Nicholas E. Davies Enterprise Award of Excellence Population Management Case Study

Transcription

1 Applicant Organization: Centura Health Organization s Address: 188 Inverness Dr. W #500, Englewood, CO Submitter: Amy Feaster, Vice President of Information Technology amyfeaster@centura.org Menu Item: Population Management Executive Summary Centura Health, founded in 1996 by our sponsors, Catholic Health initiatives and Adventist Health systems, manages 16 hospitals in Colorado and Kansas. These hospitals range from geographically remote, 30-bed hospitals, to a Level One Trauma center with 220 beds in a metropolitan area with a population of more than 2.5 million. Our mission statement states that we will, nurture the health of the people in our communities, and we knew that we needed to think outside the box in order to better live this part of our mission. As we approached the challenge of providing health care services for large populations and communities rather than simply for individuals, we knew that this would require an entirely new mindset. It is a game changer for everyone involved in the care of these patients, from acute care and office practices to call centers and home health support. We have worked diligently on building effective working relationships quickly in a landscape that changes with the regularity of quicksand. The population health model of accountable care demands new methods of storing, sharing, and acting on patient related data; we have had to find creative ways to assist with the project from within our EHR, which is built for the traditional model of taking care of one patient at a time. Additionally, we have had to identify and incorporate other types of software such as a registry tool and a customer relationship management (CRM) tool, and we have needed to push our HIE for new services as well. Our IT experts have been challenged at multiple points to find creative solutions for the new issues and needs of these programs. We would not be successful without a strong senior leadership team with unwavering commitment to learning new ways of providing care to our communities. One of the principles espoused by our Centura leadership is the concept of being a Learning Lab. We are willing to commit resources and time trying out good ideas in different environments, with the goal of defining methods and solutions that work. We have approached learning about population health management using this concept. Below we describe two different approaches that are both showing early process and outcome successes. One story addresses population health impact in a geographically isolated population; the other describes an approach more suited to a large metropolitan area. Although the goals of population health management are similar in both projects, the tools and methods differ. Both are addressing controlled populations and are considered proof of concept activities. Centura looks forward to applying learnings from both of these projects to other populations and communities in the future. Case study: Mercy Family Medicine CPC Program Background knowledge Mercy Family Medicine (MFM) in Durango, Colo., is one of 75 practices in the state of Colorado participating in the CMS sponsored Comprehensive Primary Care (CPC) initiative. From the CMS website: Page 1 of 16

2 The Comprehensive Primary Care (CPC) initiative is a four-year multi-payer initiative designed to strengthen primary care. Since CPC s launch in October 2012, CMS has collaborated with commercial and State health insurance plans in seven U.S. regions to offer population-based care management fees and shared savings opportunities to participating primary care practices to support the provision of a core set of five Comprehensive primary care functions. These five functions are: (1) Risk-stratified Care Management; (2) Access and Continuity; (3) Planned Care for Chronic Conditions and Preventive Care; (4) Patient and Caregiver Engagement; and (5) Coordination of Care across the Medical Neighborhood. The initiative is testing whether provision of these functions at each practice site supported by multipayer payment reform, the continuous use of data to guide improvement, and meaningful use of health information technology can achieve improved care, better health for populations, and lower costs, and can inform future Medicare and Medicaid policy. Visit for more information. Over 490 practices across the nation are participating in this initiative, and MFM is the only practice in Centura Health chosen through an application process to participate. We are currently in the third year of the program, which is scheduled to end December 31, The CPC project requires work in the following areas: care management for high risk patients, access and continuity for all patients, patient engagement, quality improvement, care coordination across the medical neighborhood, shared decision making, participation in several learning collaboratives and health IT innovation. Local Problem being Addressed and Intended Improvement: Mercy Regional Medical Center is an 82- bed community hospital located in Durango, Colo., a geographically remote area with limited ambulatory care access. As the sole full-service hospital in the area, MRMC owns both primary care and specialty practices, including Mercy Family Medicine (MFM). The EHR is shared across acute care and ambulatory practices. In order to serve our community more efficiently, we began to transform our primary care practices to NCQA-recognized patient centered medical homes (PCMH). This transformation coincided with our acceptance into the CPC Initiative, and we consider that foundational in achieving the goals of the CPC project. As part of this project, we have been committed to identifying a primary care provider (PCP) for every patient whom we treat, to develop a risk stratification tool for identifying at-risk individuals, and to design care coordination services for those patients. In January 2012, 79% of MRMC s patients had a PCP identified in the EHR. Our unplanned readmission rate for Medicare Fee for Service (FFS) discharges was 141/1000 in January 2012, and our Medicare expenditures for attributed Medicare patients in the highest risk quartile was $979 per member per month (PMPM). We set goals to reduce unplanned readmissions, and reduce overall cost of care by providing active case management in ambulatory settings. Design and Implementation: The work of rapid transformation from a traditional primary care clinic to a NCQA Level III recognized patient centered medical home was accomplished in 2013 and recognized in October This occurred while simultaneously gathering baseline data and meeting milestones set forth by the CPC project. That transformation was instrumental to our success. The initial requirement of the CPC initiative was to empanel every one of our 14,000 patients to one of our PCPs. We started this process in Through reporting enabled by our IT partners, we were able to identify which patients needed a PCP, and using a four cut method were able to attach patients to a provider within our clinic. Our four cut method is loosely modelled after Murray and Davies (1): 1. Patients who have only seen one provider are assigned to that provider. Page 2 of 16

3 2. Patients who have seen multiple providers are assigned to the provider who cared for the patient the majority of the time. 3. If there is a tie between providers, than the patient will be assigned to the provider who last cared for them. 4. For remaining patients, the business support staff will contact the patient via phone to have them identify their primary care provider. At the end of 2014, we had empanelled 99.98% of our patients. This percentage has now been sustained for nearly a full year. Risk stratification was the second undertaking for our team, which was comprised of a physician champion, a project coordinator, clinical staff and administrative support staff in the clinic. Additional resources were provided by our parent hospital in the form of IT infrastructure, training, planning and education. Risk stratification is accomplished by clinic staff using a combination of discrete data populated in the EHR, collaboration with the PCP, and personal interviews with our patients to identify more subtle psychosocial risks. After surveying the literature, we developed a proprietary risk tool to be sure we captured both objective and subjective aspects of risk. We also included our pediatric population from the inception of the project, and developed a risk tool appropriate for that population. Elements of the adult risk tool include: number of hospitalizations, ED visits and office visits, number of active prescriptions (including oxygen), language, health literacy assessment, number of chronic medical conditions, BMI, tobacco history, number of mental health diagnoses, significant life stressors, social determinants, and self-management abilities. Elements of the pediatric risk tool include: age, number of hospitalizations, ED and office visits, active medications (excluding antibiotics), family language and health literacy, number of chronic medical conditions, tobacco exposure, growth parameters, family or patient mental and behavioral health conditions or major life stressors, and family/caregiver social determinants and self-management abilities. The objective data is captured prior to a patient s visit and the subjective data is collected at the time of the face-to-face encounter. All patients with appointments in the practice are screened using these risk tools and given a risk score, with a higher number indicating a higher risk for health problems requiring intervention. Each patient is reassessed at periodic intervals, depending on the risk score. The risk stratification can be reviewed in the EHR by any provider or clinical staff. The following describes the care coordination activities provided to patients at each risk score: Risk Score Reassessment interval 1-2 Annually 3 Every 6 months 4-5 Every 3 months 6 Monthly Care Management action steps Primary prevention: Care management that is aimed toward keeping the patient educated about his/her health and preventing illness by incorporating preventative screenings/immunizations Secondary Prevention: Care management that is aimed toward keeping the patient s diseases in control; providing them with educational and service opportunities that will maintain or improve their disease state Tertiary Prevention: Care management that provides intensive education, referral to other resources for the patient s/family s wellbeing Complex Care Management: Care management of patients with multiple co-morbidities, multiple medications, frequent ER visits and hospitalizations Ninety-five percent of our 14,000 patients are risk stratified and care management has been available and active since 2013, using monthly contacts, coaching, education, and leveraging community support Page 3 of 16

4 services for particular needs. The care coordinators actively manage all Level 6 patients, and all diabetics with poor control (HbA1C greater than 9 mg/dl). They receive regular reports generated from the EHR identifying high-risk individuals in order to guide their activities. The RN Care Coordinators (RNCC) document the patient s care plan in the EHR in notes that are clearly labeled as RNCC notes. These are viewable by the providers and clinical staff members, so everyone is aware of the risk score and the care activities for each patient. Our Risk Stratification Tool has proven to be highly reliable. Twenty percent of our population falls into higher risk and one percent falls into the highest risk category. These statistics parallel the literature regarding risk levels of our national population. The Mercy Adult and Pediatric Risk Stratification Tools are described in the CMS CPC implementation guide. Our lead project coordinator provides hands-on training using the risk scoring tool with all clinical staff, including providers, MAs and other nursing support staff. The risk score requires about three minutes per patient visit, and is generally done by the MA for the objective elements, and is done in conjunction with the provider for the subjective elements. Our providers felt strongly that the subjective elements of risk scoring contribute to appropriate identification of people in need, and they know they can influence the provision of services by recording such things as social determinants, health literacy, and life stressors in the EHR. This type of engagement was accomplished due to the fact that we had a physician champion and provider involvement in each step of this project. When a patient is determined to be a level 6, a referral is made to the RN Care Coordinators. Care coordinator activities include the following: Care manage with phone calls, office visits, home visits by means of motivational interviewing, goal setting, health coaching, patient education and setting up individualized care plans. Care Coordinators visit each hospitalized patient and ensure a post-discharge follow up visit within seven to fourteen days of discharge with their PCP. Care Coordinators also call each discharged patient within 48 hours post-discharge. They do medication reconciliation for all post hospitalized patients during that time. Coordinating services with other community health resources and public and private assistance agencies. Collaboration with the public health department, providers, hospital case managers, nurse navigators, and social workers to manage high Emergency Department (ED) utilizers in a coordinated fashion. The patient is asked to participate in setting personal goals. These are referred to in phone contacts and office visit contacts to keep patients engaged in their care. The patient is given a written summary of their goals and action steps after each visit. The CPC initiative is a multipayer coordinated effort. For the Colorado coalition, the payers have coordinated their reporting efforts with the CPC clinic stakeholders to deliver a unified report to the participating practices and identify at-risk patients based on claims data. This has significantly reduced the time and effort needed to receive and interpret claims data in paper reports from multiple companies and cross reference those to our own risk assessment. With the electronic multipayer tool, we can now cross reference patients identified as at-risk from insurers claims data against our assessments to ensure that we have truly identified all people in our service area who are at risk. With this tool we can identify high ED utilizers regardless of whether they present to other sites for care, and have included those people in our active care management program. Page 4 of 16

5 How was Health IT Utilized? Rather than develop or buy new tools, this project relied on getting the maximum use out of the information and features already available in our EHR. We use the EHR encounter summary, problem list and medication lists to assess the number of visits to the office, ED hospital, number of chronic problems, and active medications. The problem list will auto-populate with a status of Acute or Chronic depending on the ICD code that was utilized, therefore allowing for the chronic diagnoses to be counted easily. Our risk stratification tool flows exactly as the screens in the EHR flow to minimize the number of clicks and screen changes for those performing the risk assessment. Our care coordination documentation (named RNCC Care Plan and RNCC note) is filed in an identifiable manner in the encounter summary and available to all care givers: Health Information technology is at the heart of supporting patient centered medical homes. The population is managed in a registry to provide regular reports to the care coordinators so they can prioritize their contacts and work. Daily reports are available from the EHR so the care coordinators know which patients have been seen in the ED or admitted in the past 24 hours. Care coordinators work in the same offices as the providers and keep in close contact for the high-risk patients. A multitude of reports have been created by our IT team to assist our practice in the overall management of our patients. MFM was able to report electronically on 23 of the 24 clinical quality measures defined by the CPC project the first year. Of the 14 practices in our regional cohort, we were the only clinic to manage this at the onset. This past year, the nine selected CQMs were reported via an electronic file submission (QRDAIII), which other practices were not able to do, therefore reporting will be submitted via attestation in the future. Mercy Family Medicine would not have been able to accomplish these required tasks without the use of our EHR and data team. Page 5 of 16

6 We provide regular feedback of both process and outcome data to the entire practice staff. Process measures include rates of contacts made in the expected time frame, diabetes care measures and multiple other condition oriented process measures. Physicians and clinic staff work in teams to optimize the measures. Value Derived/Outcomes: Despite a higher than median Hierarchical Condition Category (HCC) patient risk score for our population, our annual report from CMS reflects the following successes: Per member per month expenditures are going down, from a four quarter average of $659 in January 2014 to $603 in January 2015 (median value among Colorado CPC practices was $686). Hospital admission rates per 1000 beneficiaries for any cause went from 226 in January 2013 to 177 in January 2015 (median value among Colorado CPC practices with a similar risk profile was 253). Hospital admission rate per 1000 for Ambulatory Care Sensitive Conditions (ACSC) went from 36 per 1000 in 2013 to 26 per 1000 in 2015 (median value among Colorado CPC practices with a similar risk profile was 49). Our rate of unplanned 30 day readmission, while not going down, remains below the median for matched practices, both for the whole cohort and for the highest quartile risk group. Page 6 of 16

7 Page 7 of 16

8 The rise in unplanned readmissions may be explained by several factors, including continued growth with a tripling of providers, as well as an influx of Medicare beneficiaries from the entire geographic region as other providers opted out of caring for Medicare clients. We have four times the average number of Medicare clients as other CPC participants in our state. In addition, our population has a higher proportion of high-risk and dual Medicare-Medicaid insured clients than the other CPC practices. Despite caring for a population heavily weighted toward high-risk by several measures, we have reduced overall and ACSC hospitalizations. Compared to the four quarters prior to the start of the CPC project, unplanned readmissions have decreased from 120/1000 discharges to 97/1000 in the most current four quarters. The project has had a positive impact, and the recent rise in unplanned readmissions may be within our acceptable limits of variability. Our care coordinator staff has stabilized recently, and we will monitor this closely as those staff members develop ongoing stable relationships with their clients. Our number of hospitalizations/1000 Medicare beneficiaries decreased by 22%, from 226 in Q2 (Jan March 2013) to 177 in Q10 (Jan March 2015)( This correlates to 122 avoided hospitalizations year over year. Lessons Learned: The Mercy team struggled with the rapid pace of change and several adjustments in leadership in 2012 and 2013.s The project coordinator served as a strong and committed local leader who was able to stabilize the team and create enthusiasm for the results. The physician champion s commitment to the project has been unwavering and without his leadership the other providers would not have followed suite. The responsiveness of the IT support systems in extracting data for the Clinical Quality Measures for the past 3 years and electronically filing reports with CMS has been monumental. Our PCMH structure and commitment leant itself well to incorporating care management activities. Our determination to use the tools we have at hand rather than waiting for development or purchase of IT tools helped us stay on schedule and meet deadlines; in addition, it helped our clinical staff learn about EHR capabilities that they were not aware of. We did have an unusual alignment of common interest with our insurers that resulted in coordination of reporting efforts. We acknowledge the value of those pre-existing collaborative relationships with several of the major insurers in the state that laid the groundwork for this very beneficial work. Financial Considerations: The process of risk stratification adds about one hour per day to our MA time, therefore costing approximately $5,000/year. The annual cost of the care coordinators is $90,000 each. Training in the role requires about two weeks. Interface with QRDAIII and reports development together cost $8,000. Our PCMH structure includes bi-monthly quality team meetings with 8-12 staff members, who work on workflow and optimization. This is required for NCQA recognized PCMH clinics and therefore was a part of our routine budgeting process. Mercy Family Medicine (MFM) clinics receive $500,000 to $700,000 per year from the various payers and CMS to accomplish the milestones outlined in the project. A project coordinator works closely with the practice to improve the quality of care with rapid cycle improvement methods such as PDSA. A requirement of the project is use of an ONC certified EHR; Centura had made this investment at the system level in prior years. Page 8 of 16

9 We have received one shared savings payment from a payer group to date. CMS will award shared savings based on how the entire region performs statewide (for each individual state involved with CPC). Based on a locally determined cost per hospitalization of $26,000, and our estimate that we have avoided 122 hospital admissions in the past year, we have avoided costs totaling approximately $3,172,000 from January 2013 to January References [1] Panel Size: How Many Patients Can One Doctor Manage?, Mark Murray, MD, MPA, Mike Davies, MD, Barbara Boushon, RN, Fam Pract Manag Apr; 14(4):44-51 Case study: Integrated Care Network: Centura s ACO Story Background knowledge Sweeping reimbursement changes created an environment in which Centura decided to participate in the Accountable Care Organization structure from the early stages. We joined the Medicare Shared Savings Program in 2012 and have been learning and growing in our expertise since. Centura leads the market in involvement in innovative networking with health care partners, including insurance companies. Our clinically integrated network, Colorado Health Neighborhoods (CHN) is comprised of 3,100 providers, of whom approximately 400 are employed. We participate in a total of seven ACOs currently, covering about 270,000 lives, and more are in the planning stages. Local Problem being Addressed and Intended Improvement: Payers have begun to ask for high performing narrow networks based on quality metrics and a proven ability to reduce the overall total cost of care. In our ACO populations, we targeted several conditions for active care planning and management, including COPD, heart failure, and diabetes. Interventions include population identification, risk stratification, identifying care gaps and addressing them, and ensuring careful transitions of care. Because we were committing to caring for an unknown population, we did not have baseline data regarding any of the target conditions. However our baseline rate of 30-day readmission for Heart Failure discharges for our whole system in 2013 was 9.89% (all payers included). We knew that in an accountable care model this rate would lead to unfavorable results. Design and Implementation: We created a dedicated team of post-acute care experts, care management experts and informaticists whose responsibility was to develop and maintain our ACO products, contract with payers under this model, and deliver results that are both financially favorable and good for our patients. This group of experts had no other clinical or operational duties, in acknowledgement of how deep and transformational this change is anticipated to be for our organization. However, the fact that this was a separate and distinct team necessitated development of close working ties with our hospital leadership, group practices and integrated health network practices. Many hours of conceptual discussions were needed to be sure that our whole community understood the goals and purpose of our Second Curve team. We continue to develop and mature the distinct tasks and action steps needed in the ACO construct to care for patients seamlessly across the care continuum. The initial challenge was identification of the people who are covered in each ACO in near real time. Then, risk stratification was necessary, and a basket of services was developed, offered and delivered to those individuals who fall in higher risk scores. Entry into a condition-based registry is accomplished using EHR problem list data or claims data, Page 9 of 16

10 depending on timing and EHR interface status. Coordination of care so that care is delivered in the most appropriate setting rather than the most expensive setting was the next goal. This involved prompt notification of ED visits and of inpatient admissions, both to Centura hospitals and to external hospitals. Next, communication links with the patient s primary caregivers were needed along with education and a toolkit for those PCPs. Both of these aspects required building toolkits, developing lines of communication, and constructing feedback loops with performance data. Collaborating with the physicians of CHN, we developed clinical pathways that include patient action plans for heart failure, COPD, asthma, CAD, and diabetes. Our action plans use the familiar red light/yellow light/green light model to guide patients. Providers then have standing orders associated with each stage of each action plan to guide their care decisions. Personal visits are made to the practices both by the physician Director of Accountable Care and, when a practice is struggling, by a performance improvement specialist. Members of the team at the practice are taught to ask, What is the guideline, what is the measure, how do you document it in your EHR, and how do you use the clinical pathways and action plans? As might be predicted, our level 3 NCQA certified Primary Care Medical Homes need less coaching and support than practices who have not achieved this recognition. As we grew, it became clear that our Care Coordinators required a tracking tool; a customer relations management software tool was identified for this purpose. 24/7 availability of care coordination services is provided using a call center model. Page 10 of 16

11 FullWell, our Care Management arm, is devoting significant time, energy and resources to data management and reporting and analytics capabilities to support our ability to deploy targeted, personalized interventions in response to key insights about the health of the populations we manage. Development is now underway for an online portal where providers will be able to access the clinical guidelines, tools and quality reports in one place. Process measures are tracked in four domains: patient/caregiver experience, care coordination/patient safety, preventive health, and at risk population. These may be seen on the CMS website at Service-Payment/sharedsavingsprogram/Downloads/ACO-Shared-Savings-Program-Quality-Measures.pdf Very similar metrics are established for the other six ACO projects and for our self-insured populations. All metrics, generated by our team or by the insurers, are shared with the involved practices by quarter. We prepare reports to clearly display individual provider, practice, ACO, and benchmark data so that each provider and practice understands their performance in comparison to expectations and to a benchmark. These discussions are frequently done in person, with coaching available about how to improve. Page 11 of 16

12 How as Health IT Utilized?: Committing to a population of patients defined by an insurance plan means monitoring care and affecting outcomes that may be happening in an external hospital and in multiple EHRs. Since we are on a standardized single platform EHR for Centura hospitals and practices, and had traditionally only managed those patients who had encounters with our services, this meant new and creative thinking about IT support for such a venture. We receive membership files monthly from each insurance carrier and use third party software to risk stratify the population using claims data. We then identify the primary care home for those high-risk individuals who will receive interventions. Care transitions from Centura hospitals are managed in a post-acute care system. Care coordination is documented and tracked in a CRM tool. Where the primary care of the client occurs in the system EHR, we document the care coordination in the patient EHR record as well. We work with practices using approximately ten different EHRs, and for that reason have opted to centralize the care coordination in a CRM rather than in any one EHR. We are working with our HIE to deliver ADT data from external hospitals to our database. Centura hospital admissions are identified by a flag in the system EHR designating patients who are participating in these accountable care programs. In this manner, caregivers and case managers are aware of these patients at each point of contact. Initially, this flag was added manually at the time of registration; we have since implemented an electronic data feed that we receive from the payers. For each admitted ACO client, within 15 minutes of the admission order, an automated page is sent to the inpatient case manager. A report is run every day for the hospital-based case managers to identify the ACO patients in order to proactively work on smooth and seamless transitions of care. Data is fed from our EHR to our post-acute care database so that our ambulatory based care coordinators are aware of risk level and issues and participate in care documentation. Risk stratification is achieved using claims data and refined using clinical data at the time of first contact post discharge to determine what level of services the client will be offered. Those in the highest risk categories are offered care management across the continuum, including home visits, monthly phone contacts and tools and education for managing their condition. We use condition-based registries to monitor care gaps and outcomes. The ACO flag in the admissions module: Page 12 of 16

13 The flag displays on the ED and the inpatient and the case management status boards: And on provider rounding lists: The ACO flag is meant to trigger all caregivers to inform the patient that they are eligible for extra services, and that a care coordinator may call them after departure or discharge to explore ways that we can help them navigate their condition and the care needed for it. Below is an example tracking report from the ACO registry which is available to all practicing providers in Centura Health Physician Group and Colorado Health Network: These reports are available on demand in the provider s office and are meant to guide the provider and her/his staff to close any gaps in care that are identified. Electronic systems are also utilized to identify and track gaps in care based on information in our EHR. This allows us to identify needed services for our patients in a timely manner which in turn improves the health of the population we serve. Identification of gaps in care on this large scale would not be possible in a paper-based world. Numerous condition-specific process metrics are tracked to close these gaps in care. The care coordinators will contact the patient to encourage compliance with recommended care Page 13 of 16

14 processes, and the PCPs receive regular reports of patients with gaps in care so they can also encourage the patient to participate. For those patients who are on a remote monitoring program, if data is not sent from the in home monitor in a set period of time, a message is sent to the care coordinator who will make phone contact with the client. Value Derived Outcomes: By reducing care gap indexes and providing coordination of care across all venues of care, we have successfully positively impacted the metrics for three conditions for our Medicare Shared Savings Program (called Colorado Accountable Care or CAC). For our heart failure patients, this program has achieved a 15.65% reduction in readmission for heart failure, and has reduced the overall annual PMPY cost of care by 7.4%. Chronic Disease Management: CHF* Months Benchmark (Norm) CAC 2013 Avg CAC 2014 Avg CAC 2015 YTD % Change From 2013 Utilization Data PMPY $33,830 $31,073 $30,274 $28, % Readmission Rate 22.65% 20.37% 18.50% 16.81% % Quality Data Care Gap Index % Efficiency Indices Cost Index (Actual Costs/Predicted Costs based on % Relative Risk Score) Admission Utilization Index (Actual Admits/Predicted Admits) % ED Utilization Index (Actual/Predicted ED Visits) % We have achieved similarly promising results in addressing diabetes with a 18.32% reduction in readmissions, and COPD with a 11.25% reduction in readmissions year over year: Chronic Disease Management: Diabetes* Months Benchmark (Norm) CAC 2013 Avg CAC 2014 Avg CAC 2015 YTD % Change From 2013 Utilization Data PMPY $16,360 $13,878 $13,878 $14, % Readmission Rate 18.85% 15.56% 13.80% 13.97% % Quality Data Care Gap Index % Efficiency Indices Cost Index (Actual Costs/Predicted Costs based on Relative % Risk Score) Admission Utilization Index (Actual Admits/Predicted Admits) % ED Utilization Index (Actual/Predicted ED Visits) % Page 14 of 16

15 Chronic Disease Management: COPD* Months Benchmark (Norm) CAC 2013 Avg CAC 2014 Avg CAC 2015 YTD % Change From 2013 Utilization Data PMPY $23,887 $19,579 $19,568 $19, % Readmission Rate 20.37% 16.84% 15.91% 14.14% % Quality Data Care Gap Index % Efficiency Indices Cost Index (Actual Costs/Predicted Costs based on % Relative Risk Score) Admission Utilization Index (Actual Admits/Predicted Admits) % ED Utilization Index (Actual/Predicted ED Visits) % The national median composite quality score for MSSP is 84% to 86%, while ours is 91.73%. We attribute this to establishing personal relationships with both our PCPs and our highest risk clients to gain their trust and cooperation with the process measures and action plans that we recommend. Lessons Learned: Committed business leaders and passionate clinical leaders are keys to our success in the ACO model. We have maximized the support we deliver through our system EHR, but have been willing to look to other more specialized software for unique aspects of the project. Knowing that care coordination across many locations of care and many EHRs is the key, we have learned to use a suite of IT tools for care management, client tracking, and condition-based registry management. Our next step is to apply lessons learned to other applicable populations. Financial Considerations: Leaping into the world of health care financing by ACO is not for the faint of heart. Centura has invested heavily in designing the processes to support these activities. Our annual operating cost for four FTEs is roughly $500,000. Annual maintenance fees for software tools are about $250,000. Changes to our system EHR and content development required about 100 hours of planning and consensus building, and reporting development cost approximately $11,000. We are beginning to see some payoff, as evidenced by the following: PQRS payment incentives: Our Centura owned group practice (CHPG) will receive approximately $71,000 in PQRS incentive payments this fall for CY Since the ACO project files GPRO on behalf of CHPG, there will be no payment adjustments for CY15 forward. For one of our private insurance ACOs, this program assisted in raising our PMPM from $2.50 to $7 last year. For another of our private insurance ACOs we were successful in reducing total cost of care by 6 percent. In our self-insured program, we have generated a $1M reduction in total cost of care for clients with frequent ED use. For CAC, we generated a $1.7M reduction in total cost of care for the chronic disease population. Conclusion: Centura Health has designed two models for managing the health of a population. One is tailored to a large metropolitan area and one to a more rural and geographically isolated area. Both Page 15 of 16

16 contain key elements, namely close knit relationships with our PCPs, an aspect of risk that requires a personal conversation with the client or patient; active care management using longitudinal relationship building with the highest risk individuals; engaging our patients in their own care using goal setting and motivational interviewing; making the most of our existing IT tools and EHR for care and data management; and working with partners such as insurers, HIEs, and government agencies with a shared vision and goal. The major difference between the two projects is that in the metropolitan ACO model we needed a method to identify and follow clients that had not yet had an encounter with any part of our health care system; this necessitated using a suite of software tools. Using the HIMSS Steps model of demonstrating HIT value, we can attest to improvements and enhancements in each area: Satisfaction: We have improved patient satisfaction by improving communication with our patients and among our caregivers. Additionally, we have improved provider satisfaction by establishing personal relationships and developing clear expectations and toolkits to help providers be successful. Treatment/Clinical: We have demonstrated improved clinical metrics in both models, such as: improved blood pressure control, improved compliance with disease specific interventions and monitoring, or improved preventive surveillance. Electronic information/data: We have accomplished population health reporting and improved quality measures reporting and are learning how to aggregate and format the data to effectively communicate with our caregivers in both projects. Prevention and Patient Education: We have developed effective patient and provider education and have demonstrated improved condition specific surveillance and improved patient compliance with care recommendations in both projects. Savings: We have reduced unplanned ED visits and unplanned hospitalizations in both projects. Although we have achieved gratifying success in both projects, we have much work to do to optimize these services. In addition to those activities, our next tasks include planning for sustainable and efficient funding and staffing models so that we can apply our lessons learned to nurture the health of people in all of our communities. Page 16 of 16