ACO Congress. Using Analytics to Improve ACO Performance November 5, 2013

Transcription

1 ACO Congress Using Analytics to Improve ACO Performance November 5, 2013

2 Introductions Deb Davis OPTUM General Manager, West Region Accountable Care Solutions Jay Hazelrigs OPTUM National Lead, ACO Actuary Accountable Care Solutions 2

3 Analytics for ACO s High Risk Claimants Provider Performance 3

4 It may feel like you are scaling Mount Everest! You need critical information to help make critical decisions 4

5 The challenges in transitioning to value-based care Which patients are at high risk? What conditions exist in my population? What s happening to my patients outside of my four walls. Are my patients getting better? Who is managing the patient? What is the optimal treatment plan for the patient? Where am I on the quality standards? What is my bonus? How do I compare with my colleagues? 5

6 Electronic medical records are only the first step in the evolution, a technology for the fee-for-service model Longitudinal Population Health Management Prediction Stratification Fee for value Technology Analysis Reporting Transactional Monitoring Fee for service Low Health Status High 6

7 The next evolution is enabling technology that adds value to a fee- for-service and fee-for-value world Longitudinal Population Health Management Prediction Stratification Fee for value Technology Analysis Reporting Transactional Monitoring Fee for service Low Health Status High 7

8 Key Capabilities Needed Risk/Value-Based Contracts Advanced Clinical Risk Prediction Advanced Care Management Cost/Risk Analytics (Predicted Cost, Total Spend, Leakage, Etc.) HCC RAF/Risk Score Optimization Mobile Patient Education & Remote Monitoring PCMH Clinical Integration (HIE, Registries) Basic Clinical Analytics (Clinical Stratification) Patient Relationship Management Transition Management P4P Patient Registries & Gaps-In-Care Management Tasking and Alerting for Physicians in the EMR Physician Management (Scorecards, Dashboards, etc. Measure Reporting (ACO, PQRS, HEDIS, Etc.) 8

9 Population Segmentation

10 Population Segmentation Commercial High Level Example 10

11 HCC Risk Score Stratification by Beneficiary Member Key Death Date Attributed Physician Paid 7/11 to 6/ Final Score Total HCC Count HIV/ AIDS Septicemia/ Shock Opportunistic Infections Metastatic Cancer and Acute Leukemia 917 6/25/2012 Deceased $211, JOHN DOE, MD $143, /14/2012 Deceased $171, JOHN DOE, MD $182, /29/2012 Deceased $203, JANE DOE, MD $231, JANE DOE, MD $73, JOHN DOE, MD $82, JANE DOE, MD $100, /30/2012 Deceased $130,

12 Hidden Risk/Low Utilizers All Members Low Utilizers % of Total Living Beneficiaries 29, % Gender Female 17, % Male 12, % Race White 28, % Black % Asian % Hispanic % Native American % Unknown % Other % Age Category <65 3, % , % , % , % , % , % , % % Risk Score < , % , % , % , % , % % % % % % % Low utilizers are defined as beneficiaries with a risk score greater than and less than $1,000 in claims Identifies at risk populations by various categories Beneficiary detail is also provided 12

13 Frequent Flyers Emergency Room Visits People with Emergency Room Visits in November, 2012 Member Key Death Date Primary Diagnosis (highest cost) Paid 7/11 to 11/12 # ER Visits /23/ Cirrhosis of liver without mention of alcohol $42, Cellulitis and abscess of hand, except fingers and thumb $20, Epistaxis $4, Obstructive chronic bronchitis, with (acute) exacerbation $143, Anxiety state, unspecified $43,101 5 People with Multiple Emergency Room Visits in October and November, 2012 Member Key Death Date Primary Diagnosis (highest cost) Similar report will be provided for frequent admissions/readmissions Paid 7/11 to 11/12 # ER Visits Encephalopathy, unspecified $43, Lymphocytic choriomeningitis $35, Chest pain, other $42, Malignant otitis externa $13,

14 Frequent Flyers Monthly Summary Diagnosis Member Total ER Specific Key Primary Diagnosis Visits ER Visits 11/1/12 10/1/12 9/1/12 8/1/12 7/1/ Anxiety state, unspecified Migraine, unspecified without mention of intractable migraine without mention of status migrainosus Chest pain, unspecified Sprain and strain of unspecified site of back Early syphilis, latent, serological relapse after treatment Nasal bones, closed fracture Lumbago Headache Unspecified backache Shortness of breath Abdominal pain, unspecified site V681 Issue of repeat prescriptions Other dyspnea and respiratory abnormalities Unspecified sinusitis (chronic) Similar reports provided for frequent admissions/readmissions 14

15 High Cost Beneficiaries Death Total Paid Total Paid Total Paid Member_KEY Age status Date Hospice? Highest Cost DRG or Condition IN 2011 in 2009 in 2010 in Disabled Factor VIII Recombinant NOS $423,156 $558,483 $467, ESRD 0907 Other O.R. Procedures for Injuries with MCC 130, , , Aged 0003 ECMO or Tracheostomy with Mechanical Ventilation 96+ Hou 2,163 10, , Aged 0003 ECMO or Tracheostomy with Mechanical Ventilation 96+ Hou 1,378 61, , ESRD 0003 ECMO or Tracheostomy with Mechanical Ventilation 96+ Hou 62,689 73, , Aged 0003 ECMO or Tracheostomy with Mechanical Ventilation 96+ Hou 34,267 84, , ESRD 0329 Major Small and Large Bowel Procedures with MCC 132, , , Aged Oct Tracheostomy with Mechanical Ventilation 96+ Hours or Prin 6,666 2, , ESRD Dec Other Respiratory System O.R. Procedures with MCC 8,407 9, , Aged 0003 ECMO or Tracheostomy with Mechanical Ventilation 96+ Hou 1,445 45, , Aged 0003 ECMO or Tracheostomy with Mechanical Ventilation 96+ Hou 14,571 1, , ESRD 0252 Other Vascular Procedures with MCC 19,567 18, , Aged 0005 Liver Transplant with MCC or Intestinal Transplant 7,802 13, , Aged Jan Pulmonary Edema and Respiratory Failure 11,924 30, , Aged 0003 ECMO or Tracheostomy with Mechanical Ventilation 96+ Hou 7, , , Disabled 0003 ECMO or Tracheostomy with Mechanical Ventilation 96+ Hou 1,734 1, , Aged 0001 Heart Transplant or Implant of Heart Assist System with MC 110, , , ESRD 0239 Amputation for Circulatory System Disorders Except Upper Li 349, , , Disabled 0003 ECMO or Tracheostomy with Mechanical Ventilation 96+ Hou 1,359 37, , Aged Jan Other Digestive System O.R. Procedures with MCC 22,899 20, , Aged 0003 ECMO or Tracheostomy with Mechanical Ventilation 96+ Hou 1, , ,498 Report will be provided using paid expenses and HCC risk scores as basis for high risk claimants 15

16 Interventions Using Analytics

17 Ambulatory Care Sensitive Conditions Potentially Avoidable Admissions Data Conditions Number of Admissions Percent of Admits Rate/1000 Members Asthma % 3.19 CHF (non-htn) % COPD % DiabMel no c % 0.76 DiabMel w/cm % 5.37 Epilepsy/cnv % 3.69 Gastroent % 1.76 HTN % 1.68 PID % 0.11 Pneumonia % Tuberculosis % 0.08 UTI % Grand Total 3, % Potentially avoidable admissions or ambulatory care sensitive conditions defined by the Agency for Health Research and Quality. The metrics are based on admissions for conditions that may be avoided through well-managed, coordinated primary care. Additional detail based on attributed provider, servicing provider and facility can be used to identify opportunities to reduce admissions. 17

18 Preference-Sensitive Conditions: Admissions Acute Inpatient Admits Medicare Paid PMPM Admits per 1,000 unique Admits Paid (for admits State Percentiles of FFS Medicare people per 1,000 per Admit with the condition) Min 20th 40th 60th 80th Hip fracture surgical repair $15,064 $ Total hip replacement $17,549 $ Total knee replacement $15,344 $ Total shoulder replacement $13,771 $ CABG $48,287 $ Back Surgery $18,251 $ Carotid endarterectomy $15,760 $ Lower Extremity Bypass $24,586 $ Angiography $24,421 $ Mastectomy $5,949 $ PCI/PTCA $21,628 $ Colectomy $28,046 $ Cholecystectomy $17,589 $ Prostatectomy $11,272 $ Carotid Endarterectomy, Lower Extremity Bypass, and Mastectomy each appear to have relatively high admit rates, as compared to state percentiles of FFS Medicare 18

19 ACO System Care Retention by Member, Physician and Hospital Who has attributed patients seeking care within the delivery system? Attributed Physician Physician NPI Name Members Total Inpatient Outpatient -PCP - SPEC DME Deceased Deceased % 31% 20% 67% 4% 48% JOHN SCOTT MD % 0% 5% 77% 1% 26% JOHN BECKNER MD 415 8% 3% 1% 68% 1% 35% JOHN MANLEY MD % 42% 44% 90% 1% 44% JANE LEAHY MD % 61% 50% 88% 3% 65% JANE AUSTIN MD 401 8% 1% 1% 66% 1% 54% JOHN REESE MD % 14% 19% 87% 6% 31% JANE AUSTIN MD 338 9% 2% 1% 70% 1% 45% JOHN BRAKEBILL MD % 52% 36% 89% 4% 56% JOHN THOMAS MD % 65% 69% 86% 4% 66% JANE MONTGOMERY MD % 51% 48% 88% 5% 57% Beneficiaries of the largest attributed physician have no innetwork inpatient admits This attributed physician has lowest PCP care retention potential loss of beneficiaries from CMS attribution ACO may have additional opportunities to manage care by expanding their network, formally or informally, to include specialists 19

20 Provider Detail Tool Provider Summaries Provider NPI/Site Number ACO Average Attributed Members Attributed Member Months 7, ,475 Average HCC Risk Score Average PMPY $ 7, $ 7, Average PMPM $ $ Member Utilization ACO Average No. of Low Utilizers No. of Top 100 Frequent Flyer ER Members No. of Top 100 Frequent Flyer IP Members No. of High Cost Beneficiary Members No. of Readmits No. of Avoidable Admits No. of Preference Sensitive Conditions IP Admits/1, ER Visits/1, Provides comparison for ACO average or ACO- provided benchmarks Chronic Conditions Member Count % of Total ACO Average Diabetics % 27.5% CHF % 8.8% COPD % 11.4% CAD % 6.0% AMI 5 1.2% 1.0% Vascular Disease % 9.2% Dialysis or Renal Failure % 9.9% Mental Health/Psych 5 1.2% 4.1% 20

21 Provider Detail Tool Assigned Beneficiaries Characteristics Member Summary ACO Assigned Record Attributed Member Key Number Risk Score Paid $ 185, $ 105, $ 84, $ 80, $ 74, $ 70, $ 68,292 Utilization Summary By Member Key Low Utilizers ER Frequent Flier IP Frequent Flier High Cost Beneficiary Report also includes detailed claims of attributed beneficiaries 21

22 Provider Detail Tool Chronic Disease Registries NPI: Provider: JOHN DOE, MD Beneficiaries: 401 Members with Specific Chronic Condition by Member Key Vascular Disease Diabetics CHF COPD CAD AMI Dialysis or Renal Failure Mental Health/Psych Over 30% of this physician s beneficiaries have diabetes Shows members with multiple chronic conditions 22

23 Provider Detail Tool Physician Performance Avoidable Admits Preference Sensitive Conditions Summary Summary Avoidable Admits Count Paid Preference Sensitive Condition Count Paid Bacterial pneumonia 10 $ 39,045 ER visit for abdominal pain 12 $ 4,647 Cellulitis 1 3,352 ER visit for acute bronchitis 3 1,035 CHF 3 21,328 ER visit for asthma COPD 2 10,465 ER visit for back pain 4 1,036 Diabetes 3 11,588 ER visit for fever Kidney/urinary infection 1 3,047 ER visit for head ache ER visit for nausea/vomiting 3 1,021 ER visit for UTI 16 9,325 ER visit pharyngitis 1 44 IP visit for Angiography 4 65,569 IP visit for Back Surgery 2 9,246 IP visit for CABG 1 35,390 Indicates opportunity for physician education to Indicates opportunity for increase savings physicians to educate beneficiaries on appropriate healthcare choices Report also includes detail of which beneficiaries are having avoidable or preference sensitive admits 23

24 Project Potential Impact of Interventions 24

25 Operational and Programmatic Strategies

26 Care Management Programs Care Management Deployment Services Avoid the cost, time and risk to develop, design and delivery population health management services independently Referral Management Avoid revenue losses as a result of patient engagement outside the health system Primary Care Clinic Patient Eme rgency Room Emergency Department Redirection Avoid costly delivery of services in the ER that are more appropriately delivered by PCP Care Access Services Readmission Prevention Reduce medical costs driven by redundant or inappropriate utilization Specialty Care Clinic Inpatient Hospital Avoid revenue losses as a result of non-payment for readmissions within 30 days Home Complex & Chronic Patient Management Drive quality compliance with EBM, improve patient outcomes and generate medical cost savings to improve pay for performance or risk bearing opportunities 26

27 Case Study Steward Health Care Network Steward Health Care Network Client Overview: Steward Health Care Network (SHCN) is the second largest physician network in Massachusetts with over 1,100 member physicians comprised of both primary care physicians and a full range of specialists. Steward participates in the CMS Pioneer ACO Program where they are responsible for managing the health their attributed Medicare fee for service beneficiaries. 27

28 2013 Pioneer Covered Lives Local Chapter by Month by Covered Lives 2013 Pioneer Jan 2013 Pioneer Feb 2013 Pioneer Mar 2013 Pioneer April 2013 Pioneer May Cape Cod Pioneer June Greater Boston West 5,320 5,268 5,223 5,195 5,149 Greater Brockton 11,136 11,085 11,040 11,001 10,945 Greater Dorchester 2,316 2,303 2,291 2,276 2,267 Greater Fall River 3,124 3,096 3,078 3,064 3,044 Greater Haverhill Greater Methuen 5,712 5,681 5,663 5,634 5,605 Greater Nashoba 2,489 2,475 2,466 2,449 2,442 Greater Norwood 4,386 4,353 4,339 4,321 4,296 Greater Quincy 2,010 1,996 1,986 1,973 1,964 Greater Taunton 8,270 8,229 8,196 8,165 8,131 Out of Area Data Opt-Outs* 2,261 2,261 2,261 2,261 2,261 2,261 Total 47,161 46,884 46,680 46,476 46,241 46,002 28

29 Develop Interventions: Complex & Chronic Patient Management Value Proposition: 1-2% cost savings for the managed population with primary drivers of admit reduction and ER visit reduction Targeted Population: High Cost = Healthcare cost >/= $10,000; High Risk = No PCP/multiple physicians, 4 or more ER visits in 12 months, 2 or more unscheduled inpatient admissions in 6 months, Inpatient length of stay >/= 10 days, Readmission within 30 days, Coordination of multiple services (PT, OT, ST, HH, and high risk DME), Poly Pharmacy >/= 10 prescriptions, Co-morbid mental health condition, and multiple trauma Focused Interventions: Medication reconciliation Compliance with EBM and treatment plan Reduction in total cost of care for targeted population Method of Engagement: Local telephonic RN s aligned to System Provider organizations with optional face to face engagement in home/physician office Assumptions: Savings valid for Optum defined program with no variation to targeted population, method of engagement, focused interventions, staffing or workflow/processes Physician communication, engagement and participation in support of program and patient participation is expected and should be facilitated by client IP admit and ER visit reductions serve as a prospective proxy for the achievement of 1 2 % cost savings for the managed population Retrospective measurement of program effectiveness is calculated by comparing medical costs across 2 time periods prior to program implementation (baseline group) and after the program is initiated (intervention group). A identification approach is used, whereby individuals are identified for the baseline and intervention periods using 24 months of data but claim costs are calculated using the most recent 12 months of experience (with 3 months of claim run out) Value/Savings Model: Complex and Chronic Patient Management: 2% Assumptions Value Overall Membership 26,000 Outreach 2% 520 Engagement 70% 364 Driver #1 ER Reduction ER visits per 1, ER Visit Reduction of Engaged 20% 65 Average cost per ER visit for CCPM patients $ 1,000 Driver #1 Savings $ 65,302 Driver #2 Admit Reduction Admits per 1,000 1, Admit Reduction of engaged 30% Average cost per admit for CCPM patients $ 15,000 Driver 2 Savings $ 2,358,720 Total Gross Savings $ 2,424,022 Utilization Adjustment Factor 15% $ (363,603) Net Savings $ 2,060,418 % Gain Sharing to System 50% $ 1,030,209 CCPM 2% Total $ 1,030,209 29

30 Steward Program Components Complex & Chronic Patient Management Covers 6 provider organizations approximately 100 PCPs Includes 9 Optum nurses with varied clinical backgrounds and 1 nurse with psychiatric nursing subspecialty; meets language needs of beneficiaries (Russian and Portuguese) Community-based program with geographical considerations in beneficiary/provider organization assignment allows for home- and office-based contract Beneficiary identification and stratification uses a combination of predictive modeling and real-time data Physician community makes direct referrals 30

31 Complex and Chronic Patient Management - ER visits declined 16% All Chapters Total Member Population Engagement Summary since Program Inception (July ) Count % Target Total High Risk Prospect in CCPM Program Identified High Risk through Algorithm Other Referral Total Member Population Outreached Enrolled (currently enrolled or discharged) Members: ,500 31

32 Steward Program Components Readmission Prevention Program Deployed at SNF level of care (versus IP acute) Provides coverage for 7 markets (aligned with major IP acute care facilities) consisting of 43 SNFs Jointly staffed with 6 Steward nurses and 2 Optum nurses Makes initial contact face-to-face while beneficiary is in SNF Conducts post-discharge followup care via phone call 32

33 Readmission Prevention (RAP)- Readmission rates declined 25% All Chapters Total Members Discharged From Steward Hospital to SNF (Estimated for Period Nov 2012 Mar 2013) 1,237 Engagement Summary Since November 2012 Patients referred to RAP (from Pioneer Patient Tracker) Count % % N= 964 N= 676 N= 370 N= 702 N= 491 N= Readmissions for beneficiaries with an acute discharge to a SNF 33

34 Financial Analysis

35 Basic Business Model Historic Benchmarks, Performance Year Targets and Actual Expenditures Are we making any money? Updated Benchmark/Performance Year Targets Gross Shared Savings 35

36 Financial Modeling Considerations One step at a time.. need to understand and quantify Medicare Benefit Cost Sharing Impacts ACO Program Definition Impacts (i.e. paid through dates, interim billing, etc.) Duplicate Claims IBNR / Speed of Payments Truncation Local and National Trends by Eligibility Type Medicare Unit Cost Fee Schedule Changes Local and National (i.e. AWI & GPCI) Excluded Beneficiaries and Medical Expense Impact Opt-out Beneficiary Claims Substance Abuse Claims Beneficiary Churn Risk Adjustment Impacts 36

37 ACO Average Claims Cost Projection vs. Updated Benchmark Base projection almost 3% below updated benchmark ACO claims net increase 3.7% from base projection once all adjustments considered Initial estimate showed ACO revenues, final projection does not Updated Benchmark / Target = $10,300 37

38 Conclusions

39 Conclusions Segment Population: Leverage Payer Data Understand Your Risk - Strong Actuarial Analysis Quantify Interventions: By Condition By Physicians Care Network Location Develop Care Management Programs: Condition-Based Use Dynamic Registries Measure Impacts Maintain Financial Discipline: Understand Your Contracts and Targets Understand & Update Financial Results Constantly 39

40 Contact Information Deb Davis Jay Hazelrigs