Paying Medi-Cal Managed Care Plans for Value: Quality Goals for a Financial Incentive Program APRIL 2019 Contents About Bailit Health 3Executive Summary Bailit Health Purchasing, LLC (Bailit Health) is a health policy consulting firm dedicated to 5Introduction and Purpose ensuring insurer and provider performance 5Approach and Methodology accountability on behalf of public agencies. The Measure Selection Process firm primarily works with states to take actions that positively influence the performance of the 0Performance Measure Set Recommendations 1 health care system and support achievement of measurable improvements in health care quality 2Performance Evaluation Methodology 1 and cost management. Design Considerations and Recommendations For more information, visit 6Next Steps and Key Considerations for 1 www.bailit-health.com. Implementation 17 Appendices About the Foundation A. dvisory Group Members A The California Health Care Foundation is B. Comparison of DHCS EAS and CHCF Advisory Group dedicated to advancing meaningful, measur- Measure Selection Criteria able improvements in the way the health care delivery system provides care to the people of C. High-Opportunity External Accountability Set Measures California, particularly those with low incomes for Rate Year 2019 and those whose needs are not well served by D. Measures Considered by the Advisory Group and the status quo. We work to ensure that people Summary of Discussions have access to the care they need, when they E. Recommended Measure Set, by Domain and need it, at a price they can afford. Population Age Group CHCF informs policymakers and industry lead- 25 Endnotes ers, invests in ideas and innovations, and connects with changemakers to create a more responsive, patient-centered health care system. For more information, visit www.chcf.org. California Health Care Foundation 2 Executive Summary A cross the country, many states are establish- $$ Have systemic impact on health if performance ing clear performance expectations for their improves Medicaid managed care plans (MCPs) and $$ Be outcome-based, preferably adopting financial incentives tied to quality of care and other measures of performance.1 California is not one $$ Be pertinent to the Medi-Cal population of them, despite an abundance of poor MCP scores on $$ Be feasible to collect with existing infrastructure many measures of quality and consumer experience.2 Moreover, when Medi-Cal MCPs are able to reduce the $$ Align with other measures currently in use in cost of care, the state reduces their capitation rate — a California, with special attention to measures in phenomenon known as “premium slide” — even if they the Department of Health Care Services’ External have improved quality of care and made health-related Accountability Set (EAS)4 investments to address social determinants affecting individuals and communities. After considering hundreds of measures and weighing options for the size of the measure set, 12 measures were A 2018 report from the California Health Care Foundation selected. The 12 measures span six domains: preventive (CHCF) provided a recommendation for a performance care / early detection, care coordination, chronic illness incentive program that would address premium slide in care, maternity care, medication management, and Medi-Cal managed care.3 This report picks up where that patient experience (Table 1). one left off. Specifically, this report provides recommen- dations for a measure set and performance evaluation Table 1. Recommended Measure Set, by Domain methodology to encourage improvement in the quality of care provided to Medi-Cal beneficiaries by MCPs. These MEASURE recommendations were developed over a series of four Preventive Care / $$ Breast Cancer Screening meetings with an Advisory Group representing a diverse Early Detection $$ Cervical Cancer Screening array of Medi-Cal stakeholders, including Medi-Cal MCP $$ Childhood Immunization Status – leaders, consumer advocates, provider representatives, Combo 3 or 10* and other experts listed in Appendix A. $$ Chlamydia Screening $$ Immunizations for Adolescents – Combo 2 Recommended Performance Care Coordination Plan All-Cause Readmissions Measures $$ Before selecting a set of performance measures for a Chronic Illness Care $$ Controlling High Blood Pressure financial incentive program, the Advisory Group estab- $$ Comprehensive Diabetes Care: lished criteria after considering those adopted by national HbA1c Poor Control bodies and other states. They agreed that measures Maternity Care $$ Cesarean Rate for Nulliparous incorporated into a financial incentive program should: Singleton Vertex Birth $$ Be meaningful to patients $$ Prenatal and Postpartum Care $$ Be meaningful to providers Medication $$ Asthma Medication Ratio Management $$ Be amenable to plan or provider influence Patient Experience $$ Consumer Assessment of $$ Represent an opportunity for improvement Healthcare Providers and Systems (CAHPS) – Rating of Health Plan $$ Be nationally vetted or vetted by a California orga- nization charged with measure development for *Some Advisory Group members recommended Combination 3 because it includes the most important vaccinations, but others favored Combo 10, supporting evidence, validity, and reliability as it is the most complete and is used for NCQA accreditation and widely among other states. On March 7, 2019, after the Advisory Group’s final meeting, DHCS announced its intention to move to Combination 10 for EAS Measurement Year 2019. Paying Medi-Cal Managed Care Plans for Value: Quality Goals for a Financial Incentive Program 3 All of the recommended measures are found within the potential new Department of Health Care Services (DHCS) Performance Evaluation Measurement Year 2019 EAS measure set announced on Methodology Recommendations March 7, 2019, at the Medi-Cal Managed Care Advisory There are many approaches states have taken when Committee meeting. applying performance measures to a financial incen- tive program for Medicaid managed care. The Advisory The Advisory Group also determined that some impor- Group considered and addressed four key design ques- tant areas of Medi-Cal MCP performance that should be tions in constructing its recommended methodology. measured lack measures that meet the selection criteria. Table 2 presents four design questions considered by the It recommended that California take the following steps Advisory Group and the resulting recommendation. to strengthen the incentive measure set: $$ Include a depression measure when a valid and Key Considerations for operationally feasible measure is available. Implementation $$ Include a statin measure once clinical guidelines Adopting financial incentives tied to performance would have stabilized. focus the attention of Medi-Cal MCPs on key state pri- $$ Continue to stratify measurement results by orities and could accelerate the MCPs’ performance subpopulation to identify priorities for reducing improvement efforts and improve health outcomes for disparities. over 10 million Californians enrolled in Medi-Cal man- aged care. The performance evaluation methodology $$ Conduct the CAHPS survey annually. discussed by the Advisory Group and recommended in $$ Incorporate access data from the DHCS timely this report does not benefit from knowing a specific state access survey and/or the California Department of goal or a defined financing method (e.g., from state reve- Managed Health Care (DMHC) access report when nues or from MCP savings or capitation withholds). These there are mature methodologies and available are critical factors, however, and some modifications to benchmarks. the recommendations may be desirable once the spe- cific program goal and financing method are established. Moreover, before the financial program is implemented, California would need to consider any changes in clini- cal guidelines, changes to measure specifications, and changes to measure endorsement from national orga- nizations made between the time of this report and implementation. Table 2. Performance Evaluation Methodology Design Questions RECOMMENDATION Should performance scores be used as a “gate” that a Medi-Cal should use a combined “gate-and-ladder” model for Medi-Cal MCP must pass to qualify for financial incentives, assessing performance for the allocation of incentives. or as a “ladder” in which the state tiers its financial rewards based on level of MCP performance? Should California evaluate MCP performance for high Medi-Cal should reward both high achievement and improve- achievement, improvement over time, or performance ment over time. superior to the competition? How high must be performance to be evaluated positively? When awarding achievement, Medi-Cal should set targets that are measure-specific. For improvement, Medi-Cal should set targets at an achievable level annually so that plans have a meaningful incentive to generate ongoing improvement. Should DHCS weight some measures more than others? Medi-Cal should give all measures equal weight for the purpose of allocating incentives. California Health Care Foundation 4 Introduction and Purpose Medi-Cal has many of the necessary building blocks for a financial incentive program tied to MCP performance, California has the largest Medicaid managed care pro- including collection of a robust set of access, quality, and gram in the country by far. With over 10 million enrollees, patient experience measures, and a Medi-Cal managed it is twice the size of the next largest Medicaid managed care performance dashboard.7 In all counties where ben- care program. Close to 90% of Medi-Cal enrollees with eficiaries have a choice of two or more plans, Medi-Cal full-scope coverage, and one in four of all Californians, uses six performance measures to assign beneficiaries to get their care from a Medi-Cal managed care plan. Yet a plan if they do not choose one themselves. A financial despite the importance of Medi-Cal managed care to incentive program would be a positive next step, one the people of California, quality of care is highly variable already taken by many other state Medicaid programs among Medi-Cal managed care plans (MCPs), and con- and by commercial and Medi-Cal MCPs in California sumer satisfaction routinely ranks well below the national that operate provider incentive (pay-for-performance) average. For example, in 2017, on average across all programs. Medi-Cal MCPs, nearly half (46%) of women did not receive their recommended cervical cancer screening. The Department of Health Care Services (DHCS) uses Approach and many tools available to manage MCPs, such as competi- tive procurement, contract management and oversight, Methodology public reporting, and penalties. However, one tool that CHCF convened an Advisory Group, and Bailit Health many other states use that California does not is financial facilitated a series of four meetings between October 26, incentives tied to quality and other measures of Medi-Cal 2018, and February 1, 2019, to develop the measure set MCP performance.5 Moreover, when Medi-Cal MCPs are and performance evaluation methodology. The Advisory able to reduce the cost of care, the state reduces their Group included a mix of health plan, provider and capitation rate — a phenomenon known as “premium consumer representatives, along with technical measure- slide” — even if they have improved quality of care and ment experts (Appendix A). made health-related investments to address social deter- minants affecting individuals and communities. The role of the Advisory Group was to advise on key ele- ments of the performance measure set and performance In April 2018, the California Health Care Foundation evaluation methodology. Each member was encouraged (CHCF) published Intended Consequences: Modernizing to offer ideas, provide feedback, and express prefer- Medi-Cal Rate Setting to Improve Health and Manage ences. The Advisory Group members were not expected Costs,6 which recommends a gain-sharing approach to reach consensus and, as such, Advisory Group recom- that would, if adopted, establish positive performance mendations presented in this report do not imply that incentives for improving quality and reducing the cost full consensus was reached. Nonetheless, the members of care. The report did not, however, recommend which of the Advisory Group found shared agreement on most specific performance measures should be used and how points. they should be used. Picking up where that report left off, CHCF hired Bailit Health, formed an Advisory Group, An incentive program design should reflect the goal of and charged that body with developing a recommended the incentive program and the financing method. For performance measurement set and assessment strategy example, a performance incentive program designed to for a financial incentive program for Medi-Cal MCPs. ensure that quality is acceptable before shared savings are distributed may have different design characteristics This report presents the approach and outcomes of that than one that allocates bonus dollars for high achieve- project. It provides a path forward for California and dem- ment. The evaluation methodology discussed by the onstrates that a diverse group of stakeholders — one that Advisory Group and recommended in this report does included MCP and provider representatives, consumer not benefit from a specific goal or financing method. advocates, and other experts — could coalesce around Once state officials finalize those decisions, some modi- a shared vision of the measurement set criteria, size, and fication to the recommendations presented in this report measures. may be desirable. Nonetheless, these recommendations Paying Medi-Cal Managed Care Plans for Value: Quality Goals for a Financial Incentive Program 5 should generally prove themselves robust whatever goal These criteria generally align with the National Quality and financing decisions are reached in the future. Also, Forum’s criteria, with a few exceptions. Criteria eight because the goal was to develop recommendations that and ten listed above are specific to Medi-Cal. Another would undergo further review by state officials and other reflects the Advisory Group’s sentiment that it is impor- stakeholders, and not to design a methodology for date- tant to increase use of outcome measures even though certain implementation by DHCS, the Advisory Group most measures in use today are process measures. These did not develop recommendations at the level of detail criteria also align with the newly released DHCS goals for that will be required for implementation. its EAS. A full comparison of the EAS and Advisory Group measure selection criteria can be found in Appendix B. Measure Selection Process Domains and Populations The measure selection process involved seven steps: The Advisory Group was asked to identify and priori- (1) define the selection criteria, (2) identify domains and tize performance domains and Medi-Cal populations it populations, (3) identify measure sources, (4) identify wanted represented in the measure set recommenda- data sources and means to acquire data, (5) estimate the tions. It was given a comparative analysis of domains desired measure set size, (6) select the measures, and used in four other states (Massachusetts, Oregon, (7) refine the measure set. Each step is described below Rhode Island, Washington) in their measure set develop- with corresponding recommendations from the Advisory ment processes and those employed by the Integrated Group. Healthcare Association (IHA). The Advisory Group also considered the five program goals established by DHCS Selection Criteria in its Medi-Cal Managed Care Quality Strategy Report: The Advisory Group established measure selection crite- maternal and child health, chronic disease, tobacco ces- ria that served as parameters for deciding which measures sation, reducing health disparities, and fostering healthy should be included and excluded from the measure set. communities through reducing opioid misuse and over- They were selected following consideration of criteria use.8 These priorities were mapped to the candidate adopted by national bodies and by other states. The measures for consideration. Advisory Group recommended that measure selected for a financial incentive program for Medi-Cal MCPs should: The performance domains that received the most sup- port were patient experience, preventive care  / early 1.Be meaningful to patients detection, access, social determinants of health, care 2.Be meaningful to providers coordination, and chronic illness care. The Advisory Group subsequently elected to add maternity care as a 3.Be amenable to plan or provider influence domain given the large number of births for which Medi- 4.Represent an opportunity for improvement Cal is responsible, and medication management, due to interest in inclusion of an asthma treatment–related 5.Be nationally vetted or vetted by a California measure. organization charged with measure development for supporting evidence, validity, and reliability The Advisory Group also recommended that DHCS con- 6.Have systemic impact on health if performance tinue its efforts to measure and reduce health disparities improves and that equity be considered throughout performance measurement. It recommended that DHCS conduct sub- 7.Be outcome-based, preferably population analysis as an ongoing practice. This was 8.Be pertinent to the Medi-Cal population recommended in lieu of a single statewide disparity measure because regional analysis was reported to have 9.Be feasible to collect with existing infrastructure revealed that disparities differ in nature across California 10. Align with other measures currently in use in counties. California, with special attention to measures in DHCS’s External Accountability Set (EAS) California Health Care Foundation 6 Measure Sources The Advisory Group sought to align its recommended IHA Measure Set and Related Activity measure set with other measures currently in use in The nonprofit Oakland-based Integrated Healthcare California. Measure set alignment helps focus plan and Association (IHA) has developed both a recom- provider improvement efforts on high priorities and mended Medi-Cal measure set and a performance reduces some of the administrative burden associated incentive methodology for voluntary adoption by with reporting and acting upon performance measures. MCPs with their network providers. The “Align. Measure sets recommended by one or more Advisory Measure. Perform.” (AMP) program’s recommended Group members included the following: Medi-Cal managed care measure set focuses on clinical quality, patient experience, utilization, $$ Core Measure Sets jointly developed by the and cost of care measures. IHA collects data from Centers for Medicare and Medicaid Services selected Medicaid providers and calculates their (CMS) and America’s Health Insurance Plans (AHIP) performance.11 IHA also created a shared savings model for use by Medi-Cal managed care plans with $$ CMS Medicaid Adult Core Set their providers. This value-based incentive design recommends payment based on quality, cost, and $$ CMS Medicaid and Children’s Health Insurance resource use for physician organizations.12 Program Child Core Set Both the measure set and shared savings model for $$ Covered California Measure Set (plus its disparities Medi-Cal are extensions of IHA’s 15-year program measures) for over 200 physician organizations and 10 health $$ DHCS EAS for MCPs and Specialty Health Plans9 plans serving commercial HMO enrollees. The AMP Commercial HMO program is one of the nation’s $$ DHCS Managed Care Performance Monitoring largest and longest-running alternative payment Dashboard Report models and serves nearly 10 million enrollees. Provider organizations serving both Medi-Cal and $$ IHA“Align. Measure. Perform.” Measure Set — commercial HMO members benefit from a common Medi-Cal Managed Care measure “superset” that is continuously evaluated $$ Medi-Cal Managed Care and Mental Health Office and maintained by active and regular participation of health plans and provider organizations partici- of the Ombudsman pating in the program with standing academic, $$ California Department of Social Services regulator, and accreditation organizational support. (CDSS) Continuum of Care Reform Mental Although this project differs from IHA’s work in Health Care Measures10 important ways — namely, that this project is $$ CDSS Medi-Cal State Hearing Data Statistics directed at incentives from a state purchaser (DHCS) to MCPs, whereas IHA’s work focuses on incentives $$ National Quality Forum (NQF) disparities measures from MCPs to their network providers — there are, nevertheless, benefits to alignment and impor- Data Sources tant lessons to be learned from IHA’s experience. There were four primary data sources for the measures Materials and feedback provided by IHA informed found within the above-named measure sets: clinical the development of the recommended measure set and performance evaluation methodology for this data, claims or encounters, nonclaims administrative project. data, and survey data. Although data availability is often a significant constraint on measure options, the Advisory Group did not recommend eliminating from consider- ation any of the four data sources. Paying Medi-Cal Managed Care Plans for Value: Quality Goals for a Financial Incentive Program 7 Measure Set Size incentives tied to performance were just one of several The Advisory Group was asked to consider three options ways DHCS should manage MCP performance, and that for the size of the measure set (Table 3). additional performance measures would continue to be part of other improvement efforts, such as MCP-specific accreditation work, public reporting of EAS measures, Table 3. Measure Set Size Options statewide and plan-specific quality improvement proj- CONSIDERATIONS ects, and penalties for very poor performance. 5–10 measures $$ Would focus MCP improvement efforts in highest priority areas, particularly Analysis of Candidate Measures if aligned with measures used in Bailit Health analyzed over 200 measures from the ten Medi-Cal auto-assignment measure sets identified by the Advisory Group, and $$ Would not allow for inclusion of measures in all domains of interest then winnowed down the list of individual measures to be considered by the Advisory Group based on two 12–15 measures $$ Would allow inclusion of 1–2 measures considerations: in each domain of interest $$ Measures appearing in two or more measure sets. $$ Would maintain some focus on priorities, but less focus than smaller Bailit Health sorted all measures by domain and measure set calculated the number of measure sets within which each measure appeared. For most domains, Bailit 20–25 measures $$ Would allow inclusion of 2–3 measures in each domain Health only selected measures appearing in two or $$ Would signal an expectation that steps more sets. For domains with measures found in only should be taken to increase perfor- one set, Bailit Health included all measures. mance across the board rather than focused on narrow set of measures $$ High opportunity for improvement. Bailit Health reviewed Medi-Cal MCP performance on EAS measures appearing in at least one other measure It is important to consider the purpose and use of the set of interest to the Advisory Group to determine measure set when considering its size. Anticipating at the which measures had the highest opportunities for time that the measure set was to be used as a “gate” improvement.13 High-opportunity areas were those in which performance qualifies eligibility for shared sav- for which a measure has (1) a low statewide average ings, the Advisory Group recommended a more limited score, defined as weighted performance below the measure set of approximately 5 to 10 measures. The Healthcare Effectiveness Data and Information Set Advisory Group subsequently accepted that this same (HEDIS)14 HMO 50th percentile, and (2) significant size measure set could be used in a manner other than variability among plans, defined as a greater than 15 as a “gate,” such as a ladder system in which bonus pay- percentage point difference between the plans with ments are tiered according to level of performance or the third-lowest and third-highest scores. Additional performance improvement. data on high-opportunity EAS measures can be found in Appendix C. Even among Advisory Group members who agreed on a smaller measure set, some expressed concern with limiting the size of the measure set. For example, one member noted that given the breadth and depth of the Medi-Cal population, he was unsure how to limit the size of the measure set while retaining the measures most appropriate for each given subpopulation. Overall, however, the Advisory Group recognized that financial California Health Care Foundation 8 Consideration of Candidate Measures Refinement The Advisory Group considered candidate measures After conducting its initial review of candidate measures, from the subset identified by Bailit Health. They were the Advisory Group assessed the draft recommended then invited to submit “write-in” measures — that is, measure set. The Advisory Group considered: measures of high interest that had not yet been con- $$ Gaps by measure domain or population sidered by the group. Altogether, the Advisory Group age group discussed 43 measures. A summary of the outcome of this discussion is provided in Appendix D. $$ The size of the draft recommended set $$ Whether it wanted to reconsider any of the During its review, the Advisory Group applied the endorsed measures selection criteria described above and was particularly favorable toward measures for which: $$ How well the measures met the measure selection criteria $$ There was great performance variability or significant room for improvement The Advisory Group reviewed the 12 measures it initially $$ Improvement would have a significant impact recommended for further consideration and decided on patient health to endorse all measures for recommendation. These 12 measures represent a slightly larger set than the Advisory $$ Improvement would affect a large Medi-Cal population Group’s earlier recommendation of 5 to 10 measures. $$ Data were already being reported to DHCS as part of the EAS As the Advisory Group reviewed candidate measures, it identified a few measures that were not selected for the incentive measure set but which the group believed were of high importance and of value to include in measure sets used for other purposes, such as oversight, public reporting, and identifying statewide and plan-specific performance improvement projects. These measures are noted in Appendix D. Paying Medi-Cal Managed Care Plans for Value: Quality Goals for a Financial Incentive Program 9 Performance Measure 10 addressing adults (Appendix E). The populations Set Recommendations between these categories are not mutually exclusive: Asthma Medication Ratio, for example, reflects care pro- Table 4 lists the performance measures that the Advisory vided to children, adolescents, and adults. Group recommended DHCS consider using if California adopts a financial incentive program for Medi-Cal MCPs An analysis of Medi-Cal MCP scores for the recommended tied to performance. measures demonstrates the significant opportunity for improving health care quality and outcomes for California The recommended measure set contains 2 measures Medi-Cal beneficiaries (Figure 1, page 10). addressing children, 5 addressing adolescents, and Table 4. Performance Measures Recommended for Financial Incentive Measure Set, by Domain MEASURE (NQF NUMBER) RATIONALE FOR SELECTION Preventive Care / $$ Breast Cancer Screening Screening impacts a large population, there is opportunity for Early Detection (2372) improvement, and screening has a direct impact on mortality. $$ Cervical Cancer Screening Performance is poor and there are disparities in performance. (0032) $$ Childhood Immunization Status – Immunization is an important aspect of pediatric care. Combo 3 or 10* (0038) $$ Chlamydia Screening The measure focuses on reproductive-age women potentially (0033) experiencing domestic violence. $$ Immunizations for Adolescents – Concrete positive impact on outcome, with low median Combo 2 performance and high variation by plan. (1407) Care Coordination $$ Plan All-Cause Readmissions A measure involving hospitals is important because they are a signifi- (1768) cant part of the care delivery system and readmissions are costly. Chronic Illness Care $$ Controlling High Blood Pressure Performance has a high impact on morbidity and mortality, and there (0018) is room for improvement. $$ Comprehensive Diabetes Care: Clinically meaningful and high variability in performance. The HbA1c Poor Control (>9%) Advisory Group selected this measure over the HbA1c Control (0059) (<8.0%) measure because diabetes complications increase dramatically around 9%. Maternity Care $$ Cesarean Rate for Nulliparous An important measure with opportunity for improvement. Singleton Vertex Birth (0471) $$ Prenatal and Postpartum Care Important measures of access, affecting a large Medi-Cal population, (1517) with room for improvement and disparities in performance. Medication $$ Asthma Medication Ratio Asthma is an important issue for the Medi-Cal population, and this is Management (1800) the only chronic illness measure included for children and adolescents. Patient Experience $$ CAHPS – Rating of Health Plan CAHPS is currently the only standardized measure for patient (0006) experience. *Some Advisory Group members recommended Combination 3 because it includes the most important vaccinations, but others favored Combo 10, as it is the most complete and is used for NCQA accreditation and widely among other states. On March 7, 2019, after the Advisory Group’s final meeting, DHCS announced its intention to move to Combination 10 for EAS Measurement Year 2019. California Health Care Foundation 10 Figure 1. Medi-Cal Managed Care Plan HEDIS Scores for Selected Recommended Measures CHILDHOOD IMMUNIZATION STATUS CERVICAL CANCER SCREENING COMBINATION 3 100 100 Percentile Percentile 90 90 Percentile 90th 80 80 50th 90th 70 70 25th 60 60 50th 50 25th 50 40 40 30 30 20 20 10 10 0 0 COMPREHENSIVE DIABETES CARE CONTROLLING POOR HbA1c CONTROL (>9.0%)* HIGH BLOOD PRESSURE POSTPARTUM CARE 100 100 100 90 90 90 Percentile Percentile Percentile Percentile 80 80 80 90th 90th 70 70 70 Percentile 50th 60 60 60 50th 25th 25th 50 50 25th 50 50th 40 40 40 30 90th 30 30 20 20 20 10 10 10 0 0 0 * Indicates measures in which lower scores reflect better quality. In all other cases higher scores reflect better quality. Notes: HEDIS is the Healthcare Effectiveness Data and Information Set. Each dot represents one plan. Percentile notations are national rankings. Sources: The source for HEDIS® Medicaid benchmark data contained in this publication is Quality Compass® 2017 and is used with the permission of the National Committee for Quality Assurance (NCQA). Any data display, analysis, interpretation, or conclusion based on these data is solely that of the authors, and NCQA specifically disclaims responsibility for any such display, analysis, interpretation, or conclusion. Quality Compass is a registered trademark of NCQA. The HEDIS® Medicaid 25th, 50th and 90th percentiles reflect the measurement year from January 1, 2016 through December 31, 2016. Data are from Managed Care Quality and Monitoring Division California Dept. of Health Care Services, Medi-Cal Managed Care External Quality Review Technical Report: July 1, 2016 – June 30, 2017 (April 2018), www.dhcs.ca.gov (PDF). Paying Medi-Cal Managed Care Plans for Value: Quality Goals for a Financial Incentive Program 11 Reflecting on key gaps in the list of 12 measures, includ- ing the absence of any measures from the access domain, Performance Evaluation the Advisory Group recommended that DHCS make the Methodology following future enhancements to the measure set: Integration of performance measurement with pay- $$ Include a depression measure when a valid and ment creates an economic motivation for managed care operationally feasible measure is available. plans to make significant investments in targeted, high- priority areas. These investments hold the potential to $$ Include a statin measure once clinical guidelines improve the quality and outcomes of care for Medi-Cal have stabilized. beneficiaries. $$ Continueefforts to stratify measurements by subpopulation. The means by which performance on selected measures is translated into financial consequences requires careful $$ Collect CAHPS survey data every year so consideration; poorly constructed evaluation method- that health plans can better understand their ologies will not motivate plan investment or influence performance. measurable improvement. This section of the report $$ Use the DHCS timely access survey and DMHC reviews key design considerations for a performance access report as potential data sources in the evaluation methodology and examples from other states, future when there are mature methodologies and and then summarizes Advisory Group recommendations available benchmarks. for the Medi-Cal program. The other domain of significant interest to the Advisory Group that is not represented in the recommended Design Considerations and measure set is social determinants of health. While rec- Recommendations ognizing the important role that social determinants play State Medicaid programs have taken many approaches in the health of individuals and communities, Advisory to evaluating health plan performance as a means for Group members noted that Medi-Cal does not cover applying financial rewards. Over time, states have studied services to address social determinants (although many the approaches of their peer states and often borrowed MCPs are drawing from savings to address social deter- approaches for application in their own state. Each state minants in targeted ways) and that there are currently no has considered and addressed the following four key nationally recognized MCP measures for this domain. design questions when constructing its methodology. Should DHCS Use Performance as a Gate to Qualify a MCP for a Financial Reward, Should It Tier Rewards Based on Performance Level, or Both? The Advisory Group was asked to consider three approaches: a qualifying “gate” that ensures that only those MCPs that meet specific performance expecta- tions qualify for financial incentives; a tiered “ladder” where the amount of the financial incentive increases as performance increases; or a combination of the two. Maryland uses a qualifying gate; Texas and the District of Columbia (see Table 5, page 13) both use a gate and a ladder.15 In some cases, the ladder can extend “below ground,” with poor or deteriorated performance gener- ating a financial penalty or an offset to rewards earned on other measures. Texas’s ladder methodology operates with increasing penalties for poor performance. California Health Care Foundation 12 Table 5. District of Columbia “Gate-and-Ladder” Incentive Design EARNED INCENTIVE (REDUCTION COMPARED TO BASELINE) WEIGHT QUALITY MEASURE (OUT OF 100%) <2% 2% 3.5% 5% Potentially Preventable Admissions 33% 0% 50% 75% 100% Low-Acuity Non-Emergent Emergency Department (ED) Visits 33% 0% 50% 75% 100% 30-Day All-Cause Readmissions 34% 0% 50% 75% 100% Source: Medicaid Managed Care: 2017 Annual Technical Report, District of Columbia Dept. of Health Care Finance, April 2018, dhcf.dc.gov. The Advisory Group recommended that DHCS use performance and limited or no opportunity for further an incentive structure with both a gate and a ladder. It improvement may find there is no opportunity to receive believed that DHCS should determine if performance financial incentives despite their high performance. is adequate to qualify a MCP for an incentive and also assess achievement at one or more tiers above the quali- Table 6. Tennessee “Improvement” Example fying gate to provide heightened incentive and rewards for superior performance. The gate would ensure that BASELINE RATE MINIMUM EFFECT SIZE MCPs with poor performance are not rewarded, and the 0–59 At least a six percentage point change tiers (ladder) would provide an incentive for MCPs whose performance exceeds the gate to achieve higher levels 60–74 At least a five percentage point change of performance. 75–84 At least a four percentage point change Should DHCS Evaluate MCP Performance 85–92 At least a three percentage point change for High Achievement, Improvement, 93–96 At least a two percentage point change Performance Superior to the Competition, or Some Combination Thereof? 97–99 At least a one percentage point change The first impulse for many states designing MCP per- formance incentive programs is to assess MCPs against Source: Contractor Risk Agreement Between the State of Tennessee and Volunteer State Health Plan, Inc., Blue Cross Blue Shield of Tennessee, high achievement16 standards or benchmarks, reasoning accessed February 13, 2019, www.bcbst.com (PDF). that only those performing well should receive a reward. There are two significant limitations to such an approach: Among MCPs that were high performing prior to creation A third option is a combination of the two: evalu- of the incentive, the financial incentives would have no ate contracted MCPs for both high achievement and impact on their motivation to improve their performance; improvement. Oregon, Texas, and Washington have all and low performers won’t invest resources to improve adopted this option.19 their performance if they find the high achievement stan- dard or benchmark unattainable. Nevertheless, some A fourth approach, used by Arizona, is to assess MCP per- states — including Maryland and New York — use this formance relative to that of plan competition in the state.20 approach.17 This approach creates winners and losers, even when multiple MCPs excel or improve, and can diminish moti- Another option, employed by Tennessee, is to reward vation to improve. Consequently, it is used less often by MCPs only for performance improvement18 (Table 6). states. It is, however, the approach DHCS employs for its This entails comparing a MCP’s own performance in a performance-based auto-assignment algorithm in which preceding year, or perhaps two preceding years, to the Medi-Cal beneficiaries who don’t choose a MCP them- most recent performance period. Yet assessing improve- selves are assigned to a plan by DHCS.21 In that context, ment alone also has a limitation: MCPs with high baseline it makes sense, as members needing to be assigned to a Paying Medi-Cal Managed Care Plans for Value: Quality Goals for a Financial Incentive Program 13 plan is a “fixed pie” that needs to be divided up among The Advisory Group recommended that DHCS reward participating MCPs that operate in the region where the both high achievement and improvement over time. It members live. By contrast, financial incentives do not thought that the goal of a value-based purchasing strat- need to be allocated from a predefined pool of funds. egy is to improve value over time22 and that DHCS should reward improvement as highly as it does high achieve- There are other important considerations when decid- ment. California needs its lowest performing Medi-Cal ing whether and how to evaluate MCP performance MCPs to improve in order to lift the performance of the for achievement, improvement over time, performance entire Medi-Cal program and to close gaps in health superior to the competition, or some combination equity. DHCS may need to take time to explain this con- thereof. These include the following: cept to key stakeholders, including legislators, as it is not always intuitive why a state would want to financially $$ Should DHCS assess whether there has been reward what appear to be poorly performing MCPs. The deterioration in performance over time for any Advisory Group also thought that DHCS should consider of the measures? States sometimes evaluate negative adjustments for MCPs with evidence of deterio- deterioration and adjust incentive rewards so that rated performance, as is DHCS’s current practice with its MCPs aren’t financially rewarded when perfor- performance-based auto-assignment model. mance deteriorates — for example, preventing a plan that improves on one measure but declines How High Must a MCP’s Performance Score on three others from being rewarded for the one to Be Evaluated Positively? improved measure, or preventing the allocation Determining what performance scores define high of an incentive to a plan that improved modestly achievement or how large an improvement must be after declining precipitously the preceding year. for incentive allocation is critical to a successful incen- $$ IfDHCS rewards a combination of high achieve- tive design. If the bar is set too low, the performance ment, improvement, and/or competitive incentive may have no impact on plan behavior or perfor- superiority, should one be rewarded more highly mance, except perhaps for the very poorest performers. than another? The impact may be similar if the bar is set too high and MCPs gauge that they cannot achieve the necessary level $$ IfDHCS assesses performance achievement, of performance for the reward or that the effort required should it utilize national benchmarks (if available to attain the high achievement target level is too great for a given measure), state benchmarks, or abso- relative to the available plan resources or the size of the lute values not pegged to a benchmark? States potential reward. often use national benchmarks from the National Committee for Quality Assurance (NCQA) for their Bailit Health recommended that states adopting financial HEDIS measures and state benchmarks for non- incentives tied to MCP performance should consider the HEDIS measures. following: $$ Ifassessing performance improvement, should $$ Ifadopting high achievement targets, set those DHCS define improvement in absolute terms targets on a measure-by-measure basis and at lev- (e.g., four percentage points) or in statistical els that are reasonably attainable for at least some terms (e.g., statistically significant improvement MCPs, so that they are motivated to reach them. at p ≤ .05)? Use of absolute terms is simpler to administer, but is far less precise and fair than $$ If adopting improvement targets, set the required statistical testing. improvement percentages at levels that seem reasonable and attainable, even if not statistically If assessing performance improvement, at what level significant. Doing so should provide MCPs with should additional improvement no longer be expected sufficient motive to invest effort in improvement. due to high achievement and/or diminution of It is exceedingly difficult to attain statistically opportunity? significant improvement year after year, and steady progress in smaller increments will produce statistically significant improvement over time. California Health Care Foundation 14 New York’s high achievement target value of the national $$ Set high achievement target values at either the 90th percentile (Table 7) would be too high for California 66th, 75th, or 90th percentile level depending for some measures, as performance on quality measures in upon baseline performance for a given measure. the nation’s northeast is generally higher than in California. The high achievement benchmark should be Medi-Cal MCP performance varies relative to national above the performance of nearly all MCPs. benchmarks when reviewed across a broad array of qual- $$ If it is less than the 90th percentile, raise the high ity measures, so it may not be appropriate for DHCS to set achievement target periodically if a considerable a single value (if expressed as a percentage of the national percentage of plans meet or exceed it. average as New York does) across all measures. $$ Set the improvement target value for each mea- sure at two or three percentage points, depending Table 7. New York “High Achievement Level” Example, with Tiers upon the proximity of general MCP prior-year performance to the high achievement target. PLAN PERFORMANCE (HEDIS BENCHMARKS) POINTS EARNED Two or three percentage points is unlikely to be significant improvement in a given year but will be <50th percentile 0 cumulatively over time. 50th to <75th percentile 50% of possible points Should DHCS Weight Some Measures More 75th to <90th percentile 75% of possible points Than Others? States sometimes weight some performance measures 90th+ percentile 100% of possible points in their MCP financial incentive program more than oth- Source: 2017 Quality Incentive for Medicaid Managed Care Plans, ers. Weighting certain measures more highly is expected New York State Dept. of Health, accessed February 13, 2019, to increase motivation to focus MCP investment in the www.health.ny.gov (PDF). related clinical areas. This will, of course, also reduce MCP motivation to attend to other measures. The dif- The Advisory Group recommended that DHCS set ferences in weighting has to be significant to change improvement targets at an achievable level on an annual behavior. basis so that plans have a meaningful incentive to gen- erate ongoing improvement. Some managed care plans Weighting decisions may reflect several factors, such as will make special staff and financial investments to when the state has: improve and sustain performance only if they perceive $$ Explicitly established health priorities for the linked financial rewards to be reasonably attainable. state population, the Medicaid program, or the Medicaid managed care program, and these Inspired in part by the approach used by the Oregon priorities are associated with a subset of the full Health Authority with its contracted coordinated care measure set organizations, the Advisory Group supported adoption of the following approach: $$ Determined that the greatest opportunities for population health impact are associated with a $$ Set the “gate” value for a given measure at no subset of measures lower than the 50th percentile level, whether using NCQA national benchmarks for Medicaid man- $$ Identified where the gap between current and aged care or state-level Medi-Cal benchmarks. target performance is greatest (On March 7, 2019, DHCS announced that it $$ Determined that more effort (or cost) is required intended to raise its Minimum Performance Level by MCPs to improve performance on some mea- expectations for MCPs from the 25th percentile sures more than others level to the 50th percentile level for Medicaid plans in the US where that information is available Another reason to weight measures differently is when and the services are delivered by MCPs.) the balance of measures across domains or populations is uneven, but the state wants to weight each domain or population evenly. Arizona’s approach, summarized in Paying Medi-Cal Managed Care Plans for Value: Quality Goals for a Financial Incentive Program 15 Table 8. Arizona “Measure Weighting” MINIMUM ASSIGNED WEIGHT FOR PERFORMANCE MEASURE PERFORMANCE STANDARD CALCULATING INCENTIVE PAYMENT Adult Measures Emergency Department Utilization ≤ 55 visits/1,000 member months 25% Readmissions Within 30 Days of Discharge ≤11% 25% Child Measures Well-Child Visits: 15 Months 65% 12.5% Well-Child Visits: 3–6 Years 66% 12.5% Adolescent Well-Child Visits: 12–21 Years 41% 12.5% Children’s Dental Visits: 2–21 Years 60% 12.5% Source: Alternative Payment Model Initiative — Strategies and Performance-Based Payments Incentive, Arizona Health Care Cost Containment System, accessed February 13, 2019, www.azahcccs.gov (PDF). Table 8, provides an example. It weights each of the four The project undertaken by CHCF to select a set of perfor- child measures half as much as the two adult measures mance measures to incorporate into a financial incentive so that equal weight is given to MCP performance across program for MCPs shows that a diverse group of Medi-Cal child and adult populations. stakeholders are able to make difficult choices together and reach general agreement on the ideal number of The decision of which measures to weight higher rela- measures, measure selection, and methodology. This tive to other measures is not easily reached, as state staff effort was undertaken, however, with the understanding and external stakeholders typically have varied opinions that the recommendations should be revisited when the on which conditions, populations, and aspects of per- Newsom administration and California legislature are formance warrant greatest attention. For these reasons, ready to move forward with a financial incentive program and also to simplify messaging to the MCPs, the Advisory for Medi-Cal MCPs. Reasons for doing so include: Group recommended consistent weighting across all $$ Clinical guidelines underlying the measures measures. may have changed. $$ Some measures may have lost national NCQA and/or NQF endorsement. Next Steps and Key $$ Measure specifications may have changed. Considerations for Implementation It is also imperative that state officials articulate the goals of their incentive program and the financing method Integrating performance measurement with payment before finalizing the performance evaluation methodol- may create an economic motivation for Medi-Cal MCPs ogy in order to ensure alignment. to make significant investments to improve performance in targeted, high-priority areas. Resulting delivery system changes hold the potential to improve the quality of care and outcomes for Medi-Cal beneficiaries. California Health Care Foundation 16 Appendix A. Advisory Group Members Organization names are included for identification only. Individuals were not required or expected to represent the views of their organizations or association members. MEMBER ORGANIZATION Bill Barcellona, MD America’s Physician Groups Greg Buchert, MD Care 1st Health Plan Sarah de Guia California Pan-Ethnic Health Network Joel Gray Anthem Blue Cross Brad Gilbert, MD Inland Empire Health Plan Giovanna Giuliani California Health Care Safety Net Institute Irina Harvey Department of Managed Health Care Susan Huang, MD Health Plan of San Mateo Kim Lewis National Health Law Program Bob Moore, MD Partnership HealthPlan Linda Nguy Western Center on Law and Poverty Andie Patterson California Primary Care Association Jeff Rideout, MD Integrated Healthcare Association Anthony Wright Health Access California Paying Medi-Cal Managed Care Plans for Value: Quality Goals for a Financial Incentive Program 17 Appendix B. Comparison of DHCS EAS and CHCF Advisory Group Measure Selection Criteria DHCS: EAS FOR MY2020/RY2021 ADVISORY GROUP CRITERIA NUMBER DHCS Goals Meaningful to the public, the beneficiaries, the state, and the MCPs #1 – Be meaningful to patients #2 – Be meaningful to providers Improves quality of care or services for the Medi-Cal population #8 – Be pertinent to the Medi-Cal population High population impact by affecting large numbers of beneficiaries or #6 – ave systemic impact on health if performance H having substantial impact on smaller, special populations improves Known impact of poor quality linked with severe health outcomes #7 – Be outcome-based, preferably (morbidity, mortality) or other consequences (high resource use) Performance improvement needed based on available data demon- #4 – Represent an opportunity for improvement strating opportunity to improve, variation across performance, and disparities in care Evidence-based practices available to demonstrate that the problem is #3 – Be amenable to plan or provider influence amenable to intervention and there are pathways to improvement Availability of a standardized measure and data that can be collected #5 – Be nationally vetted or vetted by a California organization charged with measure development for supporting evidence, validity, and reliability Alignment with other national and state priority areas #10 – A lign with other measures currently in use in California, with special attention to measures in DHCS’s External Accountability Set (EAS) Healthcare System Value demonstrated through cost savings, cost- effectiveness, risk-benefit balance, or health economic benefit DHCS Other Considerations Avoid negative unintended consequences Limiting burden and intrusion on primary care provider offices #9 – Be feasible to collect with existing infrastructure The need to retain measures on the EAS for three years for baseline and trend analysis The impact of adding and deleting measures used in the auto- assignment and default algorithm California Health Care Foundation 18 Appendix C. High-Opportunity External Accountability Set Measures for Rate Year 2019 High-opportunity areas are those for which a measure has (1) a low statewide average score, defined as weighted per- formance below the HEDIS HMO 50th percentile, and (2) significant variability among plans, defined as a greater than 15 percentage point difference between the plans with the third-lowest and third-highest scores. LOW STATEWIDE SIGNIFICANT MEASURE SCORE VARIABILITY Ambulatory Care – Outpatient Yes * Ambulatory Care – Emergency Department Yes * Annual Monitoring for Patients on Persistent Medications – ACE or ARB Yes No Annual Monitoring for Patients on Persistent Medications – Diuretics Yes No Asthma Medication Ratio Yes Yes Avoidance of Antibiotic Treatment in Adults with Acute Bronchitis No Yes Breast Cancer Screening No Yes Cervical Cancer Screening Yes Yes Childhood Immunization Status – Combo 3 Yes Yes Children and Adolescents’ Access to Primary Care Practitioners – 12 to 19 Years Yes Yes Children and Adolescents’ Access to Primary Care Practitioners – 12 to 24 Months Yes No Children and Adolescents’ Access to Primary Care Practitioners – 7 to 11 Years Yes Yes Children and Adolescents’ Access to Primary Care Practitioners – 25 Months to 6 Years Yes Yes Colorectal Cancer Screening n.d. n.d. Comprehensive Diabetes Care: HbA1c Testing Yes No Comprehensive Diabetes Care: Medical Attention for Nephropathy No No Comprehensive Diabetes Care: Blood Pressure Control (<140/90 mm Hg) No Yes Comprehensive Diabetes Care: Eye Exam No Yes Comprehensive Diabetes Care: HbA1c Good Control No Yes Comprehensive Diabetes Care: HbA1c Poor Control No Yes Controlling High Blood Pressure No Yes Immunizations for Adolescents (includes HPV) No Yes Medication Reconciliation Post-Discharge n.d. n.d. Plan All-Cause Readmission n.d. No Prenatal & Postpartum Care – Postpartum Care Rate Yes Yes Prenatal & Postpartum Care – Timeliness of Prenatal Care Yes Yes Use of Imaging Studies for Low Back Pain No Yes Weight Assessment and Counseling for Nutrition and Physical Activity for Children/Adolescents – Nutrition No Yes Weight Assessment and Counseling for Nutrition and Physical Activity for Children/Adolescents – PA No Yes Well-Child Visits in the 3rd, 4th, 5th, and 6th Years of Life No Yes *Performance to benchmark or variability of these measures not considered, as higher or lower rates do not necessarily indicate better or worse performance. Notes: n.d. indicates no data. Depression Screening and Follow-Up for Adolescents and Adults was not included since it did not appear in any sets of interest to the Advisory Group outside of the EAS. Sources: Data used for these determinations were from Medi-Cal Managed Care External Quality Review Technical Report, July 1, 2016 – June 30, 2017, DHCS, April 2018, www.dhcs.ca.gov (PDF). Statewide average scores were pulled from Table 5.7 for Rate Year 2017 (measurement year January 1, 2016 through December 31, 2016). Plan scores used to determine variability were pulled from Appendices B through Q. Data used to determine whether statewide average scores were above of below the HEDIS Medicaid HMO 50th percentile was pulled from NCQA’s Quality Compass HEDIS 2017 (calendar year 2016) data. Paying Medi-Cal Managed Care Plans for Value: Quality Goals for a Financial Incentive Program 19 Appendix D. Measures Considered by the Advisory Group and Summary of Discussions MEASURE NAME (NQF NUMBER) STEWARD RECOMMENDATION DISCUSSION Ambulatory Care National Committee Excluded Utilization measures are not good performance measures, (AMB-OP and AMB-ED) for Quality Assurance as the appropriate utilization can vary greatly by population. (n/a) This measure is also redundant with the goal of establishing performance gates, as savings can be achieved by avoiding ED visits. Annual Monitoring for National Committee Excluded Measure is being retired. Patients on Persistent for Quality Assurance Medications (2371) Antidepressant National Committee Excluded Psychiatric care is carved out of Medi-Cal MCP contracts. Medication for Quality Assurance Management (0105) Appropriate Testing National Committee Excluded This measure is no longer NQF-endorsed and performance for Children with for Quality Assurance is high. Pharyngitis (0002) Asthma Medication National Committee Endorsed Asthma is an important issue for the Medi-Cal popula- Ratio (1800) for Quality Assurance tion, and this is the only chronic illness measure included for adolescents. One Advisory Group member expressed concern about whether the measure was measuring what it intended, as people may have multiple inhalers and those with mild and intermittent asthma do not need a rescue inhaler. Avoidance of Antibiotic National Committee Excluded This measure has coding issues that have led to the inclu- Treatment in Adults with for Quality Assurance sion of patients with chronic obstructive pulmonary disease Acute Bronchitis (0058) and is not a high priority for the Advisory Group. Breast Cancer Screening National Committee Endorsed This measure impacts a large population, there is significant (2372) for Quality Assurance opportunity for improvement, and it has a direct impact on patient mortality. CAHPS Health Plan Agency for Endorsed Advisory Group members noted that CAHPS is collected Survey v5.0 – Rating of Healthcare Research by DHCS every three years. This fact, coupled with report- Health Plan (0006) and Quality ing delays, make it hard to use the CAHPS for incentive purposes. Some Advisory Group members did not like the CAHPS, as they thought results were unspecific, but noted that it is currently the only standardized measure for patient experience. The Advisory Group recommended that DHCS collect the CAHPS every year so that MCPs can better understand their performance and use rating of health plan as a measure of patient experience. Some Advisory Group members recommended including the measure Overall Rating of Health Plan. Cervical Cancer National Committee Endorsed Performance is poor and there are disparities in perfor- Screening (0032) for Quality Assurance mance. Cesarean Rate for The Joint Endorsed Advisory Group members thought that this was an impor- Nulliparous Singleton Commission tant measure with significant opportunity for improvement. Vertex Birth (PC-02) Some Advisory Group members noted that there were (0471) adverse financial incentives to deliver Cesarean sections. California Health Care Foundation 20 MEASURE NAME (NQF NUMBER) STEWARD RECOMMENDATION DISCUSSION Child and Adolescents’ National Committee Excluded The types of encounters captured by this measure are too Access to Primary Care for Quality Assurance broad and the recommended measure set includes other Practitioners (n/a) measures, such as Childhood Immunization Status, that reflect access to care among children and adolescents. Childhood Immunization National Committee Endorsed Advisory Group members thought that immunization status Status – Combo 3 (0038) for Quality Assurance represents an important area of care. Some Advisory Group members recommended Combination 3, as it included the most important vaccinations, but others favored Combo 10 as it is the most complete, and it is used for NCQA accredi- tation and widely among other states. Chlamydia Screening National Committee Endorsed The measure focuses on reproductive-age women poten- (0033) for Quality Assurance tially experiencing domestic violence. Some Advisory Group members recommended against including this measure, as they saw it as a lower priority for them than other endorsed screening measures due to the lower severity of illness. Colorectal Cancer National Committee Excluded This is not a Medicaid measure in HEDIS so no Medicaid Screening (0034) for Quality Assurance benchmarks are available from NCQA. Comprehensive National Committee Excluded This population is captured in the Controlling High Blood Diabetes Care: for Quality Assurance Pressure measure. Blood Pressure Control (<140/90 mm Hg) (0061) Comprehensive National Committee Use in a This is an important measure for diabetes care, as eye Diabetes Care: for Quality Assurance larger set disease needs to be caught early to prevent blindness. Eye Exam (0055) Advisory Group members are interested in tracking how the exam rates are impacted when Medi-Cal restores its vision benefit in 2020. Comprehensive National Committee Excluded Failure to meet this measure has less of a clinical impact for Diabetes Care: for Quality Assurance patients than Comprehensive Diabetes Care: HbA1c Poor HbA1c Control (<8.0%) Control (>9.0%). (0575) Comprehensive National Committee Endorsed Clinically meaningful and high variability in perfor- Diabetes Care: for Quality Assurance mance. The Advisory Group selected this measure over HbA1c Poor Control Comprehensive Diabetes Care: HbA1c Control (<8.0%) (>9.0%) (0059) because the diabetes complication rate increases dramati- cally around 9%. Comprehensive National Committee Excluded Process measure captured in the Comprehensive Diabetes Diabetes Care: for Quality Assurance Care: HbA1c Poor Control (>9.0%). HbA1c Testing (0057) Comprehensive National Committee Excluded Performance is already high, and thresholds keep rising, Diabetes Care: for Quality Assurance making it harder for plans to meet targets even if they are Medical Attention for already performing well. Nephropathy (0062) Concurrent Use Pharmacy Quality Excluded There are data challenges to measurement because the of Opioids and Alliance measure requires pharmacy data. Benzodiazepines (n/a) Paying Medi-Cal Managed Care Plans for Value: Quality Goals for a Financial Incentive Program 21 MEASURE NAME (NQF NUMBER) STEWARD RECOMMENDATION DISCUSSION Contraceptive Care – US Office of Excluded Contraception is an important topic. Advisory Group Most & Moderately Population Affairs members wanted to monitor performance as data become Effective Methods (2903) available. One Advisory Group member expressed concern that the measure did not account for physiological reactions to different types of contraception. Contraceptive Care – US Office of Excluded Contraception is an important topic. Advisory Group Postpartum (2902) Population Affairs members wanted to monitor performance as data become available. One Advisory Group member expressed concern that the measure did not account for physiological reactions to different types of contraception. Controlling High Blood National Committee Endorsed Performance has a high impact and there is room for Pressure (0018) for Quality Assurance improvement. One Advisory Group member noted that one could directly calculate how many lives would be saved by a reduction in blood pressure, and that if he could pick a single measure to include, it would be this one. Developmental Oregon Health & Excluded Providers do not regularly use the code in this measure. Screening in the First Science University Three Years of Life (1448) Diabetes Care for National Committee Excluded No one spoke in favor of the measure. People with Serious for Quality Assurance Mental Illness: Hemoglobin A1c (HbA1c) Poor Control (>9.0%) (2607) Elective Delivery Prior The Joint Use in a Given the low number of elective deliveries, this did not rise to 39 Completed Weeks Commission larger set as a high-priority measure. There is still an opportunity for Gestation (PC-01) (0469) improvement with regard to disparities in care. Follow-Up After National Committee Excluded No one spoke in favor of the measure. Hospitalization for for Quality Assurance Mental Illness (0576) HIV Viral Load Health Resources Excluded No one spoke in favor of the measure. Suppression (2082) and Services Administration – HIV/ AIDS Bureau Immunizations for National Committee Endorsed Concrete positive impact on outcome, low median perfor- Adolescents – for Quality Assurance mance with high variation by plan. Combo 2 (1407) Lead Screening in National Committee Use in a There are low rates of abnormal tests in California, and it Children (n/a) for Quality Assurance larger set was not recommended as a key measure. This measure is already required for Medi-Cal and some Advisory Group members said it would be difficult to provide a rationale for singling out one Early and Periodic Screening, Diagnostic, and Treatment measure over others. Plan All-Cause National Committee Endorsed The Advisory Group thought it was good to include a Readmission (1768) for Quality Assurance measure involving hospitals, since hospitals are a large part of the care delivery system and readmissions are costly. One Advisory Group member noted that it would be impor- tant to evaluate Medi-Cal performance once DHCS adopts the standard methodology used by HEDIS. California Health Care Foundation 22 MEASURE NAME (NQF NUMBER) STEWARD RECOMMENDATION DISCUSSION Prenatal and National Committee Endorsed Important measures of access, large Medi-Cal population Postpartum Care for Quality Assurance with room for improvement, and disparities in perfor- (Timeliness of Prenatal mance. It was noted that the measure lost National Quality Care and Postpartum Forum endorsement because there was no evidence tying Care) (1517) frequency of visits to outcomes despite the consensus on the importance of the visits. Prenatal Immunization National Committee Revisit at a This is a new measure that is still undergoing testing with Status (n/a) for Quality Assurance later time HEDIS and the Integrated Healthcare Association. Proportion of Pharmacy Quality Revisit at a Revisit when there is a clinical consensus (types of statins, Days Covered by Alliance later time dosage, age at which to start are in flux) on appropriate Medications: Statins guidelines. (0541) Screening for Clinical Centers for Medicare Excluded There are new and better measures available to address Depression and & Medicaid Services depression in HEDIS, but plans can’t yet operationalize Follow-Up Plan (0418) them. State Fair Hearings (n/a) California Excluded This is not an actionable measure, since it only looks at Department of Social count of hearings without an indication of the impact of a Services higher or lower count. Statin Therapy National Committee Revisit at a Revisit when there is a clinical consensus (types of statins, for Patients with for Quality Assurance later time dosage, age at which to start are in flux) on appropriate Cardiovascular Disease guidelines. (n/a) Statin Therapy for National Committee Revisit at a Revisit when there is a clinical consensus (types of statins, Patients with Diabetes for Quality Assurance later time dosage, age at which to start are in flux) on appropriate (n/a) guidelines. Statin Use in Persons Pharmacy Quality Revisit at a Revisit when there is a clinical consensus (types of statins, with Diabetes (n/a) Alliance later time dosage, age at which to start are in flux) on appropriate guidelines. Use of Imaging Studies National Committee Excluded The measure fails to capture clinically relevant information. for Low Back Pain (0052) for Quality Assurance Weight Assessment and National Committee Excluded This is a process measure with no evidence of impact on Counseling for Nutrition for Quality Assurance outcome. and Physical Activity for Children/ Adolescents (0024) Well-Child Visits in the National Committee Use in a The measure is important, but of a lower priority. The 3rd, 4th, 5th, and 6th for Quality Assurance larger set measure timing is too late to catch developmental delays, Years of Life (1516) and the measure imposes an artificial deadline for the visits. Paying Medi-Cal Managed Care Plans for Value: Quality Goals for a Financial Incentive Program 23 Appendix E. Recommended Measure Set, by Domain and Population Age Group Below is a breakdown of the recommended measure set by domain. Please note that measures can span multiple domains (e.g., the Prenatal and Postpartum Care measures — Timeliness of Prenatal Care and Postpartum Care — could be considered both preventive care / early detection and maternity care measures). The counts of measures by popula- tion age are not mutually exclusive (e.g., a measure can include both children and adolescents). POPULATION AGE GROUP NUMBER OF MEASURE MEASURES CHILD ADOLESCENT ADULT Preventive Care / Early Detection $$ Breast Cancer Screening 5 1 2 3 $$ Cervical Cancer Screening $$ Chlamydia Screening $$ Childhood Immunization Status – Combo 3 $$ Immunizations for Adolescents – Combo 2 Care Coordination $$ Plan All-Cause Readmissions 1 1 Chronic Illness Care $$ Controlling High Blood Pressure 2 2 $$ Comprehensive Diabetes Care: HbA1c Poor Control Maternity Care $$ Cesarean Rate for Nulliparous 2 2 2 Singleton Vertex Birth $$ Prenatal and Postpartum Care Medication Management $$ Asthma Medication Ratio 1 1 1 1 Patient Experience $$ CAHPS – Rating of Health Plan 1 1 Total 12 2 5 10 California Health Care Foundation 24 Endnotes 1.Making Quality Matter in Medi-Cal Managed Care: How 15. 10.09.65.03: Quality Assessment and Improvement, Maryland Other States Hold Health Plans Financially Accountable for Div. of State Documents, accessed February 13, 2019, Performance, California Health Care Foundation (CHCF), www.dsd.state.md.us; Medicaid Managed Care: 2017 Annual February 2019, www.chcf.org. Technical Report, District of Columbia Dept. of Health Care Finance, April 2018, dhcf.dc.gov; and see hhs.texas.gov (PDF). 2.Report forthcoming. 16.Some states use the term “attainment” or “excellence” 3.Intended Consequences: Modernizing Medi-Cal Rate Setting instead of “achievement.” to Improve Health and Manage Costs, CHCF, April 2018, www.chcf.org. 17.Maryland Div. of State Documents; and 2017 Quality Incentive for Medicaid Managed Care Plans, New York State Dept. of 4.The Advisory Group conducted its work based on the External Health, n.d., www.health.ny.gov (PDF). Accountability Set (EAS) in use by DHCS as of February 2019. On March 7, 2019, DHCS announced to its Medi-Cal 18.Contractor Risk Agreement Between the State of Tennessee Managed Care Advisory Committee that it would significantly and Volunteer State Health Plan, Inc., Blue Cross Blue expand the EAS. Based on the manner in which DHCS has Shield of Tennessee, accessed February 13, 2019, presented the current EAS and the new EAS measures, the www.bcbst.com (PDF). measure set is expanding from 17 measures to potentially as 19.“CCO Incentive Metrics,” Oregon Health Authority many as the 59 unique measures comprising the 2019 CMS Office of Health Analytics, accessed February 13, 2019, Adult and Child Core Sets (with seven measures found in both www.oregon.gov; HHSC Uniform Managed Care Manual: sets) for care delivered during Measurement Year 2019. Medical Pay-for-Quality (P4Q) Program, Texas Health and 5.Making Quality Matter, CHCF. Human Services, n.d., hhs.texas.gov (PDF); and Moving Apple Health to Value: Changes to Contracts for 2018, 6.Intended Consequences, CHCF. Washington State Health Care Authority, August 2017, 7.“Medi-Cal Managed Care Quality Improvement Reports,” www.hca.wa.gov (PDF). California Dept. of Health Care Services (DHCS), accessed 20.“ACOM 306, Alternative Payment Model Initiative – Withhold February 20, 2019, www.dhcs.ca.gov. and Quality Measure Performance Incentive,” Arizona Health 8.Medi-Cal Managed Care Quality Strategy Report, DHCS, Care Cost Containment System, comments.azahcccs.gov. March 28, 2018, www.dhcs.ca.gov (PDF). 21.“Auto Assignment Incentive Program,” DHCS, last modified 9.See note 4. December 14, 2018, www.dhcs.ca.gov. While DHCS does not currently provide its MCPs with a financial incentive linked 10.“Continuum of Care Reform (CCR) Data Dashboard,” to quality performance, since 2005 it has rewarded MCPs California Dept. of Social Services, accessed April 6, 2019, with preferential auto-assignment enrollment volume for www.cdss.ca.gov. (1) superior performance relative to a competing plan(s) within 1 1.“AMP Medi-Cal Managed Care,” Integrated Healthcare a county, (2) improvement, and (3) excellence. Association, accessed February 21, 2019, www.iha.org. 22.Mary Beth Dyer and Beth Waldman, Value-Based Purchasing 1 2.“Incentives,” Integrated Healthcare Association, accessed for Managed Care Procurements: A Toolkit for State Medicaid February 20, 2019, www.iha.org. Agencies, Robert Wood Johnson Foundation’s State Health and Value Strategies Program, January 2018, www.shvs.org. 1 3.The only EAS MCP measure excluded was NCQA’s “Depression Screening and Follow-Up for Adolescents and Adults.” This measure did not appear in any of the other measure sets of interest to the Advisory Group. 14.The Healthcare Effectiveness Data and Information Set (HEDIS) is the standard set of health plan performance measures in use in the US for Medicaid, Medicare Advantage, and commercially insured populations. It is maintained by the National Committee for Quality Assurance, which collects and reports performance by health plans, and which also uses performance data for health plan accreditation. For more information see www.ncqa.org. Paying Medi-Cal Managed Care Plans for Value: Quality Goals for a Financial Incentive Program 25