcontracted agency or provider. 4) If the Agency finds it is necessary to continue altering the service category classification scheme, the evaluation team recommends that the Agency does not use any pre-existent service categories in new schemes (e.g., do not use L2.3 for nursing facility service in the new scheme when it was used previously to identify assisted living services). Develop a standardized format for the data files sent to the evaluation team that maintains constant variable naming, width, and storage type conventions across data deliveries. Recommendations for Future Analysis 1) The evaluation team believes access to care begins with the services indicated in enrollees plans of care. Accordingly, the FSU team would like to incorporate this data source in future analyses. 63 Final Report (Fiscal Year ) June 30, 2017

64 2) Conduct more analyses at the plan and region level. These additional analyses would contribute added value by allowing the Agency and evaluation team to isolate effects of plan and regional dynamics. Examples include evaluating differences in services provided across plans after taking into account enrollee case mix, assessing potential barriers to access to care in rural areas, and/or identifying distinct strategies among the plans for shifting enrollees into home and community-based settings. 3) Examine in greater detail whether or not there are systematic differences in network robustness by county and across plans that might impact plans capacities to serve their enrollees. 4) Create a methodology to better understand who transitions out of nursing facilities and into home and community-based settings, specifically, one that provides insight into how plans identify enrollees for possible transition, with particular attention to whether enrollees are systematically or randomly identified for possible transition (e.g., do less frail individuals or individuals with less severe medical conditions transition back into the community). Provision of patient assessment data conducted by plans that is not included in the MDS or CIRTS data files would be useful for this analysis. Additional analysis should also aim to uncover any systematic differences between enrollees who experienced more than one location shift in contrast to those who transitioned only once during the study period. The evaluation team recognizes that enrollees are the ultimate arbiter of the choice to transition. 5) Compare pathways through and services received under the program for the less elderly, disabled population of LTC enrollees versus the elderly, frail population of enrollees. 6) Identify potential causes of low levels of nursing facility care, beyond their region of residency, among AMG enrollees. 7) Evaluate outcomes for LTC enrollees who remain in nursing facilities versus outcomes for those who transitioned from nursing facility residencies into home and community-based settings. Additional recommendations 1) Work with the Agency to understand the extent and potential effects of uncleared encounters as well as claims and encounter runout. 16 2) Work with the Agency to ensure appropriate interpretation of the PNV files. The evaluation team remains uncertain as to why there are individual providers in the PNV files reported as offering nursing facility or assisted living services and will work with the Agency to better understand these records. 16 Uncleared encounters have not met the Agency s standards for payment and are not included in the data pull for the evaluation, even if a record represents legitimate service receipt. Encounter/claims runout occurs when there are records of service provision that took place in SFY1415 that were not submitted to the Agency by the time the Agency pulls the data to send to the evaluation team. 64 Final Report (Fiscal Year ) June 30, 2017

65 II: Quality of Care Purpose The Independent Assessment of the Florida Statewide Medicaid Managed Care Long-term Care Program assesses the effect of the program implemented on August 1, 2013 on quality of care for enrollees. It uses both quantitative and qualitative methods. Key research questions (RQs) identified differences in quality of care between those eligible Medicaid recipients who were in NFs or participated in the waiver programs before the LTC program (Pre-LTC) was implemented and those LTC program enrollees who currently participate in the LTC program. Appendix 1 of this report provides further information on specific methodologies employed and data sources used. The seven RQs listed below were developed to guide the evaluation of quality of care. 1. Has the quality of services that enrollees receive changed compared to the previous LTC programs and over time? 2. What have managed care plans done to improve quality of care? 3. How has timeliness of service delivery changed compared to the previous LTC programs and over time? 4. Are enrollees rights being addressed in accordance with the waiver and contract? 5. How effective are the LTC plans complaints, grievances and appeals processes? 6. What are the rates of complaints, grievances and appeals in the LTC program and have these changed over time? 7. Is there sufficient monitoring and oversight of the program by the state? What types of monitoring are done, and how are they conducted? Quantitative Findings RQ1: Has the quality of services that enrollees receive changed compared to the previous LTC program(s) and over time? This is the essential question of this evaluation. Instrumental to answering this question is defining quality in LTC settings; however, there is no universal definition for quality of care to date. In order to address issues of quality, the evaluation team sought guidance from Agency-defined performance measures, research-based empirical literature, gray literature, such as white papers authored by organizations, reports from the External Quality Review Organization (EQRO), and expert opinion. These sources have been instrumental in developing a list of Quality measures (QM) appropriate for each of the three locations of care delivery (i.e., nursing facilities (NF), other residential settings (ORS 17 ), and home). The Agency has adopted a conceptual framework that includes 17 For the purposes of these analyses, ORS includes care in assisted living facilities (ALFs) and adult family care homes (AFCHs). ORS provide many of the same services as those required by enrollees living in their own homes. They are home-like environments that are shared with others and are considered social rather than medical settings. 65 Final Report (Fiscal Year ) June 30, 2017

66 four components of healthcare quality: a) safety; b) effectiveness; c) patient-centeredness; and d) timeliness. These components are reflected in the QMs examined to answer RQ1. For the purposes of answering RQ1, the evaluation team presents, whenever possible, each QM Pre-LTC and LTC. Whenever possible, the evaluation team included QMs that are applicable and measureable across all three settings of care. In the absence of data related to outcomes, the evaluation team looked at some measures of process and, in some instances, relied upon reports of administrative personnel regarding structural changes. Quality Measures The evaluation team, along with the Agency and DOEA administrators, continue to define healthcare quality for enrollees in the LTC program. This definition reflects a dynamic process as successes and deficiencies in the current program come to light within the context of verifiable observations and evolving knowledge. Embedded within this definition are components of person-centered care, a value espoused by the Centers for Medicare and Medicaid Services (CMS), Florida Agency for Health Care Administration (AHCA), and other stakeholders in LTC quality. Quality of Life Given that quality of life is an optimal goal of all persons, it is one indicator that is appropriate to all sites of care. These data are available for those enrollees living in ORS and for those receiving HCBS in their homes. Unfortunately, there is no specific measure of quality of life for NF residents currently available. Despite the current absence of data for NF enrollees, the evaluation team chooses to list this QM here, as it is, ultimately, the most essential QM for all older adults. Health plays an important role in how individuals experience their quality of life. How an individual perceives his or her health, i.e., as poor, fair, good, very good or excellent, is a recognized predictor of functional decline and mortality. 18 Of specific relevance to this evaluation, is whether the move to the LTC program leads to a change in perceived health status irrespective of site of care (e.g., NF versus ORS or ORS versus home). Does perception of health increase (greater reports of good, very good, excellent) or decrease (greater reports of fair or poor) from Pre-LTC to the LTC program? Table II.1 summarizes the responses of enrollees in other residential settings to questions relating to their quality of life and self-reported health and whether either of these has changed since they were enrolled in LTC. When asked how satisfied they were with their overall quality of life, those ORS residents in the LTC group were significantly more likely to indicate that they are satisfied or very satisfied compared to ORS residents Pre-LTC (70.47% v 44.34%). Self-reported overall health among ORS residents improved with the percent of those rating their health as fair or poor dropping (67.66% v 53.12%) and those reporting their health as excellent/very good and 18 Stewart, A.L., & Ware, J.E. (1992). Measuring functioning and well-being: The Medical Outcomes Study approach. Durham NC: Duke University Press. 66 Final Report (Fiscal Year ) June 30, 2017

67 good increasing (32.33% v 46.88%) from Pre-LTC to LTC. The percent reporting their health as very good/excellent nearly doubled in the LTC period (3.82% v 7.5%). The percent of ORS residents who rated their health as better and much better compared to one year ago increased between the Pre-LTC and the LTC periods (9.71% v 13.16%). Table II.1. Quality of Life and Self-Reported Health for Enrollees in Other Residential Settings (ORS) Pre- LTC and LTC How satisfied are you with your overall quality of life? Pre-LTC (N = 6,824) LTC (N = 8,707) % of % of N N Respondents Respondents Very satisfied Satisfied Neither satisfied/dissatisfied Dissatisfied/Very dissatisfied Thinking about last year, how do you feel about the way things are now? (Assess 2 only) Much better Better About the same Worse Much worse How would you rate your overall health at the present time? Excellent/Very good Good Fair Poor Compared to a year ago, how would you rate your health? 186 1,520 1, ,268 2, ,296 1, , ,429 3, Much better Better About the same Worse/Much worse , ,720 1, Source: CIRTS (Assess 1, 701B, v. 2008; Assess 2, 701B, v. 2013) Note: The pre-ltc total N is a function of both assessment 1 and assessment 2, and some variables/questions were only present in certain assessments. There is also missing data due to questions not being completed in each assessment. The italicized portion behind each question indicates if the question was only available in a certain assessment. N for Pre-LTC or LTC includes the number of conditioned respondents in each row (who answered yes ), non-conditioned respondents (who answered no ), and missing/nonrespondents. Table Il.2 summarizes the response to the same questions for enrollees residing at home and receiving home based services. When asked how satisfied they are with their overall quality of life, those home residents (i.e., enrollees living at home and receiving services) in LTC were more likely to indicate that they are satisfied or very satisfied compared to home residents Pre-LTC (67.39% v 36.90%). 67 Final Report (Fiscal Year ) June 30, 2017

68 Self-reported overall health improved significantly among home residents from Pre-LTC to LTC with those rating their health as poor dropping from 26.45% Pre-LTC to 16.16% LTC. The percent rating their overall health as good doubled from 13.77% Pre-LTC to 27.58% LTC. Table II.2. Quality of Life and Self-Reported Health for Enrollees Living at Home with HCBS Pre-LTC and LTC How satisfied are you with your overall quality of life? Pre-LTC (N = 14,651) LTC (N = 16,339) % of % of N N Respondents Respondents Very satisfied Satisfied Neither satisfied/dissatisfied Dissatisfied/Very dissatisfied Thinking about last year, how do you feel about the way things are now? (Assess 2 only) 152 1,534 2, ,319 2,667 1, Much better Better About the same Worse Much worse How would you rate overall health at the present time? ,693 8,129 1, Excellent/Very good Good Fair Poor Compared to a year ago, how would you rate your health? ,036 1, ,507 8,448 2, Much better Better About the same Worse/Much worse ,731 2, ,866 9,101 4, Source: CIRTS (Assess 1, 701B, v. 2008; Assess 2, 701B, v. 2013) Note: The pre-ltc total N is a function of both assessment 1 and assessment 2, and some variables/questions were only present in certain assessments. There is also missing data due to questions not being completed in each assessment. The italicized portion behind each question indicates if the question was only available in a certain assessment. N for Pre-LTC or LTC includes the number of conditioned respondents in each row (who answered yes ), non-conditioned respondents (who answered no ), and missing/non-respondents. The evaluation team also examined a quality of life question that was part of the Enrollee Satisfaction Survey used by each LTC plan. The Enrollee Satisfaction Survey was conducted by an independent survey vendor approved by the Agency. Figure II.1 depicts the survey results for enrollees perceptions of improvement to their quality of life since enrolling in their LTC plan. The majority of enrollees (77.4%) report that their quality of life has improved since 68 Final Report (Fiscal Year ) June 30, 2017

69 enrollment, with 30.6% reporting their quality of life has improved a lot. A much smaller minority (22.6%) report their quality of life has not improved. Figure II.1. Community Dwelling 1 LTC Enrollees Change in Quality of Life Since you enrolled in [Plan Name's] long-term care plan, has your quality of life improved? Yes, it has improved a lot 30.6% Yes, it has improved a little 46.8% No, it has not improved 22.6% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Percentage of Respondents Source: Satisfaction Survey 1 Includes home and ORS residents Client Involvement in Assessment Process The most basic principle in medical ethics is autonomy, the right to choose your own treatment. An essential element of person-centered care is involvement of the client in the assessment process. Although there is some concern that these data will be affected by the cognitive status of the enrollee, it is important to provide opportunities to all enrollees to participate as fully as possible in the assessment process. The evaluation team examined enrollee participation in the assessment process for enrollees residing in NFs. As shown in Table II.3, enrollee participation in the assessment process in NFs (83.4% v 83.3%) is essentially unchanged. The assessment process variable was not available in the Pre-LTC period for those enrollees living in ORS or at home; therefore this analysis reflects the LTC period alone. Greater than half (55.2%) of LTC enrollees in ORS participated in the assessment process; someone other than the enrollee participated on behalf of the enrollee for the remaining 44.8%. For enrollees receiving HBCS in their homes, 45.3% participated in the assessment process, while someone other than the enrollee participated on behalf of 54.7% of enrollees. 69 Final Report (Fiscal Year ) June 30, 2017

70 Table II.3. NF Enrollee Participation in Assessment Process Pre-LTC and LTC Participant Pre-LTC 1 LTC 1 Dif N % 2 N % 2 % NF Enrollee 43, , Family or significant other of enrollee 21, , Guardian or legally authorized representative of enrollee 6, , ORS Client , HOME Someone other than the client , Client , Someone other than the client , Source: Minimum Data Set (MDS) Taken from most recent (last) enrollee assessment per period. 2 Values may add to more than 100% due to multiple participants in one assessment. Unmet Need One essential outcome measure of quality is the level of unmet need. Unmet need is defined as an identified need for which appropriate services are not provided. The greater the unmet need, the poorer the quality of care. For ORS and home, unmet needs were determined from the need for assistance in ADLs (i.e., basic tasks of everyday life, such as eating, bathing, dressing, toileting, and transferring) and the availability of assistance to meet those needs. The higher the score, the greater the unmet need. A detailed description of the algorithm is available upon request through the Agency. As reported in Table II.4, the proportion of ORS residents needing and having ADL assistance most of the time declined from Pre-LTC to LTC (3.62% to 2.01%). The percent of those rarely having assistance declined negligibly from 0.28% to 0.26%. However, the percent of those never having assistance rose from 0.49% to 0.82%. For those residents living at home, the proportion needing and having ADL assistance most of the time declined from Pre-LTC to LTC by 9.82% (16.77% v 6.95%). The percent of those rarely having assistance declined by 2.71% (3.81% to 1.10%), and those never having assistance declined by.55% (1.47% v 0.92%). Thus, when looking at the three categories indicating unmet need, overall unmet need has increased. Conversely, the percentage of enrollees rarely or never having assistance has declined. Also, those living at home show an increase in the percentage needing no assistance. 70 Final Report (Fiscal Year ) June 30, 2017

71 Table II.4. Proportion of Enrollees with Unmet Need in Other Residential Settings (ORS) and Home Pre- LTC and LTC Item Pre-LTC LTC Dif ORS Home ORS Home ORS Home Needs no assistance 11.22% 16.64% 26.96% 43.72% 15,74% 27.08% Always has assistance 84.39% 61.31% 71.70% 55.73% % -5.58% Has assistance most of the time 3.62% 16.77% 2.01% 6.95% -1.61% -9.82% Rarely has assistance 0.28% 3.81% 0.26% 1.10% -0.02% -2.71% Never has assistance 0.49% 1.47% 0.82% 0.92% 0.33% -0.55% Source: CIRTS (Assess 1, 701B, v. 2008; Assess 2, 701B, v. 2013) Categories denoting unmet need (i.e., those that denote assistance is not available all of the time and assistance is required by enrollee). This table includes the percentage of the total population of enrollees experiencing unmet need. At a more granular level, the percentage of ORS residents for whom no assistance is needed or who always has assistance was approximately 95% for each ADL area in the Pre-LTC. This contrasts with the variability in the proportion of home enrollees who need no assistance or who always have assistance across ADLs, ranging from a low of 75.41% for Transferring to a high of 85.78% for Eating Pre-LTC. Enrollees who transferred into the program or are new enrollees are extremely frail. That the level of unmet need is small (ORS 2.01%, Home 6.95%), though important, is an acknowledgement of the value of the services provided. Notably, the higher degree of unmet need in enrollees living at home is of concern as more enrollees transition from NFs to the community. Much of home care is provided by informal care providers such as family members, and the loss of care provision is an issue that requires attention. As the expectation of decline is part of the aging process, continued assessment of the need for ADL assistance and the availability of that assistance is an important part of measuring quality of care. In NFs, unmet need is less transparent. However, the MDS contains data regarding the trigger of a resident assessment protocol (RAP) and presence of a Plan of Care (POC) to address that need. The lack of an appropriate POC indicates unmet need (i.e., poor quality). Figure II.2 depicts unmet need in NFs pre-ltc and LTC SFY Overall, unmet need for enrollees in NFs has decreased in the LTC period. 71 Final Report (Fiscal Year ) June 30, 2017

72 Figure II.2. Unmet Need in Nursing Facilities Proportion of cases with problems not addressed in Care Plan 0% 10% 20% 30% 40% 50% 60% 70% 01. Delirium. 02. Cognitive Loss/Dementia. 03. Visual Function. 04. Communication. 05. ADL Functional/Rehabilitation Potential. 06. Urinary Incontinence and Indwelling Catheter. 07. Psychosocial Well-Being. 08. Mood State. 09. Behavioral Symptoms. 10. Activities. 11. Falls. 12. Nutritional Status. 13. Feeding Tube. 14. Dehydration/Fluid Maintenance. 15. Dental Care. 16. Pressure Ulcer. 17. Psychotropic Drug Use. 18. Physical Restraints. 19. Pain. 20. Return to Community Referral. Average Source: Minimum Data Set (MDS) Depression Depression is a recognized problem for older adults in general and especially for frail older adults. In the Medicaid population, many enrollees have a lack of resources that may exacerbate the risk for depression. The evaluation team examined data from the nine-item Patient Health Questionnaire (PHQ-9) that is embedded in the MDS and the 701B Comprehensive Assessment to determine the presence of depressive symptoms. The nine items are summed to create a severity score, with scores of ten or higher indicating moderate to severe depression. The results are displayed in Table II.5. In NFs the prevalence of self-reported depression (i.e., a score greater than 10 as measured by the PHQ-9) was considerably less in the LTC (1.36%) versus Pre-LTC (6.60%) period. The LTC percentages reported by staff (3.04%) were similar to the Pre-LTC percentages (3.40%). Among ORS residents in LTC, scores greater than 10 rose from 1.63% Pre-LTC to 2.49% LTC. Among residents living at home the percent of those with depression doubled from 2.01% Pre-LTC to 4.64%. 72 Final Report (Fiscal Year ) June 30, 2017

73 Table II.5. Depression as Measured by PHQ-9 Scores 1 > 10 in All Sites of Care Pre-LTC LTC Site of Care Respondents N % Respondents N % NF (self) 2 13, , NF (staff) 2 16, , Other residential settings 3 2, , Home 3 5, , PHQ-9 Mental Health Index (Score: 0-27) 2 From MDS From 701B Comprehensive Assessment, version 2013 These scores, unlike earlier findings, show increases in levels of enrollee depression when compared with the Pre-LTC period for ORS and home sites of care. These findings are more consistent with the older adult population as a whole. The most recent data (2008) from the Health and Retirement Study as published in Older Americans 2012: Key Indicators of Well-being, noted that 11% of community-dwelling older men and 16% of community-dwelling older women had clinically relevant depressive symptoms. While these data are not based on the same measure, these data are more in line than those from the previous evaluation year. Also, the evaluation team conducted additional analyses regarding this QM. Table II.6 describes the distribution of depression scores for all three sites of care in the Pre-LTC period. Table II.7 describes those scores following implementation of the LTC. Table II.6. Depression as Measured by PHQ-9 Scores 1 Distribution of Scores (Pre-LTC) Site of Care N 0 (%) 1-4 (%) 5-9 (%) (%) 20+ (%) NF (self) 2 81, < NF (staff) 2 16, < Other residential settings 3 2, Home 3 5, PHQ-9 Mental Health Index (Score: 0-27) 2 From MDS From 701B Comprehensive Assessment, version n = 34 (cases) 5 n = 14 (cases) 73 Final Report (Fiscal Year ) June 30, 2017

74 Table II.7. Depression as Measured by PHQ-9 Scores 1 Distribution of Scores (LTC) Site of Care N 0 (%) 1-4 (%) 5-9 (%) (%) 20+ (%) NF (self) 2 37, NF (staff) 2 10, Other residential settings 3 8, Home 3 16, PHQ-9 Mental Health Index (Score: 0-27) 2 From MDS From 701B Comprehensive Assessment, version 2013 Of note, in all settings of care, both Pre-LTC and LTC, more than half of enrollees have PHQ-9 scores of zero. A zero score indicates that enrollees did not answer positively to any items related to depression. There also is a significant decrease in the percentage of enrollees receiving HCBS at home with scores of 0 in the LTC period (84.68% Pre-LTC v 60% LTC). While it is possible that the decrease in the percentage of home enrollees with scores of 0 represents an increase in depressive symptoms, it is also possible that screening for depression has improved. Moreover, it is possible that more enrollees are being treated for depression and exhibiting fewer symptoms. Therefore, monitoring rates of depression over a longer period of time is needed. The evaluation team will continue to examine this QM in future studies. Nursing Facility Quality Measures CMS has developed a series of QMs for NF use. According to CMS, the purpose of these QMs is to assist families in choosing a NF or in evaluating a NF where a family member currently lives. Furthermore, CMS states that NFs can use QMs to facilitate quality improvement, including discussions with staff regarding current quality issues. The QMs are specific to residents based on their length of stay. Short-stay residents are those who are in an episode that is less than or equal to 100 days at the end of the target period (i.e., the span of time that defines the QM reporting period). An episode is defined as a period of time spanning one or more stays, beginning with an admission and ending with either a discharge or the end of the target period. The current CMS-defined NF QMs used in this evaluation are: Short Stay Quality Measures Percent of Residents Who Self-Report Moderate to Severe Pain Percent of Residents with Pressure Ulcers that are New or Worsened Percent of Residents Who Were Assessed and Appropriately Given the Seasonal Influenza Vaccine Percent of Residents Assessed and Appropriately Given the Pneumococcal Vaccine Percent of Short-Stay Residents Who Newly Received an Antipsychotic Medication 74 Final Report (Fiscal Year ) June 30, 2017

75 Long Stay Quality Measures Percent of Residents Experiencing One or More Falls with Major Injury Percent of Residents Who Self-Report Moderate to Severe Pain Percent of High-Risk Residents with Pressure Ulcers Percent of Residents Assessed and Appropriately Given the Seasonal Influenza Vaccine Percent of Residents Assessed and Appropriately Given the Pneumococcal Vaccine Percent of Residents with a Urinary Tract Infection Percent of Low-Risk Residents Who Lose Control of Their Bowels or Bladder Percent of Residents Who Have/Had a Catheter Inserted and Left in Their Bladder Percent of Residents Who Were Physically Restrained Percent of Residents Whose Need for Help with Activities of Daily Living Has Increased Percent of Residents Who Lose Too Much Weight Percent of Residents Who Have Depressive Symptoms Percent of Long-Stay Residents Who Received an Antipsychotic Medication The evaluation team reviewed the 18 CMS QMs to determine the appropriateness for inclusion in this report. One NF QM, depressive symptoms, has been discussed along with ORS and home. In addition, the evaluation team had access to data contained in the CIRTS database that allowed the evaluation team to examine these QMs across settings of care using the same assessment instrument. Some CMS QMs were less helpful. For example, NF residents usually lose weight as they approach the end of life, making weight loss less meaningful in this setting. Similarly, the evaluation team chose to exclude ADL declines as a measure of quality. In addition, some of the QMs are process measures (e.g., percent of residents given influenza vaccination). While process measures are important, they are less important to an individual client than outcomes measures, such as development of pressure ulcers. In addition, most facilities are achieving high rates for such measures, so they are becoming less important to distinguishing quality providers. Using the approaches gleaned from empirical research, as well as expert opinion, the evaluation team chose to include the following CMS QMs in the current report evaluating quality in NFs for long-stay residents: 1) percent of residents experiencing one or more falls; 2) percent of residents who self-report moderate to severe pain; and 3) percent of residents with unhealed pressure ulcers. The volume of missing data regarding date of admission prevented the evaluation team from differentiating between long- and short-stay. Therefore, all three QMs were included for all NF residents. 75 Final Report (Fiscal Year ) June 30, 2017

76 Table II.8. Quality Measures Selected from CMS QMs for NFs Pre-LTC LTC Dif QM Respondents N % Respondents N % Unhealed pressure 56,935 4, ,000 4, ulcers 1 Experienced one or more 62,788 9, ,779 7, falls 2 Experienced moderate to 51,760 14, ,370 10, severe pain 3 Source: MDS Enrollees with unhealed pressure ulcers? 2 Has enrollee had a fall since admission or last assessment? 3 Self-report from enrollee: Have you had pain or hurting within the last 5 days? Pressure Ulcers As noted above in Table II.8, there were 4,384 (7.7%) NF enrollees with unhealed pressure ulcers in the Pre-LTC period. In the LTC period the number was 4,565 (8.3%) for a modest increase of 0.6%. In addition to unhealed pressure ulcers, the evaluation team examined the number of pressure ulcers that had healed since the previous assessment or upon admission (see Table II.9). There are modest to moderate changes in the percentage of pressure ulcers healed. Table II.9. Healed Pressure Ulcers 19 by Period (Pre-LTC, LTC) Pre-LTC Difference LTC (pre-post)(%) Pressure Sores N (%) 1 3,671 (6.33) 3,223 (6.27) 0.06 Stage 2 2 n(%) 860 (23.47) 695 (21.59) 1.88 Stage 3 2 n(%) 421 (11.48) 333 (10.34) 1.14 Stage 4 2 n(%) 262 (7.15) 200 (6.21) 0.94 Source: MDS Number of enrollees with pressure sores on previous assessment or upon admission 2 Number and percentage healed since previous assessment or admission Falls There were 9,167 NF enrollees (14.6%) who experienced falls in the Pre-LTC period. In the LTC period, 7,450 or 13.6% had falls recorded. These data suggest a slight difference of -1.0 percent (see Table II.8). 17 Pressure ulcers are categorized in stages to indicate degree of severity. Stage 1 is the beginning stage of a pressure sore, in Stage 2 the outer layer of skin (epidermis) and part of the underlying layer of skin (dermis) is damaged or lost, in Stage 3 there is a deep wound, and in Stage 4 there is large scale loss of tissue Final Report (Fiscal Year ) June 30, 2017

77 Pain As frequency and severity of pain are pertinent measures, it is helpful to look at these measures in tandem. Table II.10 shows that frequency of pain appears to have improved slightly; however, the reported pain level and verbal rating of pain intensity have remained the same. Table II.10. Nursing Facility (NF) Enrollees with Pain Pre-LTC LTC Dif Item N % M N % M % M Have you had pain or hurting at any time in the past 5 days? 1 Pain rating over the last 5 days, 0-10? 2 14, , , , Verbal rating of pain intensity? 3 5, , Source: MDS Dichotomous (Yes/No) variable 2 Based on numeric scale (continuous variable) of 0-10 with 0 = no pain and 10 = worst pain you can imagine. 3 Based on verbal description of pain (1 = mild; 2 = moderate; 3 = severe; 4 = very severe, horrible). Medications Use of medications, specifically antipsychotic medications, have been identified as measures of quality by CMS. In NFs, off-label use of antipsychotic medications (i.e., use of these medications for diagnoses for which they were not developed) can be considered a chemical restraint. For these reasons, it is important to monitor the use of anti-psychotic medications. Figure II.3 shows the use of antipsychotic medications as well as anti-anxiety, antidepressant, and hypnotic medications in SFY and SFY There has been a slight decline in the percent of nursing facility residents receiving antipsychotic medications. This is consistent with national trends. There has also been a decline in the use of hypnotic medications. However, there has been an increase in the percent of nursing facility residents receiving antidepressant medications, now at 58%. These data are particularly interesting when coupled with earlier findings the absence of depressive symptoms in the NF population as measured by the PHQ-9 (12-16). In future reports, the evaluation team will need to include data regarding diagnoses of depression in order to compare findings from different years and sites of care. 77 Final Report (Fiscal Year ) June 30, 2017

78 Figure II.3. Proportion of Nursing Facility Residents Receiving Specific Medications Proportion of nursing home residents receiving medications Antipsychotic Antianxiety Antidepressant Hypnotic 0% 10% 20% 30% 40% 50% 60% 70% Source: Minimum Data Set (MDS) 3.0 Taken together, these data support very slight gains in quality (e.g., modest reductions in the prevalence of pain, falls, and pressure ulcers), but these gains are not sufficient to make a firm determination of increased quality overall. Furthermore, it should be noted that plans are not responsible for medical care in NFs and do not have access to MDS data. However, it is the expert opinion of the evaluation team that these enrollee-level outcomes are the most valid measures of quality of care. Plans do indirectly influence quality and, therefore, these outcomes are affected. Home and Community-Based Service Quality Measures Defining quality of home and community-based services has been an issue and challenge of long duration, dating back, at least, to the first 1915 (c) waivers in the early 1980s. Outcome measures associated with acute care do not readily lend themselves to use with home and community-based sites, as LTC patient conditions often deteriorate over time even under the best, high quality circumstances. The United States Government Accountability Office (GAO) in a 1994 report suggested a series of outcome indicators of quality of care and grouped them into four categories: functioning, safety, health, and client satisfaction. For example, a change (worsening) of ADLs would reflect functioning. Falls, a frequent measure of quality, relates to safety as decubitus ulcers (pressure sores) relate to health. This conceptual framework of quality has been adopted by the Agency for Healthcare Research and Quality (AHRQ) as well. 78 Final Report (Fiscal Year ) June 30, 2017

79 The state of Florida and the managed care plans participating in the LTC program seek to oversee and ensure quality in two ways. The state provides licensing, certifications, and sets regulatory requirements. Both the state and managed care plans monitor care provision. The development of QMs for home and other residential community based sites is not as far along as it is for nursing facilities. Further, there is a movement nationally to develop QMs that are applicable across sites of care. The evaluation team has chosen from among current national HCBS data sources those areas (social help, safety, health, and nutrition) and select associated metrics appropriate for enrollees living at home. Enrollees living in other residential settings (ORS) are handled separately later in this report. Some, but not all, of the metrics listed below, specifically in the area of health, are reported in Table II.11. Unmet need and depression are explored separately for SFY (Unmet Need Table ll.4 and Figure ll.2, pages 72-73; Depression Tables ll.5-ll.7, pages 74-75). Social Help Percent with social help, if needed Safety Percent for whom medication review by doctor or pharmacist is recommended Percent for whom medications are not being properly managed Percent taking three or more medications Number of times fallen in last six months Health Percent who always/mostly get medical care when needed Percent very satisfied/satisfied with overall quality of life Percent indicating that their health is excellent/very good Percent indicating they are not depressed Change in unmet need related to ADL Nutrition Percent indicating that they eat at least two meals a day Percent indicating that they eat alone most of the time Social Activity Percent that talk to friends, relatives, or others (by phone, computer or other means) daily, weekly, several times a month, quarterly, never Percent that spend time with someone who does not live with them daily, weekly, several times a month, quarterly, never Percent that participate in activities outside the home that interest them daily, weekly, several times a month, quarterly, never 79 Final Report (Fiscal Year ) June 30, 2017

80 The QMs selected for this evaluation were drawn to most closely reflect the thinking represented above as well as availability within the DOEA CIRTS database. To assess Pre-LTC indicators, the evaluation team chose a 12-month look-back from the client s enrollment date. The look-back created a number of challenges. One of these challenges was the revision of the 701B Comprehensive Assessment Tool that took effect on July 16, The look-back period involved enrollees who were assessed using the earlier version of the 701B Comprehensive Assessment and/or the current version. The current version 701B Comprehensive Assessment includes some valuable variables for the assessment of quality (e.g., the number of falls, client medication management). The evaluation team elected to use the current version when it represented the last assessment done of a client in the Pre-LTC period, recognizing the implication for the study population (n-size). Each of the tables below notes when the variable is from the current version (2013) of the 701B Comprehensive Assessment. The impact on population size for analysis is as follows: In the population of ORS enrollees Pre-LTC, 65% of residents completed the earlier version (2008) of the 701B Comprehensive Assessment and 35% the current (2013) version. In the home resident population, 61% completed the earlier (2008) version and 39% the current (2013) version. All LTC program assessments were completed using the current 701B Comprehensive Assessment. Another challenge relates to those residents with multiple assessments. The 701B Comprehensive Assessment is used by the case manager for initial and annual assessments or upon significant change. The MDS 3.0 is administered by the NF upon admission and quarterly thereafter, except in instances of a significant change. The evaluation team decided to use the last assessment done of a resident in the Pre-LTC period as a project baseline and the last assessment in the LTC period for use in measuring change. The evaluation team recognizes that all the variables associated with the annual assessment are not likely to be collected during quarterly assessments, but believes that the variables of greatest interest, detailed in Table II.11, are collected across assessments. Table II.11. Quality Measures for Enrollees Residing at Home Quality Measures Social Help: If needed, is there someone who can help you? (Assess 2 only, 701B, v. 2013) N Pre-LTC (N = 14,651) % of Total N N LTC (N = 16,339) % of Total N Dif. % Yes No Safety: Should client have new medication review by doctor/pharmacist? (assessor) (Assess 2 only, 701B, v.2013) ,632 6, Yes No 75 1, , Final Report (Fiscal Year ) June 30, 2017

81 Quality Measures N Pre-LTC (N = 14,651) % of Total N N LTC (N = 16,339) % of Total N Dif. N/A % Safety: Are client s medications managed properly? (assessor) (Assess 2 only, 701B, v.2013) Safety: Current condition of dehydration 3 Safety: Number of times client has fallen in last 6 months (Assess 2 only, 701B, v.2013) Safety: Do you take 3 or more prescribed/over-counter medications? (Assess 2 only, 701B, v.2013) Yes No N/A 1, Mean , SD Mean SD Yes No Health: When you need medical care, how often do you get it? (Assess 2 only, 701B, v.2013) 15, Always Most of the time Rarely Only in emergency Never Health: How satisfied are you with your overall quality of life? Very satisfied Satisfied Neither satisfied/dissatisfied Dissatisfied/Very dissatisfied Health: How would you rate your overall health at the present time? 1, ,534 2, , ,319 2,667 1, Excellent/Very good Good Fair Poor Nutrition: Do you usually eat at least two meals a day? (Assess 2 only, 701B, v.2013) Yes Nutrition: Do you eat alone most of the time? (Assess 2 only, 701B, v.2013) No ,036 1,381 1, ,507 8,448 2,641 15, Yes , Final Report (Fiscal Year ) June 30, 2017

82 Quality Measures N Pre-LTC (N = 14,651) % of Total N N LTC (N = 16,339) % of Total N Dif. No 1, , % Social Activities: About how often do you talk to friends, relatives, or others (by phone, computer or other means)? (Assess 2 only, 701B, v.2013) Once a day 2-6 times a week Once a week/several times a month Every few months/a few times/year Never Social Activities: How often do you spend time with someone who does not live with you? (Assess 2 only, 701B, v.2013) Once a day 2-6 times a week Once a week/several times a month Every few months/a few times/year Never Social Activities: How often do you participate in activities outside the home that interest you? (Assess 2 only, 701B, v.2013) Once a day 2-6 times a week Once a week/several times a month Every few months/a few times/year Never Assessment Type 8,265 4,104 2, ,138 4,738 5,899 3,823 1, ,201 2,998 4,867 3,759 3, Assessment 1, v.2008 Assessment 2, v ,933 5, , Source: DOEA CIRTS (Assess 1, 701B, v. 2008; Assess 2, 701B, v. 2013) 1 Mean 2 Standard deviation 3 Have you been told by a physician that you have any of the following conditions (dehydration)? Yes answers are reported. Social Help Considerably more LTC program home enrollees indicate that they have someone who can help them if needed, someone other than their primary caregiver (58.95% v %). Safety Having a periodic medication review is a QM employed across acute care and LTC sites. This issue is of particular importance for clients who are taking many (i.e., three or more) medications. Among enrollees living at home 96.01% of LTC enrollees are taking three or more medications. The percent of clients whose medications 82 Final Report (Fiscal Year ) June 30, 2017

83 are being managed properly rose from 93.10% in the Pre-LTC to 96.13% in LTC enrollees per the assessor report. Being dehydrated is a QM employed across acute care and LTC sites placing a client at risk for falls. Among home enrollees the percent of those with a current condition of dehydration more than doubled from 1.26% to 3.18%. Among home enrollees, the mean number of falls dropped from 2.72 Pre-LTC to 2.45 LTC. Hence, with the exception of dehydration, each of these indicators relating to safety has moved in the desired direction for improvement in quality of care. Health The percent of home enrollees rating their health as excellent, very good, or good doubled from 15.42% Pre-LTC to 32.13% in LTC. Receiving medical care when needed may be a factor in ER visits and hospital admissions which impact quality of life. The percent of home clients who reported always getting medical care when needed rose from the Pre-LTC (91.98%) to the LTC period (94.20%). Medical care is not the purview of the LTC plans. This falls to the Managed Medical Assistance plan or to the individual s Medicare coverage; however, coordination of enrollees LTC and MMA plans is a requirement in the LTC plan contract. Nutrition The difference between enrollees residing at home Pre-LTC and LTC when reporting eating at least two meals a day was negligible (97.90% v %). When asked about eating alone most of the time the difference, 4.09%, was greater (33.17% v %). More LTC enrollees living at home report eating alone most of the time suggests a greater degree of isolation and possible lower quality of life. Social Activities 20 Over seventy-five percent (75.97%) of LTC enrollees living at home report talking with friends, relatives or others two or more times per week. Fifty percent (50.85%) talk with friends or relatives on a daily basis. Ten percent (10.41%) talk with friends, relatives or others every few months or never. Sixty-five percent (65.10%) of LTC enrollees living at home spend time with someone who does not live with them 2 or more times a week. Eleven percent (11.50%) spend time with someone less often than every few months or never. Participating in activities outside the home that interest you, an Instrumental Activity of Daily Living, occurs every few months or not at all for 44.52% of LTC enrollees living at home. Twenty-five percent (25.70%) participate in activities of interest two or more times a week. Satisfaction with Care 20 Pre-LTC data are not available. 83 Final Report (Fiscal Year ) June 30, 2017

84 Satisfaction survey data collected from enrollees living in the community addressed a question regarding the enrollees satisfaction with the care they receive from the service providers overseen by plans. Satisfaction survey responses depicted in Figure II.4, include a response to the question, Using any number from 0 to 10, where 0 is the worst long-term care plan and 10 is the best long-term care plan, what number would you use to rate your long-term care plan? A higher score reflects greater satisfaction with care. The results indicate that the vast majority (83.3%) of survey respondents highly rate their long-term care services, with a rating of 8 or greater on an 11 point scale. Figure II.4. Community Dwelling LTC Enrollees Ratings of Long-term Care Program Services Using any number 0 to 10, where 0 is the worst long-term care services and 10 is the best long-term care services, what number would you use to rate your long-term care services? Percentage of Respondents 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 83.3% 10.5% 6.2% Rating Source: Satisfaction Survey (Fall 2014) Range from 0 = worst to 10 = best. Other Residential Settings Quality Measures For the purposes of these analyses, ORS includes care in ALF and AFCH. ORS provide many of the same services as those required by enrollees living in their own homes. They are home-like environments that are shared with others and are considered social rather than medical settings. The evaluation team, using the same sources and logic for defining quality for enrollees living at home, examined the QMs noted in Table II.12 for enrollees living in other residential settings. There are additional QMs employed in ORS that are different from those for enrollees living at home. These will be described in separate tables. 84 Final Report (Fiscal Year ) June 30, 2017

85 Table II.12. Quality Measures Specific to Enrollees Living in Other Residential Settings Quality Measures Social Help: If needed, is there someone who can help you? (Assess 2 only,701b, v.2013) Pre-LTC (N = 6,824) % of N Total N LTC (N = 8,707) % of Total N N Dif. % Yes No ,977 3, Safety: Should client have new medication review by doctor/pharmacist? (assessor) (Assess 2 only, 701B, v.2013) Yes No N/A Safety: Are client s medications managed properly? (assessor) (Assess 2 only, 701B, v.2013) 24 1, , Safety: Do you take 3 or more prescribed/over-counter medications? (Assess 2 only, 701B, v.2013) Yes No N/A 1, , Yes No Safety: Number of times client has fallen in last 6 months (Assess 2 only, 701B, v.2013) 1, , Health: When you need medical care, how often do you get it? (Assess 2 only, 701B, v.2013) Mean SD Mean SD Always Most of the time Rarely Only in emergency Never Health: How satisfied are you with your overall quality of life? 1, , Very satisfied Satisfied Neither satisfied/dissatisfied Dissatisfied/Very dissatisfied Health: How would you rate overall health at the present time? 186 1,520 1, ,296 1, Final Report (Fiscal Year ) June 30, 2017

86 Quality Measures Excellent/Very good Good Fair Poor Nutrition: Do you usually eat at least two meals a day? (Assess 2 only, 701B, v.2013) Pre-LTC (N = 6,824) % of N Total N , , LTC (N = 8,707) % of Total N N , , Dif. % Nutrition: Do you eat alone most of the time? (Assess 2 only, 701B, v.2013) Yes No 1, , Yes No Social Activities: About how often do you talk to friends, relatives, or others (by phone, computer or other means)? (Assess 2 only, 701B, v.2013) 157 1, ,082 7, Once a day 2-6 times a week Once a week Several times a month Every few months/a few times a year Never Social Activities: How often do you spend time with someone who does not live with you? (Assess 2 only, 701B, v.2013) 2,031 2,039 1,376 1, , Once a day 2-6 times a week Once a week Several times a month Every few months/a few times a year Never Social Activities: How often do you participate in activities outside the home that interest you? (Assess 2 only, 701B, v.2013) 1,130 1,880 1,711 1,659 1, Once a day 2-6 times a week Once a week Several times a month Every few months/a few times a year Never Assessment Type ,426 2,370 2, Assessment 1, v.2008 Assessment 2, v ,437 2, , Source: DOEA CIRTS (Assess 1, 701B, v. 2008; Assess 2, 701B, v. 2013) NOTE: Assessment 1 refers to the first 701B Comprehensive Assessment survey instrument from September Assessment 2 refers to the revised 701B Comprehensive Assessment survey from April Mean 2 Standard deviation 86 Final Report (Fiscal Year ) June 30, 2017

87 Social Help More than half of LTC program enrollees in ORS indicate that they have someone who can help them if needed, someone other than their primary caregiver (57.16% in LTC v % in Pre-LTC), a notable increase. Safety Among LTC enrollees in ORS, 98.17% are taking three or more medications. The percent of clients whose medications are being managed properly rose from 97.89% in the Pre-LTC to 98.47% in LTC enrollees per the assessor report. Among ORS enrollees, the mean number of falls dropped from 2.48 Pre-LTC to 2.43 LTC. Health The percent of ORS enrollees rating their health as excellent, very good, or good rose from 32.33% Pre-LTC to 46.88% in LTC. Medical care is not the purview of the LTC plans; however, they are responsible for coordinating medical care as needed. This falls to the Managed Medical Assistance plan or to the individual s Medicare coverage. However, receiving medical care when needed may be a factor in ER visits and hospital admissions which impact quality of life and constitutes an area of interest for future examinations of quality. The percent of ORS clients who reported always getting medical care when needed rose from the Pre-LTC (96.66%) to the LTC period (97.44%). Nutrition As noted in Table II.12, the difference was negligible between those ORS enrollees Pre-LTC and LTC when reporting eating at least two meals a day (98.29% v %), but the percentage of ORS enrollees who ate alone most of the time increased from Pre-LTC to LTC (10.71% v %). The latter, eating alone most of the time, suggests a greater degree of isolation and possible lower quality of life. Social Activities 21 Over sixty percent (62.55%) of LTC enrollees in ORS report talking with friends, relatives or others one or more times per week. Twenty-three percent (23.33%) talk with friends or relatives on a daily basis. Twenty-four percent (24.21%) talk with friends, relatives or others every few months or never. Fifty-four percent (54.22%) of LTC enrollees in ORS spend time with someone who does not live with them 1 or more times a week, 12.98% once a day. More than twenty-six percent (26.72%) spend time with someone every few months or never. Participating in activities outside the ORS that interest you, an Advanced Activity of Daily Living, occurs every few months or not at all for 59.84% of LTC enrollees in ORS. Twenty-four percent (23.79%) participate in activities of interest one or more times a week. Satisfaction On satisfaction surveys, enrollees responded to the question, Since you enrolled in [Plan Name's] long-term care plan, has your overall health improved? The data for the study period are currently under analysis. 21 Pre-LTC data are not available. 87 Final Report (Fiscal Year ) June 30, 2017

88 Caregivers in Crisis While the focus on enrollee outcomes is paramount in measuring quality, the LTC program also seeks to support caregivers. Non-paid, informal caregivers (i.e., family members, close friends) provide most of the care described as instrumental activities of daily living (IADLs; e.g., shopping, money management) and are essential to allowing enrollees to age in place. For enrollees living at home, these informal caregivers may also provide ADL support such as assistance in bathing, dressing and toileting. Indeed, caregiver support is one of the essential elements that allows enrollees to remain in their homes. Often a caregiver crisis will trigger transition to a NF or from home to another residential setting. Therefore, support for these caregivers, as provided in the LTC period, is an important contribution of the change from the former legacy waiver programs. The evaluation team reviewed data available in the 701B Comprehensive Assessment regarding caregiver stress for caregivers of enrollees in home or other residential settings. Specifically, the assessor was asked Is the caregiver in crisis? Furthermore, if the answer was yes, then there was an opportunity for greater specificity regarding the type of crisis. Results from these data are presented in Tables II.13 and II.14. Table II.13. Number and percent of caregivers in crisis based on assessor observation (Other Residential Settings) Is the caregiver in crisis? (Yes) If so, is the caregiver in a financial crisis? (Yes) If so, is the caregiver in an emotional crisis? (Yes) If so, is the caregiver in a physical crisis? (Yes) Pre-LTC Source: DOEA CIRTS (Assess 1, 701B, v. 2008; Assess 2, 701B, v. 2013) LTC N % of total N % of total , , Table II.14. Number and percent of caregivers in crisis based on assessor observation (Home) Pre-LTC LTC N % of total N % of total Is the caregiver in crisis? (Yes) 3, , If so, is the caregiver in a financial crisis? (Yes) If so, is the caregiver in an emotional crisis? (Yes) 1, , , , If so, is the caregiver in a physical crisis? (Yes) Source: DOEA CIRTS (Assess 1, 701B, v. 2008; Assess 2, 701B, v. 2013) 1, , Final Report (Fiscal Year ) June 30, 2017

89 There are increases between the Pre-LTC rates of crisis and those reported in the LTC period for both those living in ORS (Table II.13) and those living at home (Table II.14). While it may be intuitive to assume that caregivers of ORS enrollees would have greater financial burden, these data indicate higher levels of emotional crisis for this group when compared to the Pre-LTC period. In summary, the QMs examined for location of care appear to have moved, if negligibly, in the desired direction from Pre-LTC to LTC. For NF enrollees, these QMs include reported declines in frequency of pain, falls, and unhealed pressure ulcers. However, for enrollees in home or other residential settings, caregiver crises show modest increases for caregivers of enrollees receiving services in ORS or at home. This is an important issue as caregiver crises are associated with movement into NFs. This, and other areas, suggest a need for continued scrutiny including, but not limited to, continued focus on data regarding the incidence and treatment of depression, site-specific foci (e.g., transitions from site of care) and external influences on quality of care. RQ2: What have managed care plans done to improve quality of care? In accordance with Federal regulations (42 CFR Part 438, subpart E), the Agency contracted with an external quality review organization (EQRO), Health Services Advisory Group, Inc. (HSAG), to conduct an independent assessment of the quality of each managed care plan. HSAG specifically reviewed the performance improvement projects (PIPs) conducted by each of Florida s managed care plans that are in the LTC program. Data from those reports is essential to answering RQ2. The scope of the PIPs is to monitor and improve the quality of care provided to enrollees. For the first year of the LTC program, the EQRO was limited to evaluation of the PIP methodology for each managed care plan. The PIPs were evaluated on criteria related to the study topic, question, population, indicators, sampling, and data collection. These criteria were set by CMS in the EQR Protocol 3: Validating Performance Improvement Projects (PIPs): A Mandatory Protocol for External Quality Review (EQR), Version 2.0, September Plans were required to do two PIPs (one clinical and one non-clinical), each targeting a specific area of care or service. Each LTC plan was required to conduct a PIP focused on Medication Review as well as a non-clinical PIP of their choice. See Table II.15 for a list of each plan s PIP topics. The Medication Review PIP is directed at enrollees 18 years of age and older who receive HCBS and consists of the evaluation of medications listed in the enrollee s medical record or case file conducted by a licensed nurse or a pharmacist in consultation with the enrollee s physician. Beginning in January 2016, dually eligible enrollees were excluded. Medication review has been shown to reduce the risk associated with adverse drug interactions, potentially inappropriate medications, and over or under use of prescriptions. The American Geriatrics Society Guidelines recommend conducting at least one medication review during the measurement year because it can substantially improve quality of care. 89 Final Report (Fiscal Year ) June 30, 2017

90 Managed Care Plan Table II.15. Performance Improvement Projects by Plan Performance Improvement Projects (PIPs) Technical Criteria American Eldercare Medication Review 1 Not Met American Eldercare Person-centered Care Plan Partially Met Amerigroup Medication Review Partially Met Amerigroup Improving the Number of Members with Advance Directives Not Met Coventry Medication Review Met Coventry Timeliness of Services for the Long-term Care Program Met Humana Medication Review Met Humana Advance Directives Met Molina Medication Review Partially Met Molina Reduction of Home and Community-Based Service Recipients Transferred to Nursing Facilities Partially Met Sunshine Medication Review Met Sunshine Influenza Immunization Met United Healthcare Medication Review Met United Healthcare Documentation of an Advance Directive Met Source: SFY External Quality Review Technical Report, Appendix B, Table B-3 1 Specifications changed to exclude dually eligible enrollees. Four plans met established criteria for both PIPs; one plan partially met both; and two plans did not meet one and partially met the other. When PIPs are unmet, plans are mandated to address any deficiencies. The plans have access to feedback and guidance in the PIP validation tools and may seek technical assistance from Health Services Advisory Group (HSAG). While these data support the presence and technical design of quality improvement plans, implementation of the plans and subsequent data collection will not be available for some years. The impact of the PIPs will be measurable in later evaluations. RQ3: How has timeliness of service delivery changed compared to the previous LTC programs and over time? Timeliness of service delivery relates to quality as well as to access to care. The evaluation team examined two sources of available data for information regarding timeliness of care: 1) satisfaction surveys and 2) case management reports. The breakdown by plan and site of care are measures of interest. According to the satisfaction survey results, respondents are asked about timeliness of the services they received with a question Overall, how often are your long-term care services on time? with response options of always, usually, sometimes, and never. Figure II.5 depicts answers to this question. Most enrollees reported services always on time regardless of plan. 90 Final Report (Fiscal Year ) June 30, 2017

91 Figure II.5. Percentage of Enrollees Reporting Services on Time Overall, how often are your long-term care services on time? 80% 70% 60% 50% 40% 30% 20% 10% 0% % % % % % of total Never Sometimes Usually Always No Response American ElderCare Amerigroup Coventry Humana Molina Sunshine United TOTAL Source: Managed Care Case Management File Audit Report Case Management Reports The evaluation team used data from audits of case management reports to evaluate the timeliness of contacts with enrollees. Case management audit reports are completed quarterly by the managed care plans. The results are presented in Figure II.6 through Figure II.9. Using this data is especially challenging. The data are kept in Excel files with text entries. There are inconsistent protocols for entering the data; therefore, data must be extensively cleaned before it can be entered in numerical form for analyses. 91 Final Report (Fiscal Year ) June 30, 2017

92 Figure II.6. Percentage of Enrollee Contacts Completed within Five Business Days of Enrollment for Community-Based Enrollees % Initial contact completed within 5 business days if in community 90.00% 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% AEC Amerigroup Coventry Humana Molina Sunshine United Total 10.00% 0.00% 2014 Q Q Q Q2 Source: Managed Care Case Management File Audit Report Figure II.7. Percentage of Enrollee Contacts Completed within Seven Business Days of Enrollment for Nursing Facility Enrollees % 90.00% Initial contact completed within 7 business days if in a nursing facility 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% AEC Amerigroup Coventry Humana Molina Sunshine United Total 0.00% 2014 Q Q Q Q2 Source: Managed Care Case Management File Audit Report 92 Final Report (Fiscal Year ) June 30, 2017

93 Figure II.8. Percentage of Case Manager Follow-up Calls Completed within Seven Days of Initial Assessment % Case Manager telephone follow-up call completed within 7 business days after initial assessment 90.00% 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% AEC Amerigroup Coventry Humana Molina Sunshine United Total 0.00% 2014 Q Q Q Q2 Source: Managed Care Case Management File Audit Report Figure II.9. Percentage of Calls Documented in the Case Notes % Contacts to enrollee that were attempted or made, documented in the case notes 90.00% 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% AEC Amerigroup Coventry Humana Molina Sunshine United Total 10.00% 0.00% 2014 Q Q Q Q2 Source: Managed Care Case Management File Audit Report 93 Final Report (Fiscal Year ) June 30, 2017

94 While there is some variation among plans, overall, contacts are being completed in a timely manner in most cases for community-dwelling (Home) enrollees (See Figure II.5) and enrollees residing in NFs (see Figure II.6) Follow-up calls after NF admission are not as reliable (see Figure II.7). Documentation remains high (See Figure II.8). The evaluation team will continue to monitor differences in plans in timeliness of services. RQ4: Are enrollees rights being addressed in accordance with the waiver and contract? These data are taken from the case management plan report provided by the individual plans to the Agency. Enrollee Rights The majority of enrollees in the LTC program have functional limitations. Unless the enrollee has lost decisionmaking capacity, a function that is retained is the ability to make informed medical decisions. Knowing one s rights is fundamental to exercising them. Knowing one s responsibilities is fundamental to ensuring that independence is maintained. Enrollees are often dependent upon care providers for information regarding their rights. Therefore, it is imperative that an explanation of enrollees rights be part of the LTC program enrollment process and, indeed, it is a requirement of the LTC plan contract. Documentation in the enrollee s file is the only available means of determining if this information was provided. Therefore, these data are abstracted from the enrollees case management files. Figure II.10 depicts findings from study Year indicating that plans generally have documented explaining the enrollees rights. Figure II.10. Enrollee Rights % Percentage of case files with evidence of Case Managers explaining enrollee's rights 90.00% 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% AEC Amerigroup Coventry Humana Molina Sunshine United Total 10.00% 0.00% 2014 Q Q Q Q2 Source: Managed Care Case Management File Audit Report 94 Final Report (Fiscal Year ) June 30, 2017

95 Personal Emergency Plan A personal emergency plan, also called an advance care plan, documents whether the person wants to be transferred to the hospital or ER, whether cardiopulmonary resuscitation (CPR) is desired, and the name of the surrogate decision maker. Sometimes this is in the form of a living will, an advance directive, a durable power of attorney for health care, or a Do Not Resuscitate Order. These are important discussions and documents used to ensure enrollees get the treatment they want and do not get treatment they do not want. Figure II.11 depicts that personal emergency plans are in place for over 80% of enrollees across plans. Figure II.11. Personal Emergency Plan % Percentage of enrollees with personal emergency plan in the case file 90.00% 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% AEC Amerigroup Coventry Humana Molina Sunshine United Total 10.00% 0.00% 2014 Q Q Q Q2 Source: Managed Care Case Management File Audit Report Primary Care Physician (PCP) The treatment team develops the long-term care plan in a LTC or home setting, usually without physician input. Ideally, the PCP would review and approve the plan, which would not be considered complete until it has been reviewed by the enrollee s PCP. Although this is not an LTC requirement, failure to identify the PCP could indicate a lack of appropriate communication and access. Figure II.12 indicates almost universal compliance with this procedure. 95 Final Report (Fiscal Year ) June 30, 2017

96 Figure II.12. Primary Care Physician Documentation % Percentage of enrollees with an identified PCP 90.00% 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% AEC Amerigroup Coventry Humana Molina Sunshine United Total 10.00% 0.00% 2014 Q Q Q Q2 Source: Managed Care Case Management File Audit Report Care Plan Summary The care plan summary should be the up-to-date summary of all care currently provided to the enrollee. Any provider, as well as the enrollee or named decision-maker, should be able to view it easily in the case file. A separate copy should always be given to the enrollee or their representative if the enrollee lacks decision-making capacity. Instrumental to the care plan summary is the assessment instrument. In NFs, the federally mandated instrument is the MDS, with a 701B Comprehensive Assessment also completed for Medicaid LTC enrollees. In ORS and home, the state mandated instrument is solely the 701B Comprehensive Assessment. These assessments also belong in the enrollee s case file. Figures II.13 and show that documentation of the care plan summary and assessment instrument is above 70% for all plans. 96 Final Report (Fiscal Year ) June 30, 2017

97 Figure II.13. Care Plan Summary Percentage of case files in which eligibility documents were present in case file (i.e., - LOC determinations, etc) % 90.00% 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% AEC Amerigroup Coventry Humana Molina Sunshine United Total 0.00% 2014 Q Q Q Q2 Source: Managed Care Case Management File Audit Report Figure II B Comprehensive Assessment % Percentage of case files in which the 701-B Assessment is present and completed properly 90.00% 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% AEC Amerigroup Coventry Humana Molina Sunshine United Total 0.00% 2014 Q Q Q Q2 Source: Managed Care Case Management File Audit Report 97 Final Report (Fiscal Year ) June 30, 2017

98 Screening for High Risk Persons with dementia or cognitive problems, psychiatric conditions, or functional limitations may be at high risk for receiving poor quality care because they are often unable to advocate for themselves. They need to be identified and a care plan needs to guide the care provided. Therefore, special screening for high risk is an important QM. Figure II.15. Screening for High Risk % Was there evidence of special screening for and monitoring of high risk persons and conditions documented in the case file? 90.00% 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% AEC Amerigroup Coventry Humana Molina Sunshine United Total 10.00% 0.00% 2014 Q Q Q Q2 Source: Managed Care Case Management File Audit Report Note: Humana informed the Agency that internal review errors contributed to the low percentage reported in 2015 Q1 Complaints/Issues Another means of addressing enrollees rights is through the complaint process. Complaints are defined as any expressions of dissatisfaction with the Florida Medicaid program. Although most complainants report one problem when they contact the Agency, some report multiple problems. These are recorded as Issues. The Agency encourages all stakeholders to report any potential issue, concern, or complaint regarding the SMMC Program to the SMMC Complaint Operations Center. All allegations and issues are recorded, regardless of whether they are found to be accurate or substantiated. Complaints/issues were reported via the AHCA Complaint Center. Data related to complaints and resolutions are noted below under findings for RQ5 (see Figures II.16, II.17, and II.18). 98 Final Report (Fiscal Year ) June 30, 2017

99 RQ5: How effective are the LTC plans complaints, grievances and appeals processes? Information on the Agency s website indicates that, Once a complaint or issue is submitted online to the Agency s complaint/issue center, one of our Agency staff will contact the complainant within one business day of submitting a critical or high priority complaint. (Retrieved from: The Agency reports from July 2014 to June 2015 provide the average or median days to resolution of complaints/issues raised by enrollees or providers. In any given week, the number of complaints can vary, from low to high. Small numbers of complaints in these figures are particularly sensitive to those few cases with very high or very low values, and do not necessarily reflect poorly on the entity reporting. Figure II.16 depicts data from SFY Figure II.16. Average Days to Resolution, Beneficiaries Issues Average Days to Resolution Jul-14 Aug-14 Sep-14 Oct-14 Nov-14 Dec-14 Jan-15 Feb-15 Mar-15 Apr-15 May-15 Jun-15 American Eldercare Amerigroup Coventry Humana Molina Sunshine United Source: Data Submitted by LTC plans to the Florida Agency for Health Care Administration During SFY , disposition resolutions were divided into two categories: pending and resolved. By June 2015, American Eldercare reported 5, Amerigroup reported 41, Coventry reported 4, Humana reported 8, Molina reported 33, Sunshine reported 12, and United reported 107 complaints with dispositions pending. By June 2015, American Eldercare reported 10, Amerigroup reported 175, Coventry reported 4, Humana reported 15, Molina reported 32, Sunshine reported 28, and United reported 69 complaints with dispositions resolved. During the evaluation period, the number of unique LTC enrollees included 106,413 individuals. This included 25,676 (United), 42,813 (Sunshine), 6,926 (Molina), 5,884 (Humana), 5,881 (Coventry), 5,716 (Amerigroup), and 15,904 (American Eldercare). Nine types of grievances an expression of dissatisfaction about any matter other than an adverse benefit determination were reported in SFY : quality of care, access to care, medically necessary, excluded 99 Final Report (Fiscal Year ) June 30, 2017

100 benefit, billing/claims dispute, enrollment/disenrollment, out of plan service authorization, in plan service authorization, and pharmacy benefits. Five grievance disposition types were reported in SFY : claim paid, procedure scheduled, referral for non-plan Medicaid benefits, unable to contact enrollee, and enrollee withdrew grievance/appeal. Coventry (n=1) and Molina (n=7) reported dispositions ending with claim payments. Humana reported one complaint being resolved by having a procedure scheduled. Sunshine resolved five complaints by making referrals to Medicaid field offices for non-plan Medicaid benefits. United reported 9 complaints being closed out because they were unable to contact the enrollee. Three plans had dispositions of the enrollee withdrawing his/her grievance: Amerigroup (n=1), Humana (n=2), and Sunshine (n=1). RQ6: What are the rates of complaints, grievances and appeals in the LTC program and have these changed over time? For the LTC program period, the evaluation team is reviewing plan-reported rates of complaints/issues. No data on complaints, grievances, or appeals were available for the period preceding the rollout of the LTC program. This report only includes first year full report (SFY14-15) of the LTC program. Future reports will compare subsequent program years. Figures II.17 and II.18 depict complaints/issues the plans received and reported to the Agency for the July 2014 to June 2015 time period. During the evaluation period, the number of unique LTC enrollees included 106,413 individuals. This included 25,676 (United), 42,813 (Sunshine), 6,926 (Molina), 5,884 (Humana), 5,881 (Coventry), 5,716 (Amerigroup), and 15,904 (American Eldercare). Figure II.17. Complaints/Issues Reported by Plan from July 1, 2014 to June Issues Reported from July 1, 2014 to June 30, 2015 by Plan United 176 Sunshine 40 Plan Names Molina Humana Coventry Amerigroup 216 American Eldercare Number of Issues Reported Source: Data Submitted by LTC plans to the Florida Agency for Health Care Administration 100 Final Report (Fiscal Year ) June 30, 2017

101 Figure II.18. Numbers of Complaints/Issues per Quarter by Plan Number of Complaints per Quarter by Plan American Eldercare Amerigroup Coventry Humana Molina Sunshine United Jul-Aug 2014 Oct-Dec 2014 Jan-Mar 2015 Apr-Jun 2015 Total Source: Data Submitted by LTC plans to the Florida Agency for Health Care Administration Based on input from the Agency and findings from the administrative focus group, the evaluation team believes that tracking of complaints may be significantly enhanced with the introduction of the LTC program. Given the fragmentation of complaints data prior to this implementation, the evaluation team recognizes that collection of complaints data was not ideal. Now, with one contract manager for each plan who monitors complaints daily, more complete and valid data should be available in the future for this highly important quality measure. A revised complaint form was posted online in April It now allows for the uploading of documents with a complaint. This revision was requested by both recipients and providers. The revised form contains targeted questions that solicit information specific to each type of complaint. The questions on the new form were specifically designed from lessons learned during LTC implementation, to better solicit responses that allow the Agency to route complaints more quickly and accurately. Reporting complaints is not limited to enrollees or their responsible parties. The Agency has made efforts to address potential challenges people may face in trying to report a complaint including making the complaint form available in English, Spanish and Creole. Further, in recognition that the complaint form (available at may be difficult to access and complete, in January 2015 the Agency began encouraging people to call rather than go online to report an issue in hopes of making reporting easier. Furthermore, how to pursue a complaint is explained in the Enrollee Handbook and case managers are responsible for explaining enrollee s rights and responsibilities. However, despite these efforts, qualitative interviews found that enrollees generally do not recall how to register a complaint. RQ7: Is there sufficient monitoring and oversight of the program by the state? What types of monitoring are done, and how are they conducted? The evaluation team continues to note that the Agency and DOEA personnel are committed to providing the best possible services to LTC enrollees and their care providers. The roll-out year (SY ) was a period of transition in which a number of facets of the program were challenging. In this report regarding SFY , 101 Final Report (Fiscal Year ) June 30, 2017

102 the evaluation team replicated the analyses conducted during the Year report using transition period data. While the Agency leadership is no longer examining complaint reports on a daily basis, the LTC Oversight Unit and their corresponding contract manager within the Bureau of Medicaid Plan Management Operations review their plan s complaints weekly and take appropriate action to ensure compliance. The Agency continues to value enrollee satisfaction and feedback and remains committed to meeting the enrollee needs across all three sites of care. In addition, the Agency monitors plan performance using multiple reports routinely provided by plans. For example, Case Management File Audit Reports and Case Management Monitoring and Evaluation reports are submitted quarterly. Other monthly reports include Missed Services, Denial, Reduction, Termination or Suspension of Services Report, and Enrollee Roster and Facility Residences Report. The Agency also has established performance measures (e.g., case manager training and satisfaction with long-term care plan) that are used to evaluate plan performance. Failure to meet benchmarks established for performance measures may result in monetary fines or other enforcement actions. In addition, the Agency currently provides information that is extremely helpful in the oversight of the SMMC LTC programs. The Agency continues to issue the Quarterly SMMC Reports, each focusing on a specific area of care. Also, the DOEA Consumer Independent Report provides a detailed description of systems that monitor complaints. The final portion of that report suggests that efforts are underway to consolidate complaint data from various sources so as to meet CMS requirements. Despite earlier recommendations, there are some important data that are collected by the plans, but are not made available to the Agency. These data, especially enrollee outcome data, could provide an important source of information that transcends sites of care and better addresses the heart of the program, the health and well-being of individual enrollees. The evaluation team chooses to restate the following recommendations that were part of the Year report. Based on data from SFY , there have been no indications that these recommendations should change. Recommendations to Enhance Agency Oversight Use Enrollee Outcomes As noted above, the Agency uses a series of reports to evaluate quality of care provision. These reports rely on processes and documentation of implementation. However, throughout the literature on measuring the quality of care provision, individual (i.e., enrollee) outcomes are the gold standard. Therefore, enrollee outcomes, assessed on an annual basis, ought to be the gold standard for measuring the quality of care within SMMC LTC plans, independent of who provides the care. That is, these outcomes are important regardless of which entity (e.g., NF, ORS staff, or plan) administers and/or implements the actual provision of care. For example, one enrollee outcome of interest is pressure sores. Increasing prevalence of pressure sores is an indicator of poor healthcare quality. While plans do not provide medical care in NFs, plans can begin monitoring for pressure sores among NF clients as soon as an LTC enrollee is in their plan, particularly since the majority of LTC program enrollees are already in NFs. Additionally, plans indirectly affect enrollee location of care through the plan s choice of NFs in their networks. Enrollees or their authorized representatives can choose from among those NFs in the plan network. It should be noted that plans have limited ability to exclude NFs from their network. While most NFs meet minimum quality thresholds, if plans contract with NFs that provide low quality care, they can, in effect, contribute to the increased risk of developing pressure sores. Also, plan case managers monitor care within all 102 Final Report (Fiscal Year ) June 30, 2017

103 settings, and, therefore, indirectly impact quality of care. The evaluation team notes that it is imperative that these enrollee outcomes are the gold standard rather than focusing solely on processes or documentation. This goal is closely aligned with the following recommendation(s). Collect and Utilize Existing Data The evaluation team has noted that there are areas in which the DOEA as delegate of the Agency is monitoring the plans using enrollee data. For example, plan personnel are required to complete face-to-face, annual evaluations with all enrollees. This evaluation is completed using a form of 701B Comprehensive Assessment for ORS and home enrollees. Form 701T, Non-Community Placement Assessment, is completed for enrollees residing in NFs. The plans are submitting completed assessments to CARES. Assessments are reviewed by CARES staff and DOEA s Person-Centered Monitoring Team and these reviews provide a picture of the psychosocial and functional needs of enrollees; data from Form 701T are not provided to the Agency. This is an important omission as these data are used in making critical decisions regarding location of care, specifically transitioning from the NF to the community. Furthermore, if these data were available through the Agency, the evaluation team could review these data while evaluating quality indicators for enrollees who move from one site of care to another. Another important source of quality information is the annual level of care (LOC) redetermination. Until recently, DOEA had a performance measure that assessed plan compliance with the requirement to conduct annual LOC redeterminations. These data had been reported to the Agency up until July 30, However, CMS discontinued the federally-mandated performance measure. Since both sources of data (701B Comprehensive Assessment and LOC Redeterminations) are being collected and submitted, the former (701B) by the plans to DOEA, and the latter (LOC Redeterminations) collected by DOEA this recommendation is not overly onerous to implement. Finally, the evaluation team suggests that the 701B Comprehensive Assessment replace the 701T Non- Community Placement Assessment for enrollees living in NFs. The 701B Comprehensive Assessment includes outcomes, specifically items regarding quality of life (items G. 76 & 77) and unmet need (Items D. 70, a-f; E. 72, c, d, f, g, & h), that are not included in the 701T and are appropriate for NF enrollees. Quality of life is considered a single, overarching outcome that is essential in measuring the success of care provision. It is important to measure quality of life and unmet need once enrollees are in the community as compared to responses during NF residency. Recommendations for Future Analyses Attend to vulnerable sub-populations The stated goals and financial incentives of the SMMC LTC program include, indeed concentrate on, moving enrollees from NFs as a site of care back into the community for ORS or home care provision. Transitions for older adults are an established source of stress, often resulting in depression, cognitive decline, and even mortality. Therefore, those enrollees undergoing transitions are especially vulnerable. While the move from a NF to the community appears to be a positive event, there needs to be special attention paid to the quality of life, 103 Final Report (Fiscal Year ) June 30, 2017

104 health, and cognition of these individuals. Using the 701B Comprehensive Assessment for enrollees in NFs as recommended above, would allow for monitoring the same outcomes within the NF as would be available upon return to the community. The evaluation team also recommends more frequent assessments during the first year after any transition. In addition to these recommendations, the evaluation team notes that sub-populations may be differentially affected by the LTC service model. The transition year ( ) and the current study year ( ) have appropriately focused on comparisons with Pre-LTC care as measured by established QMs. Hence, analyses have been conducted at the population levels to establish trends related to a significant structural change. However, in future years, the evaluation team seeks to look at differences among groups within the larger universe of enrollees. For example, since the plans vary according to regions, enrollees in certain geographic areas may receive differing quality of care. Indeed, comparisons can be made among plans to determine best practices or protocols. Furthermore, enrollees moving from one site of care to another are of particular interest in that the moves can affect quality of care. This is especially important for those moving from NFs to communities, where much care is provided by informal care providers such as family members. As movement into communities is a goal of the SMMC LTC, examination of the effect on quality of care is paramount. Also, further examination of the role and capacity of case managers is essential. This component of care provision is subject to triangulation by examining case management audit reports, encounter data, and cost effectiveness data. Questions in the qualitative interview also relate to case manager effectiveness. As case managers provide an interface between the SMMC LTC and medical portions of health care provision, understanding their roles is essential. For example, the evaluation team can examine numbers of enrollees assigned to each manager by plan and relate these data to QM outcomes. While the evaluation team does not suggest determining a causal relationship, performing more detailed and nuanced analyses will provide greater insight into areas of improvement and action items for implementation. Continue to review and analyze data in greater detail Differences among plans during the transition year ( ) and first full year of implementation ( ) suggest that reporting for future years be presented by plan, as well as by site of care. Further analyses in future evaluations also will illuminate differences based on demographic categories and individual characteristics. Finally, these data suggest further examination of target populations (e.g., enrollees transitioning from NFs to Home; race and ethnicity; urban vs rural) in order to determine the effects of these transitions and factors on quality of care. Another method of analyzing data at the facility level will address issues of facility choice as well as enrollee outcomes. Furthermore, as data for additional years become available, the evaluation team will look at trends over time. Summary The evaluation team notes that Agency and DOEA administrative personnel are committed to providing outstanding services to enrollees. Current monitoring practices to assure that outstanding service is received can 104 Final Report (Fiscal Year ) June 30, 2017

105 be enhanced by: a) using enrollee outcomes (both medical and psychosocial) as the gold standards when measuring quality; b) collecting and utilizing existing data to an even greater extent; and c) attending to vulnerable populations, especially those undergoing transitions. Qualitative Findings The inclusion of a qualitative component (i.e., enrollee interviews) presents a unique opportunity for enrollees to voice their perceptions of the services they are receiving. Qualitative methods allow for understanding phenomena from the emic or lived experience. Indeed, qualitative interviews allow enrollee respondents to identify issues that are important to them that others (i.e., experts ) may not know about and, therefore, not ask about. Overall, this component can give enrollees a sense of empowerment as well as provide important information not available from any other source. For this study year (SFY ) the Agency and evaluation team collaboratively revised the interview guide. The newly revised guide is available in the Appendix. The evaluation team interviewed thirty-one enrollees across all three sites of care. Figure II.19 on the following page depicts the data collection (interview) process. Of the 1,274 potential participants, thirty-one people agreed to be interviewed about their LTC services. Five of these individuals received services in NFs, 12 in ORS, and 14 in home. Most home participants interviewed received medication assistance and assistance with completing ADLs. Few participants remembered changes to their LTC services from three years ago, and were, therefore, unable to remember if there were any substantial changes to their care. Notably, some of the participants were not enrolled in LTC services at the time of the change. Overall, the enrollees who were interviewed had positive feelings towards their LTC services. However, there remain areas for improvement. 105 Final Report (Fiscal Year ) June 30, 2017

106 Figure II.19. Qualitative Data Collection Analyses of the interview transcriptions led to the findings regarding services, enrollee rights, plans of care, case managers, and the plans under which they receive services. The findings described below also are reflected in the recommendations. Services Overall, enrollees responded positively regarding their services. However, almost half indicated they had unmet needs. These included needs for additional help, additional equipment, more doctors, and transportation. Despite these unmet needs, most indicated positive responses to both quality of care and quality of life. Enrollees in NFs and ORS generally did not indicate that they were involved in choosing their plan or facility in which they reside. Rather families made that decision for them or they did not know they had a choice. For example, one NF enrollee stated The long-term service provider we never really had a choice We didn t choose it they chose it for us. However, those that did make the choice felt empowered such as a resident who stated I make my decisions on things I feel confident enough that I can. Others preferred a shared-decision making model as exemplified by I didn t choose it, my daughter found it for me and she found this plan. I trust my daughter...and 106 Final Report (Fiscal Year ) June 30, 2017

107 she oversees everything. As noted above, most did not know there had been a change in the delivery of services or recognize the SMMC LTC program as part of the delivery system. Rights In response to the direct question Do you know your rights as an enrollee in the Long-term Care Program? most respondents answered no. A typical response might be Well I don t know about that but probably I do [have some rights] or Do I have rights? When their rights were provided (read) to them by the interviewer, all enrollees indicated that their rights had been respected, not violated. However, few respondents knew how to report an issue to the appropriate entity. Generally, those enrollees living in ORS stated they would report issues to the facility administrator. Some enrollees living in NFs indicated that they would look to nursing staff to resolve issues; others thought they should report to the Council on Aging. Still other enrollees sought to contact their case managers. None mentioned a call-in line or online portal for reporting issues. As one enrollee stated They probably gave me something, but I can t keep up with all that. Plan of Care Most enrollees were comfortable with their plans of care, yet they also reported that they were not involved in developing one. A small number said their families were involved. Case Manager The case manager is the one individual who represents the LTC plans to all enrollees, regardless of site of care. Therefore, enrollees were asked questions that were specifically about their case manager, apart from other services. Almost two-thirds of enrollees reported being contacted by their case managers. One enrollee who lived at home stated, I see Pierre uh once every three months but he calls me and checks on me like every couple of weeks. All enrollees reported that the case manager was respectful. Some indicated turnover in case managers as My caseworker has changed and got another one is what I m trying to say I haven t met her face to face. I ve talked to her twice on the phone and she was very kind and friendly on the phone. Others indicated a desire for more contact as this enrollee living at home it s been comfortable talking to her and stuff, you know, you, I have no complaints I would just like to see her. But I don t know how many patients she s got and she started in the middle of January and it s fifty miles for her to come and I understand, you know. A much smaller percentage (less than one fourth) indicated they did not know their case manager or said the case manager was absent. Plans Enrollees were generally comfortable with the plans administering their care. About a third said they were involved in choosing their plans. Others stated that their families chose the plan with a small number stating that they had no choice. Despite statements regarding plans and their care, about half did not know the name of their plan. Summary Of the 31 enrollees interviewed, most indicated that they are comfortable with their services and plans of care. However, it appears that many enrollees are not directly involved in decisions that impact their health and well- 107 Final Report (Fiscal Year ) June 30, 2017

108 being. The gold standard of person-centered care is demonstrated by patient (enrollee) involvement in setting and evaluating goals of care and quality of life. Information received from these interviews indicate many enrollees are not participants in the decision-making process. Enrollees appear to lack knowledge regarding rights and how to report issues that arise. Most enrollees in NFs and ORS appear to look to facility staff to resolve issues which may present a conflict of interest should the issue reflect poorly on the facility. Coupled with the enrollees reports of unmet need, it is imperative that these deficits be resolved. These interviews affirmed the importance of the case manager as the point of contact with plan personnel. In future evaluations, the evaluation team will focus on case management. While these interviews are conducted with a small proportion of enrollees, they provide a platform by which enrollees can directly articulate their experiences. Conclusions The evaluation team, along with the Agency and DOEA administrators, continue to define healthcare quality for enrollees in the LTC program. This definition reflects a dynamic process as successes and deficiencies in the current program come to light within the context of empirical and evolving knowledge. This report for Year confirms some of the findings from the transition year ( ). Embedded within this definition are components of person-centered care, a value espoused by CMS, AHCA, and other stakeholders in LTC quality. The evaluation team will, whenever possible, continue to rely on enrollee outcomes rather than process measures and value enrollee perspectives. The key findings reflect that many indicators of quality in SFY are similar if not identical to findings in the transition year, SFY Furthermore, when compared to the pre-ltc period, most QMs have improved, although improvements are slight. Other Residential Settings Self-reported overall health among ORS enrollees improved with the percent of those rating their health as fair or poor dropping, and the percent of those reporting their health as excellent/very good and good increasing from Pre-LTC to LTC. The percent of ORS enrollees indicating they had someone who could help them if needed, beyond their caretaker, increased. Home When asked how satisfied they are with their overall quality of life, those enrollees residing at home in LTC were more likely to indicate that they were satisfied or very satisfied compared to enrollees residing at home Pre-LTC. Self-reported overall health improved significantly among home enrollees from pre- LTC to LTC with those rating their health as poor dropping and the percent rating their overall health as good nearly doubling. More LTC enrollees residing at home indicate that they have someone other than their primary caregiver who can help them if needed (i.e., social help), increasing from 49.51% Pre-LTC to 58.95% LTC. Most LTC respondents gave high ratings to LTC services and LTC plans. Overall, unmet need in home based enrollees increased. 108 Final Report (Fiscal Year ) June 30, 2017

109 Nursing Facility QMs These data indicate a very slight change in the preferred direction for pressure ulcers, falls, and pain when LTC Year is compared with Pre-LTC. Depression The evaluation team notes that findings regarding the incidence and prevalence of depression based on scores form the PHQ-9 across all sites of care were not consistent with Year findings; however, they remain low in comparison with national statistics. The evaluation team will continue to target depression as a QM and triangulate the PHQ-9 scores with prescription of anti-depressant medications and encounter data. Caregiver Crisis As caregiver crisis is a significant risk factor for institutional care, it is of concern that Year data indicate a modest increase in caregivers in crisis for caregivers of enrollees in ORS and home. These data suggest the need for further study, perhaps interviews with caregivers to provide an understanding of the dynamics of care provision as related to the SMMC LTC program. This extended examination is especially important for enrollees living at home as much care is provided by informal caregivers such as family members. Recommendations Below are recommendations that the evaluation team continues to endorse as important. The evaluation team, along with the Agency and DOEA administrators, continue to define healthcare quality for enrollees in the LTC program. This definition reflects a dynamic process as successes and deficiencies in the current program come to light within the context of verifiable observations and evolving knowledge. Embedded within this definition are components of person-centered care, a value espoused by the Centers for Medicare and Medicaid Services (CMS), Florida Agency for Health Care Administration (AHCA), and other stakeholders in LTC quality. These recommendations reflect those values. 1. Perform comprehensive annual assessments for all enrollees using 701B Comprehensive Assessment instrument or modification of the 701T for NF enrollees. The Agency mandates that case managers perform an annual assessment for all enrollees. However, in the case of NF enrollees, the 701T Non-Community Placement Assessment is used rather than the 701B Comprehensive Assessment. Using the 701B Comprehensive Assessment would provide a more complete picture of NF enrollees, especially regarding self-reported, psychosocial measures. Notably, case managers, plans and the Agency do not have access to MDS 3.0 assessments, so these once-ayear assessments represent the only data regarding enrollee outcomes that are accessible. 2. Require that plans submit all findings from annual assessments to the Agency. While plans are required to conduct annual 701B Comprehensive Assessments and 701T Non- Community Placement Assessment for NFs, the plans are not required to submit all data to the Agency. These data, especially those used for determining transitions, represent a rich source of information regarding enrollee welfare that could serve as outcome measures for continuous evaluation. These data 109 Final Report (Fiscal Year ) June 30, 2017

110 would be relevant to the Agency for monitoring enrollee welfare, and could be shared between the agencies through appropriate protocol. 3. Review the protocol and training for conducting the 701B Comprehensive Assessment, specifically as related to measures of depression. The inconsistent nature of these findings regarding screening for, diagnosing, and treating depression suggests that some attention be paid to all aspects of this important quality measure. Training materials need to state clearly the implications of these items and the importance of asking each question regardless of the case manager s impression of the enrollees mental health status. Training materials reviewed by the evaluation team ( emphasized appropriate responses to enrollees who have suicidal thoughts. Obviously, this is an important safety issue; however, depressive symptoms vary in intensity and assessment administrators need to attend to each question. Also, they need to understand that it is normal that enrollees report some incidence of behaviors contained in the PH-Q9 items. What is not normal is the incidence and prevalence of enrollees with no positive responses. 4. Continue to evaluate the indicators of quality that have been identified in this report. Quality of life, unmet need, satisfaction with care, diagnosis and treatment of depression, falls, pressure ulcers, pain, and social needs are important indicators of quality in LTC services. These indicators were chosen for inclusion in this evaluation following extensive literature reviews and based on expert evaluation. Where these data are available, they should be examined by the Agency for abnormalities (e.g., the extremely low scores on the PH-Q9). Any future revisions to the 701B Comprehensive Assessment, the 701A Condensed Assessment and the 701T Non-community Placement Assessment should be tailored to include or expand upon data related to these specific QMs. Where data are not currently collected, as in quality of life for NF enrollees, processes need to be altered to enable data collection as noted in Recommendation Interview specific subpopulations. In addition to enrollees, the evaluation team recommends interviewing stakeholders such as case managers and plan administrative personnel. These individuals could be rich sources of data for improving implementation of plan contracts, interactions with enrollees, and areas in need of attention. While the enrollees themselves are the gold standard of care evaluation, process issues may be best identified by plan employees and administrators. For example, when contacting enrollees, the evaluation team has spoken with NF and ORS administrators. When told of the reason for the enrollee interview, administrators usually state at least one recommendation for improving the program or issue that consistently arises (e.g., transportation). Formalizing this input would be helpful in identifying issues for further study. 110 Final Report (Fiscal Year ) June 30, 2017

111 III: Cost-effectiveness of Care Purpose The Independent Assessment of the Florida Statewide Medicaid Managed Care Long-term Care Program assesses the cost-effectiveness of care of the program implemented on August 1, 2013 using quantitative methods. A detailed report of specific methodologies employed and data sources used for the assessment are provided in Appendix 3 of this report. The following research questions were used to guide the assessment of cost. 1a. How has the LTC program affected the growth of Medicaid costs for LTC program enrollees? 1b. How has the shift between HCBS and nursing facilities under the LTC program affected enrollees' Medicaid costs? The cost analyses take into account the LTC program s programmatic change. Prior to the LTC program, the Agency paid Medicaid providers of LTC services primarily on a fee-for-service basis, with the exception of Nursing Home Diversion (NHD) and the Frail Elder Option. The LTC program implements a capitated payment system. 22 In order to determine how the costs of the capitated LTC program compare to the costs of LTC services prior to the LTC program s implementation, the following calculation was performed: 23 Cost-Ratio = Average Monthly Inflated Costs for LTC Services Provided Pre-LTC Program / Average Monthly Costs for LTC Services Provided Under the LTC Program This ratio is used throughout the cost analyses to compare the cost of LTC services before implementation of the LTC program (in particular SFY and SFY ), to the cost of LTC services beginning with the full implementation of the LTC program in the first full LTC Program Year SFY The implementation year SFY is omitted for reason of the staggered implementation. The interpretation of the cost ratio mentioned is straight-forward (as seen in Table III.1). When the cost-ratio is greater than 1.0, then the payments made by the Agency for providers of Medicaid LTC services before the introduction of the LTC program were higher than the payments which were being made under the LTC program. For example, if the cost-ratio is 1.30, then the Agency s payments to providers of Medicaid LTC services before the implementation of the LTC program were 1.30 times higher than that of providing LTC services under the LTC program. Placing the calculation in US dollar terms, if the cost-ratio is 1.30, then for every $1.30 the Agency paid to providers for Medicaid LTC services before the LTC program was implemented, $1.00 is now being paid under the LTC program. 22 During the first year of the LTC program, there was one FFS PSN that became capitated on September 1, The Cost-Ratio and the Cost-Difference are defined here as measures for Cost-effectiveness, contrary to the usual measure of effect or product in relation to its cost. 111 Final Report (Fiscal Year ) June 30, 2017

112 In addition to the cost-ratio, the estimated cost-differences were also calculated with the following equation: Cost-Difference = Average Monthly Inflated Costs for LTC Services Provided Pre-LTC Program - Average Monthly Costs for LTC Services Provided Under the LTC Program The interpretation of cost-difference is straight-forward. Cost-neutral equals zero. A positive number would reflect the average per-member per-month (PMPM) savings associated with the LTC program. A negative number would indicate service costs under the LTC program were higher than the service costs in the Pre-LTC program period. Findings RQ: 1a. How has the LTC program affected the growth of LTC Medicaid costs for LTC program enrollees? Two methodologies are applied to the RQ; a difference analysis (as seen in Table III.1), and an interrupted time series approach (as seen in Figure III.1). Interrupted time series (ITS) is a tool to see whether the implementation of the LTC program changed the trend in the PMPM claims. Throughout this section PMPM is defined as costs related to the appropriate LTC-buckets. 24 Table III.1 (following page) provides aggregate mean monthly Medicaid claim amounts paid by the Agency for the Pre-LTC and LTC program populations, using the Inflation Factors Methodology and Difference Analyses. 25 Table III.1. Pre-LTC Program Per Member Per Month (PMPM) Claims SFY , SFY , and LTC Program Per Member Per Month (PMPM) Claims SFY Item PRE-LTC Program SFY PRE-LTC Program SFY LTC Program SFY SFY to SFY Differences PMPM SFY to SFY Differences PMPM Total: PMPM x Average Utilization per month $ 297,021,087 $ 307,603,171 $ 297,911,131 $ 3, $ 3, Comparative $ 297,021,087 $ 307,603,171 $ 3, $ 3, Cost Difference - $ 10,582,084 $ 9,692,040 - $ $ 0.89 Cost Difference due to Utilization - $ 4,414,705 $ 9,767,496 - $ $ Cost Difference due to Price changes - $ 6,167,379 - $ 75,456 - $ $ 0.89 Total Cost Difference - $ 10,582,084 $ 9,692,040 - $ $ 114,42 Cost Difference due to Price changes - $ 6,167,379 - $ 75,456 - $ $ 0.89 Cost Difference due to Inflation - $ 6,279,475 - $ 77,931 - $ 71,78 - $ 0.92 Real Cost Difference $ 112,096 $ 2,475 $ 1.28 $ 0.03 *Note: Negative values indicate a higher cost, positive values a lower cost or cost-savings. Source: Medicaid Claims Data 24 PMPM is defined as the costs related to in the following buckets: 6 Skilled Nursing Care, 7 ICF I Services and 8 ICF II Services, combined in bucket NH (Nursing Home), bucket 33 Hospice Services, 35 Home and Community Based Aging, 36 Home and Community Based DS, 37 Aids Waiver Services, 42 Personal Care Services, 61 Adult Congregate Living Facility, 66 Nursing Home Diversion Waiver, 67 Brain and Spinal Cord Injury Waiver, 71 Assistive Care Services, 73 Channeling, 79 Alzheimer s Waiver, 81 Adult Day Care Waiver, 89 PACE, 93 LTC Admin Fee and 94 Prepaid LTC. 25 For methodology and equations used, see Appendix Final Report (Fiscal Year ) June 30, 2017

113 The average nominal increase per month is $0.89, i.e., a rise in claim value PMPM, comparing the Pre-LTC SFY with the full statewide implemented LTC Program SFY Multiply this by the average monthly enrollment during the statewide LTC Program Implementation, and the expected monthly cost increase in aggregate becomes $75, However, adjusting for inflation between the two periods, the average real monthly claim amount decrease, or savings, is $ 0.03 per member. A perspective on the nominal trend in Medicaid costs during both periods aforementioned is provided in Figure III.1, next to the results of the interrupted time series (ITS) analyses as per Table III Figure III.1 shows the actual nominal values before and after the LTC program implementation (observations from the implementation period during SFY are omitted). PMPM values for the months of July and August 2014 were disregarded, as these two months are obvious outliers from the perspective of the ten remaining LTC program months. Figure III.1. Interrupted Time Series Results on Nominal Per Member Per Month (PMPM) Claim Values, Pre-LTC Program SFY , SFY , and LTC Program Per Member Per Month (PMPM) Claims SFY Source: Medicaid Claims Data The red or bold straight lines represent the overall ITS-regression model. The dotted red line reflects the expectations under business as usual (i.e., without implementation of the LTC program). The difference between the trend lines is the expected cost-savings following LTC program implementation ,703 average monthly enrollees * $0.89 estimated value increase per- individual per- month = $ 75, estimated monthly cost increase. 27 The ITS equations are given in Appendix Final Report (Fiscal Year ) June 30, 2017

114 The lines for Pre-LTC program months and LTC program months are different, both in PMPM and direction of PMPM change. 28 This change in slope is both statistically and substantively significant. The changes in averages over the expected values for the three SFY are given in Table III.2. Table III.2. Average Expected Per Member Per Month (PMPM) Values Based on Interrupted Time Series. SFYs PMPM SFY SFY SFY without LTC program ( *X 1, t )/12 = $ 3, $ 3, $ 3, with LTC program* ( *X 1, t *X 2, t *X 3, t )/12 $ 3, $ 3, $ 3,535.68** *Disregarding the months of July and August 2014 in the IT Analysis **Based on July and August SFY as without LTC program and the remaining ten months of SFY as with LTC program As shown in Table III.2, the average expected increase between SFY and SFY without the LTC program was an increase of $34.83 PMPM. 29 The PMPM difference between LTC program implementation (2 nd row) versus without LTC program (first row) is a decrease of $31.20 PMPM. The latter difference (-$31.20 PMPM) multiplied by the 84,703 enrollees in SFY makes the expected costs for the LTC program $2.6 million less expensive per month in the short-run. Where the comparative analyses show a virtual cost neutrality, the outcomes calculated with ITS differ. This is because the comparative analyses compare values in SFY with values in SFY (i.e. two data points in time with a one-year gap), while ITS analyses compares expected values in SFY (i.e. data points at the same time/year) under changed scenarios. In addition, in the ITS analyses three years of monthly data is used, instead of annual averages, the values for the LTC program months July and August were excluded as outliers, and overall the ITS comes with a rather low fit (or R 2 -value) given rather volatile data (especially in the pre-ltc program data). The comparative analyses show some important breakouts as causes for cost differences (e.g., utilization, price and inflation). The ITS analyses do show a significant change in the trend of cost, not only to PMPM but also to the direction of PMPM. The caveat clearly lies in the short timeframe under consideration. Consecutive years of added SFY data will be needed to verify the consistency of these preliminary findings. RQ: 1b. How has the shift between HCBS and nursing facilities under the LTC program affected enrollees Medicaid costs? Table III.3 provides average PMPM claim details based on a selection of service buckets (or profile combinations) taken over SFY Frequencies of profiles are provided in the head column, average PMPM in the 28 The slope of the pre-intervention is (b1), given that X2,t and X3,t assume zero values, while the slope of the post-intervention line is (b1 + b3 or ). 29 The two-year change between SFY (without the LTC program) and SFY is $69.65; which is $34.83 PMPM per annum. 114 Final Report (Fiscal Year ) June 30, 2017

115 second, and bucket breakouts in the subsequent columns. 30, 31 Red or shaded cells show fields with relative higher average PMPM costs. FREQ Table III.3. LTC Program Claims SFY Per Member Per Month Bucket Profile. LTC PMPM NH ,507 $ 3,527 $ 3,527 9,959 $ 6,040 $ 159 $ 5,731 $ 151 8,977 $ 1,064 $ 137 $ 927 6,112 $ 1,382 $ 127 $ 161 $ 1, $ 5,746 $ 155 $ 5,432 $ $ 1,410 $ 129 $ 162 $ 1, $ 3,786 $ 3,679 $ $ 6,714 $ 3,693 $ 3, $ 153 $ $ 5,451 $ 160 $ 2,864 $ 2,276 $ $ 5,664 $ 163 $ 5, $ 1,230 $ 129 $ 1, $ 1,377 $ 160 $ 586 $ $ 3,933 $ 3,659 $ $ 4,360 $ 3,475 $ $ 4,268 $ 136 $ 3,197 $ 150 $ $ 2,933 $ 2,794 $ $ 7,860 $ 3,648 $ 4,212 9 $ 5,042 $ 3,195 $ 1,847 8 $ 4,771 $ 145 $ 3,784 $ 150 $ $ 5,399 $ 129 $ 5,270 5 $ 7,408 $ 3,742 $ 3,667 5 $ 1,570 $ 108 $ 167 $ 630 $ $ 1,175 $ 159 $ 1,017 3 $ 3,071 $ 119 $ 2,003 $ 150 $ $ 4,288 $ 3,306 $ $ 3,956 $ 3,669 $ $ 1,704 $ 194 $ 150 $ 1,300 $ 60 1 $ 3,375 $ 157 $ 1,220 $ 1,828 $ 170 Source: Medicaid Claims Data As shown in the table, of the total 1,016,202 care months provided in SFY , no clear allocation concerning cost could be made to service categories HCBS and/or NFs. Enrollees in some combination with bucket 94 Prepaid LTC, constitute 97.3 percent of the care months provided. Similarly nursing facility accounts for only 10,463 service months out of the total care months, or slightly over 1.0 percent. Therefore, in answering research question 1b, Medicaid encounter data was used. The encounter data included 77,585 members in the LTC program 32 on average per month. The location of services and the claims generated by the LTC population are presented in Table III.4 and Table III.5, first by service category variable available, then by simply consolidating service types into the categories of nursing facility and HCBS. 30 Buckets used are bucket 6 Skilled Nursing Care, 7 ICF I Services and 8 ICF II Services, combined in bucket NH (Nursing Home), bucket 33 Hospice Services, 35 Home and Community Based Aging, 36 Home and Community Based DS, 37 Aids Waiver Services, 42 Personal Care Services, 61 Adult Congregate Living Facility, 66 Nursing Home Diversion Waiver, 67 Brain and Spinal Cord Injury Waiver, 71 Assistive Care Services, 73 Channeling, 79 Alzheimer s Waiver, 81 Adult Day Care Waiver, 89 PACE, 93 LTC Admin Fee and 94 Prepaid LTC. Not mentioned buckets in Table II.3 indicates a zero sum or PMPM members with negative total monthly amounts are omitted, including four with an erroneous entry in buckets 93 and or 91.6% of the 84,703 members 115 Final Report (Fiscal Year ) June 30, 2017

116 Table III.4. Average Service Utilization by Location of Service for SFY (in SFY $). 33 Service Type Service Utilization by Location Months by SFM Amounts by SFM (in nominal $)** Average Total by SFM (in nominal $)** 10 Nursing Facility 36,813 $ 217,037,813 $ 5, Assisted Care Services 4,969 $ 7,314,592 $ 1, Hospice 2,723 $ 15,810,867 $ 5, Home Health Agency 1,789 $ 7,991,017 $ 4, HCBS Waiver 10,488 $ 17,109,882 $ 1, Aging and Adult 1 $ 2,033 $ 2, DME 428 $ 955,617 $ 2, Other* 20,375 $ 3,852,529 $ Total 77,586 $ 270,074,351 $ 3, * Remainder of smaller service categories in the Encounter data base, and cost of clients who received services in more than one main service category in the same month. ** SFM is State Fiscal Month and nominal $ are values adjusted for inflation Source: Plan Encounter Data Table III.5. Average Service Utilization by Location of Service for SFY (in SFY $). 34 Service Type Service Utilization by Location Months by SFM Amounts by SFM (in nominal $)*** Average Total by SFM (in nominal $)*** Nursing Facility (10+14) 39,536 $ 232,848,681 $ 5, HCBS ( ) 15,458 $ 24,426,507 $ 1, Other** (65+90+Other*) 22,592 $ 12,799,163 $ Total 77,586 $ 270,074,351 $ 3, ** Remainder of smaller service categories in the Encounter data base, and cost of clients who received services in more than one main service category in the same month. *** SFM is State Fiscal Month and nominal $ are values adjusted for inflation Source: Plan Encounter Data Figure III.2 on the following page shows the trend in average individual encounter costs during the LTC program period. 33 Nursing Facility in Table III.5 is taken to be the sum of Nursing Home (10) and Hospice (14) in Table III.4. HCBS in Table III.5 is taken to be the sum of Assisted Care Services (14), HCBS Waiver (67) and Aging and Adult Waivers (80) in Table III.4. Other** in Table III.5 are the remaining categories in Table III.4. Both Other categories used constitute clients receiving services from some smaller not specified service types (remainder categories) or clients who received services attributable to more than one mentioned specific main service category within the same month. Although an algorithm was set up by the research team to allocate clients to one main service category per month, based on a number of days per month, days per month alone proved to be an insufficient criteria to properly allocate specific dated cost associated with a service category. The setup in both Tables III.4 and III.5 as well as in Figures III.3 and III.4 is kept as straightforward or clean (to main service categories mentioned) as possible, in order to get a perception on categorical services cost (not blurred by dual or mixed services received). The aim here is not to provide an overall average cost per client, but an as clear as possible average cost per client on the two main categories; Nursing Facility and HCBS. 34 Ibid Final Report (Fiscal Year ) June 30, 2017

117 Figure III.2. Per Member Per Month Nominal Encounter Values, SFY (in $) $4,000 ENCOUNTER CLAIMS PMPM NOMINAL $3,750 $3,500 $3,250 $3,000 Jul-14 Aug-14 Sep-14 Oct-14 Nov-14 Dec-14 Jan-15 Feb-15 Mar-15 Apr-15 May-15 May- 15 Jun-15 PMPM Source: Plan Encounter Data As can be taken from the monthly encounter values, there is a decline which, based on regression values, amounts to approximately percent annually. The evaluation team computed a Medical Cost Ratio (MCR) on the basis of the following formula: Medical Cost Ratio = Average Monthly Inflated Costs for LTC Services Provided Pre-LTC Program / Average Monthly Costs for LTC Services Provided Under the LTC Program PMPM claims (as taken from the provided Medicaid Claims Database) averaged slightly higher than the derived encounter costs reported by the LTC plans (as taken from the plan encounter databases). The Medicaid claims were $3, on average (from Table III.1), while the encounter costs averaged $3, PMPM (from Table III.5). This results in a Medical Cost Ratio (MCR) of This means that for SFY , the finding is close to cost-neutral for the LTC program. The relative distribution between care in an HCBS setting (at home, in an ALF, or in an AFCH) and care in a nursing facility or hospice is displayed in Figures III.3 and III The Medical Cost Ratio (MCR) used here is defined as the average per-individual per-month encounter costs relative to per-individual per-month Medicaid claims (x 100%). Hence it is duly noted that the MCR definition used here is not the same as the more generally used Medical Loss Ratio, as Plan revenues are not part of the equation. 117 Final Report (Fiscal Year ) June 30, 2017

118 Figure III.3. Absolute Distribution of LTC Program Enrollees by Location, SFY Data: Plan Encounter Data Figure III.4. Relative Distribution of LTC Program Enrollees by Location, SFY Data: Plan Encounter Data Conclusions If one considers the LTC program months July 2014 to June 2015, the results presented in the comparative analyses show that the cost of the LTC program introduced in August 2013 has been cost neutral as measured by the comparison of the enrollee s average per member per month costs for long-term care services in the preprogram period to the average PMPM costs after the program s full implementation. The average monthly nominal LTC program cost was estimated at $3, versus $3, in the Pre LTC period, or an increase of $ However, when corrected for inflation, the average monthly LTC program costs decreased by $0.03. Based on the interrupted time series (ITS), comparing expected values in SFY under two scenarios (LTC program versus unchanged), the cost savings per month is calculated at $31.20 per member, or $2.6 million per 118 Final Report (Fiscal Year ) June 30, 2017

119 month in total. However, given the short timeframe under consideration, consecutive years of added SFY data will be needed to verify the consistency of these preliminary findings. The relative distribution between care in an HCBS setting (at home, in an ALF, or in an AFCH) and care in a nursing facility or hospice could not be determined satisfactorily. Yet, the cost for individuals served in nursing facilities or hospice is roughly 3.5 times more expensive. The continuation of downward MCR trend will depend upon continued rebalancing from more expensive nursing facilities and hospice to care in HCBS settings. In summary, results reported herein reflect Long-term Care program performance across relatively distinct periods: Pre-LTC and LTC. Both reflect total costs of the Long-term Care program for July 2014 to June For full implementation under the SMMC LTC period, however, based on the comparative analyses the Long-term Care program showed a nominal cost increase of $0.89. Corrected for inflation between the two periods, the average real monthly claim amount decrease is $0.03 per individual (cost-neutral). Based on the interrupted time series there is a cost savings per month of $31.20 per member, to a total of $2.6 million per month, constituting a marked change in both level and direction of cost. The MCR is, in the LTC program period being considered, under downward influence. Evaluation Limitations The ideal research design for assessing the impact of policy change would involve random assignment to the intervention (LTC) and comparison (Pre-LTC) group. It is a well-known limitation in this line of work that random assignment is rarely possible due to practical and, more importantly, ethical concerns. When this approach is not possible, program evaluators rely on statistical techniques to mimic experimental designs, the gold standard for permitting causal claims about programmatic effectiveness. However, due to time constraints, the evaluation team was unable to run any models to impose necessary statistical controls, which would mitigate barriers to measuring the true effects of the LTC program on enrollees access to and quality of care. 36 Given this limitation, the evaluation team cautions readers not to draw any firm conclusions about the LTC program based on this report and encourages them to consider the analysis herein as a baseline overview of the program. Last year s evaluation report assessed a non-representative transitional year, so it did not provide an accurate baseline for the LTC period. This year s report only summarizes head-to-head descriptive comparisons without any controls for factors that may inhibit accurately measuring true differences between the LTC and Pre-LTC period. Hence, as an assessment of the first complete program year that only includes descriptive results, the evaluation team considers it a baseline against which to compare future program years. Starting in next year s report and beyond, the evaluation team intends to employ statistical models that analyze trends with controls for enrollee risk scores, demographics, etc. The evaluation team has already taken the preliminary steps required to employ appropriate statistical techniques, such as longitudinal data analysis, the results of which will be included in future evaluation reports. 36 Because the cost-effectiveness evaluation pertains to overall costs between the two-time periods, these concerns do not apply to that component of the report. 119 Final Report (Fiscal Year ) June 30, 2017

120 The models will impose statistical controls for differences between the Pre-LTC and LTC population that confound measuring the impact of the program. Most notably, analysts will risk adjust these models, controlling for any changes to enrollees health statuses between the two-time periods. Risk adjustment, including controls for changes in enrollee frailty levels, is necessary if the population of Florida Medicaid enrollees receiving long-term care services has become increasingly ill or frail over time. Thus, the evaluation team anticipates that future evaluation reports should allow stakeholders to draw more definitive, i.e., more valid, conclusions about the program. Lastly, some plans encounter data limitations are jeopardizing the validity of the results due to missing data in reported encounter record fields. The evaluation team hopes that further validation of plan data submissions will improve the quality of the encounter records in future years. One area of particular concern is case management reporting. The evaluation team was unable to identify expected instances of case management in the LTC encounter data. While it could be that the absence of case management records in the utilization data reflects a true absence of case management service delivery, the evaluation team suspects that this absence reflects a data limitation/quality issue. The evaluation team will work with the Agency to resolve this issue, as it is considered a major barrier to assessing compliance with case management requirements. The evaluation team also believes that access to individual recipient plans of care is essential to fully understanding realized access to care, as all utilization in the HCBS population is driven by those plans. 120 Final Report (Fiscal Year ) June 30, 2017

121 Appendix 1 Access to Care Methodology Study Design The access to care evaluation used quantitative methods to produce comparisons of descriptive measures between the Pre-LTC and LTC periods, namely, simple head-to-head comparisons of the same measure calculated for each time period. This evaluation is analogous to an observational study, whereby the evaluation team annually assesses the impact of the LTC program as it unfolds over time. In true experimental research designs, the investigator randomly assigns subjects to either an intervention or a control group. The intervention (experiment) is conducted, and data is collected prospectively. Hence, the intervention group in an observational study is typically already determined before the evaluation begins and the data is collected, which is often for administrative purposes rather than for the purpose of evaluation. The intervention group of interest in this evaluation is the LTC program population age 18 years or older who received LTC services under the program during state fiscal year Hence, LTC enrollees were followed from their first through their last day of eligibility between July 1, 2014 and June 30, Observational studies are very common for policy analysis where the principal goal is to determine whether a new policy (the intervention) had a positive, negative, or neutral effect on measures of interest to the policymakers, as compared with a comparison group that was not exposed to or, in this case, existed prior to the new policy intervention. The comparison group in this evaluation is Medicaid recipients age 18 and older who were enrolled in the legacy Assisted Living, Aged and Disabled, Channeling or Nursing Home Diversion Waivers, received services under the Frail Elder Option, or resided in a NF with the Florida Medicaid agency as the primary payer between July 1, 2012 and June 30, Therefore, Pre-LTC Medicaid recipients were followed for at least one month and at most 12 months, depending on when they became eligible for legacy HCBS waiver or NF services. Observational designs encounter the complex task of carefully choosing a comparison group that is a good match to the policy intervention group on many relevant factors and/or using statistical techniques to adjust for observed differences. The validity of observational studies is therefore very sensitive to the completeness and quality of the available data as well as the length of time the policy intervention and comparison population groups were observed. This quantitative evaluation uses retrospective data collected for administrative purposes by the Agency, DOEA, and the seven managed care plans during SFY1213 and SFY1415. Data Sources The quantitative analysis for each access to care RQ relied on the measures and data sources outlined in Appendix 1 Table Final Report (Fiscal Year ) June 30, 2017

122 RQ # Appendix 1 Table 1. Access Research Questions, Associated Measures, and Data Sources Research Question Measures Data Sources Have there been changes in the accessibility of services for enrollees compared to the previous LTC programs and over time? How has the population being served in the LTC program shifted (characteristics of the population and service utilization) between nursing facilities and HCBS over time? What LTC plan strategies are impacting these shifts? Do plans offer additional (expanded) benefits and ways to access services, including a Participant Directed Option (PDO), and to what extend to enrollees use these services? i. Provider network size: Number of providers offering each LTC service; facility counts by county ii. Service utilization (HCBS only): LTC versus Pre-LTC proportion of enrollees receiving each service iii. Time to first service delivery (HCBS only): Comparison of LTC with Pre-LTC mean number of days between program enrollment and first service receipt i. Location of Care: LTC versus Pre-LTC monthly enrollee residency rates in nursing facility and HCBS settings ii. Transition rates and transition into the community success: Comparison of the proportion of enrollees and success rates of transition into the community for enrollees who switched residential settings in the LTC versus Pre-LTC period iii. Nursing facility risk: Comparison of scores for enrollees residing in HCBS settings in the LTC versus Pre-LTC period iv. CDPS risk scores: Mean score by residential setting in the LTC versus Pre-LTC period v. Contracted bed counts: Number of facility beds contracted by each plan and within each region List of expanded benefits offered by each plan i. PDO enrollment rates: PDO enrollment over time and comparison of the proportion of eligible enrollees by plan who enrolled in the PDO ii. Length of PDO enrollment: Comparison of PDO enrollment spans/duration by plan i. Provider Network Verification files ii. FSU created residency file, Medicaid claims, NHD encounter data, LTC encounter data iii. Eligibility data, FSU created residency file, Medicaid claims, LTC encounter data i. FSU residency file created from eligibility data, MDS data, Medicaid claims, and FL Center inpatient data ii. FSU created residency file iii. FSU created residency file, DOEA CIRTS assessment data, MDS data, Finder file iv. FSU created residency file, CDPS risk score variables from the eligibility data v. PNV files, facility license files LTC plan contracts i. Plan PDO Roster Reports ii. Plan PDO Roster Reports 122 Final Report (Fiscal Year ) June 30, 2017

123 RQ # Research Question Measures Data Sources iii. PDO service utilization: Rates of PDO service utilization by service category iii. LTC encounter data, Plan PDO Rosters 4 Are there disparities by racial and ethnic groups in enrollees placements in certain settings and utilization of services (LTC period only)? i. Location of care disparities: residency rates in nursing facility and HCBS settings, stratified by race/ethnicity ii. Service utilization disparities (HCBS only): Proportion of enrollees receiving each service, stratified by race/ethnicity iii. Time to first service delivery disparities (HCBS only): Mean number of days between program enrollment and first service receipt, stratified by race/ethnicity iv. Facility accessibility: GIS Maps of enrollees home addresses and nearest nursing facilities, stratified by race/ethnicity i. FSU created residency file, Finder file ii. FSU created residency file, Medicaid claims, LTC encounter data, Finder file iii. Eligibility data, Medicaid claims, LTC encounter data, Finder file iv. Geocoded facility list, geocoded enrollee addresses, Finder file Data Sources for the Quantitative Evaluation The following section provides additional details on the data sources listed in Appendix 1 Table Pre-LTC Administrative Files Collected by the Agency a. Pre-LTC enrollment and program eligibility data: The Agency collects this information at the time of application to Medicaid and continuously updates it over time. It is the primary source of information on enrollment spans and recipient demographic information for the Pre-LTC period. b. Pre-LTC claims: Providers filed these records for reimbursement from the Agency for long-term care services provided during the Pre-LTC period. Claims data, critical to Pre-LTC service utilization metric calculations, includes information on the types of services provided, units of service, place of service delivery, provider specialty, provider type, diagnoses, beginning and end dates of service, and payments made to the provider. The financial buckets in this data also provide a monthly index of legacy waiver program recipients. c. Nursing Home Diversion encounter data: NHD encounters are the other source critical to calculating Pre-LTC utilization metrics. NHD program members constituted the greatest proportion of the Agency s Pre-LTC population of Medicaid recipients with long-term care needs. Service providers for recipients enrolled under this program were reimbursed via a capitated payment system, but providers also submitted encounter records to the Agency. These records contain a limited number of variables; nevertheless, those that are available provide enough information to generate Pre-LTC access metrics, including type of service rendered, service date, units of service, and provider codes. 123 Final Report (Fiscal Year ) June 30, 2017

124 2. Agency Generated Data and Reports a. Finder File: This file consists of all Medicaid recipients with at least one month of eligibility in the Pre-LTC and/or LTC period. b. Agency Quarterly SMMC LTC Reports: Public reports of SMMC LTC statistics and trends. 3. LTC Encounter Data: These data are plan-specific records of enrollee encounters with the contracted network of long-term care service providers. Encounter records are first submitted to the plans by network providers for reimbursement purposes. In some cases, network providers may be paid via fixed or capitated rates instead of on a fee-for-service basis. Plans then submit all encounter records to the Agency s fiscal agent, Hewlett-Packard Enterprise. Plans must submit complete, accurate, and timely encounter data to the fiscal agent, as defined in their contract and in accordance with generally accepted industry best practices. The plan is held responsible for errors or noncompliance resulting from its own actions or the actions of an agent authorized to act on its behalf. Plan administration must verify all encounter information it submits to the Agency in accordance with 42 CFR Accordingly, the Agency requires an attestation file with every encounter data submission, which is due within two business days of the file upload. Encounter records contain many variables analogous, and in some cases identical, to those in Medicaid claims data, including the type of service rendered, service date, service location, units of service, diagnosis codes, and provider information. Long-term care encounter data is central to calculating access to care metrics for the LTC period. Claims from the LTC period: The evaluation team uses Medicaid claims from the LTC period for calculating service utilization metrics for AEC through September 2014, where after the plan switched over to a capitated system of payment. Monthly capitated payments from the Agency to the plans for each enrolled member are also contained in this data. 4. SMMC LTC enrollment and eligibility data: These records contain plan-specific enrollment spans and enrollee demographic information collected by the Agency. They are updated on a continuous basis. 5. Plan Specific Data, Documents, and Required Reports from the Report Guide Dated April 1, a. Plan Specific Provider Network Verification files: These reports contain a list of all LTC provider types, specialties, and locations contracted by each LTC plan. The files standardized by the Agency include a Provider Group/Hospital (PG) file with information for unique individual providers, group providers, and hospitals. The Service Location (SL) file provides information about all locations for individual providers and group network providers. Medical and LTC type service providers are included in these files. Managed care plans are required to update these files on a weekly basis whenever a network provider leaves or joins the LTC plan s network. b. Enrollee Roster and Facility Residence Report: The purpose of this report is to provide information on the current physical location of each enrollee, which may be used for disaster recovery planning and relief. This report is also designed to track individuals who are transitioning 124 Final Report (Fiscal Year ) June 30, 2017

125 between settings (e.g., NF to community and vice versa). The report is due monthly, within 15 calendar days after the beginning of the reporting month. While the evaluation team does not use the report for analytic purposes, analysts intend to cross-check this source with the FSU created residency file to provide the Agency with an assessment of the scope of these reports reliability, especially for January c. Participant Directed Option (PDO) Roster Report: These reports identify which community-based enrollees opted into the PDO, their enrollment (and disenrollment) date, and binary indicators that signify which of the five allowable services the participant opted to receive under the program. 6. Assessments a. Minimum Data Set (MDS): The Minimum Data Set is a major component of the federally mandated process for clinical assessment of all residents in Medicaid/-care certified nursing facilities. It is administered upon admission and discharge as well as regularly throughout the duration of a stay. The Care Area Assessment under the MDS process, which includes the decision process for determination of functional status, is the basis of a client s care plan. Because the MDS was also administered during the Pre-LTC period, it is a principal resource for gauging location of and changes in care between the Pre-LTC and LTC program periods. b. 701B (from DOEA CIRTS database): The Florida Department of Elder Affairs (DOEA) 701B comprehensive assessment is administered as part of CARES, a federally mandated preadmission screening process for all Medicaid recipients who apply for nursing facility services. DOEA uses the assessment to determine appropriate placement based on an individual s functional capabilities and required levels of care. Additionally, for enrollees who transition from nursing facilities to receive LTC services in HCBS settings, case managers conduct the 701B within 12 months of transition into the community or the last date of the previous location of care, depending on how long the client resided in the nursing facility. Case managers use the results of this assessment to update the client s plan of care. It is administered every year thereafter to enrollees who remain in HCBS settings. Because the 701B form was also used for client assessment under Pre-LTC programs, it is a principal resource for gauging changes in care between the Pre-LTC and LTC program periods. 7. LTC Model Contract and Individual LTC plan contracts: These comprehensive documents define the managed care plans responsibilities to enrollees, network providers, and the Agency within the context of the SMMC LTC program. The contractual obligations outlined in these documents include specifying LTC benefits, approved LTC provider types and specialties, plan reporting requirements, and capitation payments received by the plans. 8. Florida Center Inpatient Data: The Agency s Florida Center for Health Information and Transparency collects inpatient claims data from acute care hospitals as well as short- and long-term psychiatric facilities. Because LTC enrollees frequently experience acute events, analysts use this dataset to track 125 Final Report (Fiscal Year ) June 30, 2017

126 enrollees when they are discharged from nursing facilities. This process is instrumental for determining whether an enrollees transitions into HCBS settings or experiences an acute care even when he/she is discharged from a NF. The Florida Center collects all claims regardless of payer; hence, all records are available for recipients with single or dual eligibility. 9. Geocoded enrollee addresses and long-term care facility locations Enrollee Residency File The access to care team relied on an internally created enrollee residency file entirely for RQ2 metrics i. & ii. and RQ4 metric i., as well as partially for RQ1 metrics ii. & iii., RQ3 PDO metrics, and RQ4 metrics ii. & iii. Determining an enrollee s location during the Pre-LTC period is problematic because there is no equivalent of the Enrollee Roster and Facility Residence Report for the Pre-LTC period. While determining an enrollee s location during the LTC period is problematic, too, since this file has known reliability issues, especially for certain months (e.g., January 2015). Additionally, even determining enrollee eligibility in the Pre-LTC period for those who resided in nursing facilities is a complicated process, as there is no equivalent of the LTC program eligibility indicator for the Pre-LTC period. Fortunately, it is possible to identify legacy HCBS waiver enrollees during the Pre-LTC period with the Bucket variable in the claims data. For these reasons the access to care team developed a methodology for determining enrollee location in both time periods, largely based on MDS assessment records and the Medicaid eligibility files. This process includes an eligibility determination for Medicaid enrollees who resided in NFs at any point during the Pre-LTC period. The analyst has made every effort to provide clarity but concision in outlining this complex process below in Appendix 1 Table 2. Data Source MDS data Appendix 1 Table 2. Enrollee Residence Location Determination Methodology Step/ Purpose Step 1: Identify which enrollees resided in a nursing facility when Medicaid was the primary payer Process 37 1) Search for all MDS records within the relevant time period (SFY1213 or 1415) for enrollees in the Finder File. 2) Bring in any records that occurred immediately before or after (+/- 93 days*) the corresponding Pre-LTC/LTC period. 3) Count the number of days between each assessment record and the next record. 4) Exclude any days between a discharge and subsequent entry or after a final discharge.** 5) If the number of days between assessments is less than or equal to 93 days, then create a NF indicator for each day. 6) If the number of days between assessments is greater than 93 days, create a NF indicator for the target date of the earlier record, but not for the subsequent days until the next record. 37 The Pre-LTC period currently includes some nursing facility days that were actually post-acute care episodes where Medicare was the primary payer. The analyst will resolve this omission and adjust all metrics/ analysis accordingly for inclusion in the final report. 126 Final Report (Fiscal Year ) June 30, 2017

127 Data Source LTC eligibility data Pre-LTC claims data MDS data from Step 1 Files from Step 3 & 4 Files from Step 1 & 2 FL Center Inpatient data FL Center Inpatient data File from Step 7 Step/ Purpose Step 2: Identify months with LTC program enrollment Step 3: Identify months with Pre-LTC legacy HCBS waiver enrollment Step 4: Identify Pre-LTC months with nursing facility residency Step 5: Create preliminary Pre-LTC residency file Step 6: Create preliminary LTC residency file Step 7: Carry forward the NF and/or HCBS waiver indicator throughout an inpatient stay, LTC period only Step 8: Fill in any gaps in the Pre-LTC period for NF residents with inpatient stays, and carry forward the NF and/or HCBS waiver indicator Step 9: Process 37 7) If only 1 assessment was found that is not a discharge (throughout the entire window), then create a single indicator for that day. 1) For enrollees in the Finder File, search for any month in the LTC eligibility data where the SFY is and the LTC enrolled variable contains an N or W indicator. 2) Create a LTC program enrollment indicator for all days in an eligible month. 1) For enrollees in the Finder File, search for HCBS waiver months where the Bucket = 35, 61, 66, 73 (A/DA, ALW, NHD, Channeling waivers, respectively) or months of Frail Elder Option program enrollment where the provider ID is or ) Create a day-level HCBS waiver indicator for all days in the month where waiver enrollment was detected. 1) Limit the file from Step 1 to person-days found in months where the enrollee was a Medicaid recipient in SFY1213 (based on the Pre-LTC eligibility data). 1) Merge the files from Step 3 and 4 by enrollee Medicaid GUID and day. 1) Merge the MDS data from Step 1 with the file output in Step 2 by Medicaid GUID and day, limited to months with LTC program enrollment. 1) Identify any days on and between an inpatient admission/ discharge and create an inpatient stay indicator for each day within SFY ) Merge this file with the preliminary LTC residency file from Step 6 by Medicaid GUID and day. 3) Carry forward the most recently observed nursing facility/hcbs waiver indicator throughout the inpatient stay. 1) Identify any days on and between an inpatient admission/ discharge and create an inpatient stay indicator for each day within SFY ) Vertically stack this file with the Pre-LTC residency file from Step 5 3) Deduplicate the file by Medicaid GUID and day (sum down all variables after grouping by person-day) 4) Eliminate any days that occurred before the first day of Pre-LTC service receipt 1) Set any day where a recipient was not determined to be located in a nursing facility as an HCBS day (an indicator for when a recipient resided in a home-community based setting) 127 Final Report (Fiscal Year ) June 30, 2017

128 Data Source File from Step 8 Step/ Purpose Create final LTC enrollee residence location file Step 10: Create final Pre-LTC enrollee residence location file Process 37 2) Create an indicator that identifies all records in this file as from the LTC period 1) Set any day where a recipient was not determined to be located in a nursing facility as an HCBS day 2) Create an indicator that identifies all records in this file as from the Pre-LTC period * Incorporates CMS s 3 day grace period. ** For coding purposes only, an enrollee death is considered equivalent to a discharge. The files output using the methodology in Appendix 1 Table 2 provide a day-level indicator of where an enrollee resided within any Pre-LTC or LTC eligible month. For the purposes of determining monthly residence location in this report, if a person switched residential settings within a given month, then that month is considered a nursing facility month when the number of days spent in a NF was greater than or equal to the number of days spent in an HCBS setting. Consequently, a month is considered a month with home and community-based residency for any enrollee when the number of HCBS days is greater than the number of days spent in a NF for that month. The analyst crosschecked this location file with nursing facility and HCBS records in the LTC encounter data, with the understanding that a few thousand LTC enrollees are missing encounter records in any given month, and found 95% concordance between the two sources. Time to First Service Delivery Appendix 1 Table 3 presents the different reasons for exclusion from RQ1 metric iii. for enrollees who entered into a Pre-LTC waiver program at some point during SFY1213 or entered into the LTC program at some point during SFY1415. Appendix 1 Table 3. Frequency of New Program Entrants in HCBS Settings Who Meet First Service Delivery Exclusion Criteria (when Case Management Services Excluded) Exclusions Pre-LTC (only those whose first Pre-LTC month was in an HCBS waiver program) LTC (only those whose first LTC month was in an HCBS setting) Has an encounter record after enrollment but no valid LTC service N/A 101 Has an encounter record prior to enrollment but not after N/A 31 Has no LTC encounter/claim* record in SFY1415 (or before enrollment) N/A 1,021 No claim found 504 N/A Enrolled in the Nursing Home Diversion program 5,273 Total** 5,777 1,153 *Claim records used for AEC in July and August of 2014 **Numbers cannot be directly compared Sources: AHCA s LTC service category crosswalk, FSU created enrollee residency file (see above), Medicaid FFS claims, LTC encounter records, NHD encounter records 128 Final Report (Fiscal Year ) June 30, 2017

129 Nursing Facility Residency Risk Model Appendix I Table 4 presents the results of the logistic regression model the analyst developed in order to derive the predicted NF residency probabilities for RQ2 metric iii. The evaluation team originally intended to use ADL ratings from the MDS and CIRTS assessment data to control for functional status in this model. However, too many enrollees receiving services in community-based settings in the Pre-LTC period were missing an assessment, which led to inaccurate estimates of residency probabilities in the Pre-LTC period. This issue persists in the LTC period to a lesser degree, as it appears CIRTS assessment reporting has significantly improved since LTC program implementation. See Appendix 1 Table 5 for a breakdown of missing assessment rates by residential setting in both time periods. These results are reported for enrollees who experienced at least 60 days of enrollment in their respective evaluation periods. Appendix 1 Table 4. Logistic Regression Model Estimates for Nursing Facility Residency Risk in the Pre-LTC and LTC Period Parameter Estimate Base Model Standard Error Wald Chi- Square P > Chi-Sq Estimate Interaction Model Standard Error Wald Chi- Square P > Chi-Sq Intercept Female < <.0001 Male Black < <.0001 Hispanic < <.0001 Other race < <.0001 White Age < <.0001 Age < Age <.0001 Age < <.0001 Age < <.0001 Age < <.0001 Age 85+ Region < <.0001 Region < <.0001 Region < <.0001 Region < <.0001 Region < <.0001 Region < <.0001 Region < <.0001 Region < <.0001 Region < <.0001 Region < <.0001 Region 11 Died < <.0001 No death Inpatient event < <.001 No inpatient event LTC Period < <.0001 Pre-LTC Period LTC*Female <.0001 LTC*Male LTC*Black LTC*Hispanic LTC*Other race Final Report (Fiscal Year ) June 30, 2017

130 Parameter Estimate Base Model Standard Error Wald Chi- Square P > Chi-Sq Estimate Interaction Model Standard Error Wald Chi- Square P > Chi-Sq LTC*White LTC*Age <.0001 LTC*Age <.0001 LTC*Age <.0001 LTC*Age <.0001 LTC*Age LTC*Age LTC*Age 85+ LTC*Died LTC*No death LTC*Inpatient event LTC*No inpatient event Appendix 1 Table 5. Reported and Missing Assessment Rates by Location of Care Setting in the Pre-LTC and LTC Period NF * HCBS Evaluation Period Reported Missing Percentage Missing Reported Missing Percentage Missing Pre-LTC 63, % 22,123 20, % LTC 54, % 36,183 11, % *At least 60 consecutive days in a nursing facility Sources: FSU created enrollee residency file, MDS 3.0 data, DOEA CIRTS (Assess 1, 701B, v. 2008; Assess 2, 701B, v. 2013) Analytic Methods for the Quantitative Evaluation The evaluation team mostly utilized univariate descriptive measures and univariate measures stratified by time period to analyze the quantitative data. The evaluation team also developed one multivariable logistic regression model to respond to RQ2. Preliminary Risk Adjustment Categories The evaluation team developed a categorical risk adjustment method using ADL functional status ratings from the CIRTS and MDS assessment data. This method takes into account early loss (bathing and dressing), middle loss (toileting and transferring), and late loss (eating) ADLs, which researchers have determined are strongly predictive of resource utilization in the RUG-III system. 38 FSU analysts may use these categories to risk adjust statistical models, allowing for greater comparability between the Pre-LTC and LTC period. An ADL loss is assigned for a given activity when an enrollee or assessor reports the need for extensive or total assistance in a CIRTS or MDS assessment field for bathing, dressing, eating, toileting, or transferring needs. Walking was excluded because this ADL is not consistently reported and/or observed in assessments that take place in nursing facility settings. 38 Morris JN, Fries BE, Morris SA. Scaling ADLs within the MDS. J Gerontol A Biol Sci Med Sci. 1999; 54(11):M Final Report (Fiscal Year ) June 30, 2017

131 On a brief methodological note, to complete these categorizations the analyst selected the most recent assessment for enrollees with multiple assessments reported within an evaluation year. For those with two assessments reported on the same day, the analyst selected the assessment that corresponds to the location of longest residency for a given evaluation year (MDS for NF, CIRTS for HCBS settings). Nursing facility location of care was assigned to enrollees who experienced 60 consecutive days of NF residency; otherwise, a HCBS status was assigned. Accordingly, enrollees are only assigned a risk category when they have at least 60 days of observation in a given evaluation period. There are clear patterns to the risk categories. Enrollees in HCBS settings are much less likely to need extensive or total assistance with any of the five ADLS, while enrollees in NFs are much more likely to need extensive or total assistance with any of the five ADLS. Bathing is likely the first ADL for which an enrollee requires help, while eating is likely the last ADL for which an enrollee requires help. The reader should note that CIRTS assessments are reported more frequently in the LTC period than in the Pre-LTC period (see Appendix 1 Table 5). It may be the case that these assessments were systematically missing in the Pre-LTC period for enrollees who were the least functionally impaired during that time period (note the proportion in the None category for enrollees in HCBS settings in the LTC period versus the Pre-LTC period). The results are presented in graphical form in Appendix 1 Figures 1-4. Appendix 1 Figure 1. Proportion of Enrollees with HCBS Residency Assigned to Each Risk Category in the Pre-LTC Period 70% 60% 50% 48.2% 40% 30% 20% 19.0% 13.4% 10% 0% 5.9% 4.7% None Bathing Bathing and Dressing 1.0% 2.7% Bathing and Toileting Bathing, Dressing, and Toileting 0.5% 1.3% Bathing, Dressing, and Transferring Bathing, Toileting, and Transferring Bathing, Dressing, Toileting, and Transferring All Five 3.3% Other Combination 131 Final Report (Fiscal Year ) June 30, 2017

132 Appendix 1 Figure 2. Proportion of Enrollees with NF Residency Assigned to Each Risk Category in the Pre-LTC Period 70% 60% 50% 40% 30% 38.1% 31.9% 20% 10% 0% 9.6% 8.7% None Bathing Bathing and Dressing 2.6% 1.4% Bathing and Toileting 3.5% Bathing, Dressing, and Toileting 0.5% 1.5% Bathing, Dressing, and Transferring Bathing, Toileting, and Transferring Bathing, Dressing, Toileting, and Transferring All Five 2.2% Other Combination Appendix 1 Figure 3. Proportion of Enrollees with HCBS Residency Assigned to Each Risk Category in the LTC Period 70% 60% 64.4% 50% 40% 30% 20% 10% 0% 4.5% 3.8% None Bathing Bathing and Dressing 0.6% 2.1% 0.7% 0.6% Bathing and Toileting Bathing, Dressing, and Toileting Bathing, Dressing, and Transferring Bathing, Toileting, and Transferring 10.4% 9.7% Bathing, Dressing, Toileting, and Transferring All Five 3.2% Other Combination 132 Final Report (Fiscal Year ) June 30, 2017

133 Appendix 1 Figure 4. Proportion of Enrollees with NF Residency Assigned to Each Risk Category in the LTC Period 70% 60% 50% 40% 30% 39.5% 30.7% 20% 10% 0% 9.4% 8.9% None Bathing Bathing and Dressing 2.5% 1.5% Bathing and Toileting 3.5% Bathing, Dressing, and Toileting 0.5% 1.4% Bathing, Dressing, and Transferring Bathing, Toileting, and Transferring Bathing, Dressing, Toileting, and Transferring All Five 2.1% Other Combination GIS Methods The geocoded long-term care enrollee and provider data was imported into the ArcCatalog GIS data management software. 39 A feature class was created from the latitude and longitude fields for both enrollees and providers using the NAD 1983 (2001) coordinate system. The feature classes were then exported and projected into the NAD 83 Florida GDL Albers projection to minimize distortion. 40 The resulting shapefiles were used for the analysis. Analysis of the proximity of the enrollees home addresses to Medicaid certified nursing facilities was completed with the generate near table tool in the ArcMap 10.4 GIS software. 41 The analysis utilizing all enrollees revealed that 42,394 (of the 86,567 provided) were so close to facilities that they were likely residents of these facilities. Those enrollees were subsequently excluded. The remaining enrollees were then used for analysis of the proximity to nursing facilities. 39 ESRI ArcGIS Desktop: Release Redlands, CA: Environmental Systems Research Institute. 40 This map of the state of Florida may be found at 5D7DDC68EB04%7D&loggedIn=false. 41 ESRI ArcGIS Desktop: Release Redlands, CA: Environmental Systems Research Institute. 133 Final Report (Fiscal Year ) June 30, 2017

134 Appendix 2 Quality of Care Methodology Study Design for the Quantitative Evaluation Quantitative analysis of the RQs relied heavily on secondary analysis of available data from EQRO reports, relevant items from LTC enrollee surveys, and Agency-defined performance measures. DOEA 701B Comprehensive Assessment and MDS data were also examined. The evaluation team s strategy was also informed with separate reviews of the relevant published literature related to the use and limitations of consumer satisfaction surveys, such as the Consumer Assessment of Health Care Provider Systems (CAHPS), LTC enrollee surveys, and quality of care (QoC) measures, such as the Healthcare Effectiveness Data and Information Set (HEDIS) and Agency Defined Performance Measures. The evaluation team worked with the Agency to develop a comprehensive set of assessment measures related to LTC service quality. Data Sources for the Quantitative Evaluation The quantitative analysis for each quality of care RQ relied on data measures and data sources listed in Appendix 2 Table 1.The section following the table provides additional details on the data sources identified in the table. Appendix 2 Table 1. Quality Research Questions, Associated Measures and Data Sources Quality Research Questions, Measures and Data Sources Research Question Measure Data Source 1. Has the quality of services that enrollees receive improved, compared to the previous LTC programs and over time? Quality indicators, recipient clinical outcomes Pre-LTC and LTC program EQRO reports, enrollee satisfaction surveys, HEDIS and Agency-defined performance measures and Medicaid claims and encounter data, DOEA 701 B Comprehensive Assessment and MDS files 2. What have managed care plans done to improve quality of care? 3. How has timeliness of service delivery changed compared to the previous LTC programs and over time? 4. Are enrollees rights being addressed in accordance with the waiver and contract? Quality indicators, network capacity, performance improvement strategies State survey results; changes in number and nature of complaints Pre-LTC and LTC program Number and nature of complaints and grievances related to the quality of care, rates of enrollees satisfaction EQRO reports, monitoring reports, Plans policies and procedures LTC enrollee surveys, Agencydefined performance measures, Grievances and Appeals reports and enrollee complaint data LTC plan and state complaint, grievance and fair hearing appeal logs, reports and summaries and enrollee satisfaction surveys 134 Final Report (Fiscal Year ) June 30, 2017

135 Quality Research Questions, Measures and Data Sources Research Question Measure Data Source 5. How effective are the LTC plans complaint, grievance and appeals processes? Rates of reported incidents and complaints over time LTC plans grievance and appeal policies and procedures, LTC plan and state complaint, grievance and fair hearing appeal logs, reports and summaries 6. What are the rates of complaints, grievances and appeals in the LTC program and have these changed over time? 7. Is there sufficient monitoring and oversight of the program by the state? Rates of reported incidents and complaints Pre-LTC and LTC program Descriptive analysis of monitoring activities and comparison of deficiencies and Agency actions over time LTC plan and state complaint, grievance, fair hearing appeal logs and reports and summaries LTC plan level deficiency and monitoring results maintained by the Agency, EQRO reports, LTC reports Specific Data Sources for the Quantitative Evaluation Quantitative analysis relied heavily on two assessment instruments: 1) MDS for NF enrollees; and 2) 701B Comprehensive Assessment for enrollees receiving HCBS (ORS and Home). The MDS is a federally mandated assessment instrument used by NFs. These assessments are useful in measuring enrollee outcomes (e. g., depression, use of anti-psychotic medication, and pressure ulcers). The CIRTS database is comprised of data from the 701B Comprehensive Assessment developed by DOEA to determine eligibility and need for services. This measure provided similar client outcomes for enrollees not residing in NFs (e.g., quality of life, nutritional status, self-reported health). The CIRTS database is particularly rich in capturing psychosocial outcomes. Months of Medicaid eligibility came from Medicaid eligibility spans. Prior enrollment in managed care came from Medicaid managed care spans. These data sources were used for analyses in the Pre-LTC and LTC program time periods. For the period following the implementation of the LTC program, the quantitative analyses also included service utilization reports and encounter data when appropriate. The following data sources supported these analyses: 1. Case Management File Audit Report This report is submitted by the managed care plan on a quarterly basis, 30 days after the close of each quarter. Each file is an Excel workbook consisting of two worksheets, one with instructions for completing the template and the File Audit Report Template. A random sampling methodology should determine which enrollees are included. The purpose of the report is to determine the compliance totals for the Agency. This 135 Final Report (Fiscal Year ) June 30, 2017

136 is done by listing each enrollee in the sample by Medicaid ID, followed by responses to each question: Y (Yes), N (No), or N/A (Not Applicable). 2. Complaint, Grievance, and Appeals Report This report is submitted monthly by each of the 7 plans and is due to the Agency no later than 15 days following the end of the month. The purpose of the report is to track the numbers and types of complaints, grievances, and appeals by region and county and how long each remains open. 3. Enrollee Roster and Facility Residence Report This report is due monthly no later than 15 days following the close of the previous month. The purpose of this report is to provide the current physical location of each enrollee and also to track individuals who are transitioning between settings. 4. Enrollee Satisfaction Surveys The LTC plans were required to contract with an Agency-approved independent survey vendor to administer the Enrollee Satisfaction Survey. The contract with the LTC plans stipulated that the survey should be administered according to the National Committee for Quality Assurance (NCQA) mixed mode protocol (mail questionnaire and then try to reach non-respondents with telephone interview), to a random sample residing in the community (minimum survey sample of 1700 enrollees, and a target of 411 completed surveys). To be included in the survey sample, enrollees must have been enrolled in the LTC plan for at least six months with no more than a 1-month gap in enrollment. The plans were required to submit the survey results using an Excel file template with responses from each enrollee. Managed care plan data and information provided by the Agency and reviewed for this report also included copies of the managed care plan contracts with the Agency, EQRO summaries of Performance Improvement Project plans, and standard performance measures collected by the Agency and validated by the EQRO. Analytic Methods for the Quantitative Evaluation The evaluation team employed descriptive statistics, including measures of central tendency and variation. The analyses also included simple univariate and bivariate comparison of selected measures from appropriate data sources. Tests for statistical differences among defined groups (e.g., cohorts before and after implementation of managed care) were conducted using Chi-square and t-tests as appropriate to compare proportions and means. Study Design for the Qualitative Evaluation The purpose of the qualitative component is to add depth to and understanding of the underlying processes of the quantitative findings. In order to accomplish this purpose, the evaluation team utilized two models of qualitative inquiry: 1) content analysis (i.e., review of written material to determine the presence of codes or themes); and 2) semi-structured interviews (i.e., interviews including a limited number of over-arching, open-ended questions). In each of these methods of inquiry, codes (or themes) emerged from the data beyond those included in the initial 136 Final Report (Fiscal Year ) June 30, 2017

137 questions. All interview and focus group letters, scripts and informed consent forms were approved by the Agency and the Institutional Review Board of FSU. Data Sources for the Qualitative Evaluation Specific data sources include: Administrative personnel and subsamples of program recipients and care providers drawn from the universe available in the quantitative data base Choice counseling and enrollment broker materials LTC plans policies and procedures Enrollee Satisfaction Surveys Transcripts of guided interviews and focus groups Using content analysis, the evaluation team reviewed choice counseling and enrollment broker materials in order to understand the intent and focus of the program before initiating interviews. The interviewers used semistructured guides with questions and prompts based, in part, on the content analyses. The following sections provide additional details on interview sources of data. Enrollee Interviews Of the 1,274 enrollees in the sampling frame, 377 have been discharged, we were unable to contact 517, 210 were unable to participate due to health reasons, and 100 had died. Of the remaining 70 enrollees contacted, 31 enrollees agreed to be interviewed about their LTC services. Analytic Methods for the Qualitative Evaluation The evaluation team examined the interview transcripts for emerging themes and relevant codes were developed utilizing the constant comparative method. This method allowed coders to compare new information to codes identified earlier and to develop new codes if none existed for the current data. This process allowed for a structured and systematic data analysis method while optimizing the emergence of new codes to capture new ideas as they developed. Data Analysis Process The analytic process began with immersion in the data; that is, the evaluation team read the transcript multiple times to become familiar with the content and flow. The evaluation team then made notations (codes) for each small bit of data, a process called open coding. These codes were recorded in Atlas/ti as the initial code list. Atlas/ti also allowed for memoing; a process where the evaluation team was able to make and retain notations related to underlying themes during the coding process. 137 Final Report (Fiscal Year ) June 30, 2017

138 When the open coding process was complete, the evaluation team summarized the findings. The data were analyzed for both manifest and latent codes and themes. For example, a manifest code might include the establishment of a complaints hub to track all complaints. However, in addition to the obvious manifest codes, there were underlying concepts (latent codes), such as defining quality, that were less concrete. Strategies for Rigor A key element in establishing validity in qualitative research is triangulation (i.e., use of more than one data source or method of data collection). This portion of the study incorporated two methods of triangulation: data source triangulation and interdisciplinary triangulation. Further data source triangulation in future reports will include data from later focus groups and interviews comprised of enrollees and their family members. Data from the access to care qualitative component was integrated with data from the components related to quality of care and cost-effectiveness. Interview Guide We re going to start by asking you four fairly broad questions, and then ask you some specific questions about your services. There are no right or wrong answers. We want to understand from your point of view. 1. Describe the Medicaid Long-Term Care Services you receive in [name of ALF or SNF or in your home]? (Prompt with the following if no response for recipients living at home: Thinking back over the past two years, have you received any of the following services? LTC Service Examples (if needed) Companion Care Person who assists you with preparing your meals, helps with your laundry, or helps with shopping. Attendant or Nursing Care Home-delivered meals Homemaker services Trained person comes in to perform household activities such as prepare meals or do household chores. Hospice services Medication Administration Person who assists you in keeping track of and/or taking your medicine Medication Management Pharmacist or nurse reviews the medicines you take both prescriptions and over the counter. Nutrition Assessment Assessment and teaching to help you prepare and eat healthy meals. Occupational, physical, respiratory, or speech therapy Personal Care Person assists with eating, bathing, dressing, personal hygiene, and other daily living activities. 138 Final Report (Fiscal Year ) June 30, 2017

139 Personal Emergency Response System Non-emergency transportation Device or button you can use to summon help in an emergency. Transportation to and from services such as an adult day facility, a. Are you aware of any changes in the long-term care services you have received over the past three years? b. How have the changes affected the quality of the care you receive? c. How have these changes affected your access to care? 2. How has your overall quality of life changed since receiving Medicaid LTC services? a. Has your satisfaction with your situation improved? If yes, how? b. Has your standard of living improved? If yes, how? c. Has your overall ability to function improved? If yes, how? 3. Has the quality of care you receive improved? (Based on the enrollees response: What are some examples of how the quality of care has improved? Or What are some examples of how the quality of care has gotten worse?) 4. What do you see as the best part of the program? 5. If you could change one thing, what would it be? Now we re going to switch to more specific questions about your services. 6. What kind of choices were you given about your Plan (long-term service provider)? a. How did you choose a Plan? b. What is the name of your Plan? c. How comfortable do you feel with the Plan you have? 7. What kind of choices were you given about your plan of care? a. Are there any services that you need, but you are not able to get? b. How would you get services added to your plan of care? c. Are you being provided the services listed on your plan of care? d. How comfortable do you feel with your current plan of care? 8. How easy is it for you to get in touch with your service providers? How far do you have to travel? 139 Final Report (Fiscal Year ) June 30, 2017

140 9. Tell me about your case manager. a. How often do you see your case manager? Name? b. What is your relationship like? c. Does your case manager talk to you about your goals for your care? d. Is your case manager always respectful? e. Can you get in touch with your case manager when needed? f. How does your case manager help coordinate the care from other providers? 10. How often do you have someone there to help you? Who? a. How effective has your medication been in providing relief/preventing accidents/etc.? 11. Do you know if you have rights as an enrollee in the Long-Term Care program? Based on the enrollees response: a. If yes: Do you know what they are? If no: Prompt with brief description of rights. b. Have your rights been upheld/violated? 12. What would you do if you had any issues with your services? a. If you have had issues, how did you appeal a decision that was made? Was it/how was it resolved? 13. Is there anything you would like to add? 140 Final Report (Fiscal Year ) June 30, 2017

141 Appendix 3 Cost-effectiveness of Care Methodology Study Design and Data The evaluation team used data on LTC costs to create a set of measures for cost analyses and medical costratios. The data used for these measures include Medicaid claims and managed care service utilization and encounter data provided by the Agency. The time period included the months of LTC Program eligibility and Pre- LTC Medicaid enrollments (SFY through SFY ). 42 At the Agency s request, Pre-LTC cost calculations were made using LTC fee-for-service, Nursing Home Diversion (NHD) and the Frail Elder Option claims. These calculations were performed by taking the 86,869 unique individuals (on average per month) 43 and comparing them to claims for LTC services filed on 84,703 unique individuals (on average per month) during the period July 2014-June This resulted in generating in excess of two million LTC claims in the Pre-LTC program period. These claims were compared on a nominal basis. From these claims, the average monthly payment made on behalf of each individual was calculated. The evaluation team utilized all data provided by the Agency for the period July 2014-June 2015 to evaluate the research question posed on LTC Medicaid costs. The evaluation team also used the baseline data provided by the Agency for the Pre-LTC period, July 2011-June Methods The data available to the evaluation team indicates that on average 86,869 unique individuals had monthly claims over the period July June 2013, while there were on average 84,703 unique individuals per month for the period July 2014-June Cost calculations for the period July 2011-June 2013 are denoted Pre-LTC. LTC program costs for the enrollees were computed by taking amounts in the Medicaid LTC claims service buckets 44 and calculating the average monthly payment to the managed care plans for long-term care services incurred in the LTC program period. These LTC program costs and this population are denoted LTC. The determination of which Medicaid service buckets to use for cost calculations was made in consultation with the Agency. The intent was to capture all costs that were associated with the LTC program. The cost analysis takes into account the LTC program s programmatic change. Prior to the LTC program, the Agency paid Medicaid providers of LTC services primarily on a fee-for-service basis, with the exception of NHD and the Frail Elder Option. The LTC program implements a capitated payment system. 45 In order to determine 42 Excluding the timeframe needed to fully implement the LTC Program, or transition period, from July 2013 to February 2014, or effectively the whole of SFY ,258 and 87,481 on average monthly for SFY and SFY respectively. 44 Claims are assigned by Florida Medicaid into relevant service categories, referred to as buckets. For example, hospital inpatient services are grouped together in two buckets, one where Medicaid is the primary payer for the service and another where Medicaid pays for Medicare co-payments and deductibles (so-called cross-over payments) for the Medicaid dually-eligible population. 45 During the first year of the LTC program, there was one FFS PSN that became capitated on September 1, Final Report (Fiscal Year ) June 30, 2017

142 how the costs of the capitated LTC program compare to the costs of LTC services prior to the LTC program s implementation, the following calculation was performed: Cost-Ratio = Average Monthly Inflated Costs for LTC Services Provided Pre-LTC Program / Average Monthly Costs for LTC Services Provided Under the LTC Program This ratio is used throughout the cost-analysis to compare the cost of LTC services before implementation of the LTC program to the cost of LTC services beginning with the full statewide program implementation in SFY When the cost-analyses ratio is greater than 1.0, then the payments made by the Agency for providers of Medicaid LTC services before the introduction of the LTC program, adjusted for inflation in order to make the comparison valid in nominal terms, were higher than the payments which were being made under the LTC program. For example, if the cost-analyses ratio is 1.30, then the Agency s payments to providers of Medicaid LTC services before the implementation of the LTC program were 1.30 times higher than that of providing LTC services under the LTC program. Placing the calculation in US dollar terms, if the cost-analyses ratio is 1.30, then for every $1.30 the Agency paid to providers for Medicaid LTC services before the LTC program was implemented, $1.00 is now being paid under the LTC program. In addition to the cost-analyses ratio, the estimated cost-savings were also calculated with the following equation: Cost-Difference = Average Monthly Inflated Costs for LTC Services Provided Pre-LTC Program - Average Monthly Costs for LTC Services Provided Under the LTC Program The interpretation of cost-savings is equally straight-forward. Cost-neutral equals $0. A positive number would reflect the average PMPM savings associated with the LTC program. A negative number would indicate that the LTC program costs more than the inflated service costs in the Pre-LTC program period. 142 Final Report (Fiscal Year ) June 30, 2017

143 Methodology Inflation Factors and Difference Analyses = (,, ) (,, ) (,, ) x 100% where: n = number of observations of a specific service, Q = the weight or frequency (of a specific service) on each specific price level, P = the various price levels of the specific service, t = period indicator (where the suffix -1 is a previous period, no matter what time period is taken), = (,, ) (,, ) x 100% (,, ) where inflation is part of the service nominal price difference: = and = (,, ) (,, ) (,, ) x 100% where: + = or + + = or = (,, ) (,, ) (,, ) x 100% Interupted Time Series The following equation is used: = +, +, +, where: Y t = N (N = 48) time series observations on the dependent variable or PMPM 143 Final Report (Fiscal Year ) June 30, 2017

144 X 1t = a dummy variable counter for time from 1 to N (N=1-24 through 39-48) (the first two months of SFY or n=37 and n=38 were disregarded as clear outliers) X 2t = a dichotomous variable scored 0 for observations before the LTC program implementation and 1 for observations after, starting in September 2014 X 3t = a dummy variable counter of time scored 0 for observations before the LTC program implementation and 1, 2, 3. for observations after the LTC program implementation starting in September In this model, b2 estimates the post intervention change in intercept, and b3 estimates the post intervention change in slope. Results on the 36 observations (24 months of Pre-LTC program and 12 months LTC program) are: = , , , + ( ) ( ) (0.1368) ) R-squared = N = Final Report (Fiscal Year ) June 30, 2017

145 College of Medicine College of Social Work Florida State University 145 Final Report (Fiscal Year ) June 30, 2017