
Transcription
Quality Incentive Program Design Decisions:An Overview of Programs in California andSelected Other StatesFEBRUARY 2021AUTHORSJustine Zayhowski, Jennifer N. Sayles, MD, MPH,and Michael Bailit
ContentsAbout the AuthorBailit Health Purchasing, LLC (Bailit Health) isa health policy consulting firm dedicated toensuring insurer and provider performanceaccountability on behalf of public agencies. Thefirm primarily works with states to take actionsthat positively influence the performance of thehealth care system and support achievement ofmeasurable improvements in health care qualityand cost management.3 Executive SummaryIncentive StructurePerformance EvaluationPerformance MeasuresConclusion5 Introduction5 Methodology6 Key Design DecisionsFor more information, visit www.bailit-health.com.About the FoundationThe California Health Care Foundation is dedicated to advancing meaningful, measurableimprovements in the way the health care deliverysystem provides care to the people of California,particularly those with low incomes and thosewhose needs are not well served by the statusquo. We work to ensure that people have accessto the care they need, when they need it, at aprice they can afford.CHCF informs policymakers and industry leaders,invests in ideas and innovations, and connectswith changemakers to create a more responsive,patient-centered health care system.1. Incentive Structure2. Performance Evaluation3. Performance Measures22 Conclusion24 AppendicesA. Key TermsB. Program SummariesC. Quality Measures Used in Each ProgramD. Consumer Experience Survey Use46 EndnotesFor more information, visit www.chcf.org.California Health Care Foundationwww.chcf.org2
Executive SummaryThe California Department of Health CareServices (DHCS) is considering alternative quality performance incentives for use with Medi-Calmanaged care plans to improve the provision of quality care for Medi-Cal enrollees and overall Medi-Calmanaged care program performance. This whitepaper examines the approaches taken by Californiaprograms, including the new 2020 Medi-Cal managedcare plan incentive strategy, and six state Medicaidprograms across key incentive design considerations.performance. The advantage of all these approachesis that they encourage quality improvement withoutthe costs associated with financial incentives.Incentive StructureState health care agencies need to determine howplans qualify for quality incentives. Programs candecide to use a “gate,” meaning that specific performance expectations must be met to qualify forthe incentive; a “ladder,” meaning the amount anincentive increases or a disincentive decreases as performance improves; or a combination of the two. Allprograms examined used either a “gate” or “gateand-ladder” approach.In designing a quality incentive program, health careagencies must determine how the incentive will bestructured. Incentive structure considerations includethese:Gates establish minimum quality standards whereasladders incentivize achieving incrementally higher levels of quality performance. What is the form of the financial incentive? Should nonfinancial incentives be applied? How do managed care plans qualify for theincentive? Should the incentive structure allow formodifications by geography within the state? Should the incentive structure allow formodifications based on variation in clinicalrisk among plans?The most common forms of financial incentives usedby researched programs include use of a capitationwithhold or a quality bonus, with multiple programscoupling both approaches. Behavioral economicssuggests that a potential loss in income, through useof a mechanism such as a withhold, is more effective to induce behavior change than a potential gain.An advantage of using a quality bonus, however, isthat plans are not financially vulnerable if they do notachieve quality performance targets.States have coupled their financial incentives withnonfinancial ones, with many programs using autoassignment preference, requiring performanceimprovement or corrective action plans, and publicizingNone of the programs examined by the authors modified their quality incentive methodologies to accountfor geographic variation or variation in clinical risk, butprograms may elect to do so to try to avoid disadvantaging plans serving vulnerable regions or populations.Performance EvaluationA major decision in developing any performanceevaluation framework is to determine what shouldbe rewarded — achievement, improvement, or both.Based on this decision, there are further considerations for each of these approaches. Achievementalone rewards high-performing plans, whereas a combination approach incentivizes both low-performingplans to improve and high-performing plans to maintain and/or improve performance. In most cases, thoseusing a combined approach weighted achievementmore than improvement to recognize high performersand to provide incentive to maintain performance.Within achievement and improvement, the nextconsideration is how to set the benchmarks. Forachievement, a program could adopt national orregional percentile benchmarks, state percentilebenchmarks, state ranking, or other non-percentileQuality Incentive Program Design Decisions: An Overview of Programs in California and Selected Other Stateswww.chcf.org3
benchmarks. Programs have tended to use nationalbenchmarks for Healthcare Effectiveness Data andInformation Set (HEDIS) measures since they providea standardized comparison of performance acrossplans, and then employ a combination of state percentile benchmarks or non-percentile benchmarkswhere national benchmarks are not available. Statebenchmarks and non-percentile values can be usedwhere no other benchmarks are available, but particularly in the case of non-percentile values, it isimportant to clearly communicate the rationale forbenchmark selection. Programs may elect to varythe achievement benchmarks by measure or to usethe same benchmark across all measures. Althoughmaintaining the same benchmark sends a consistentmessage on performance expectations, allowing variable benchmarks accounts for performance variationby measure and allows for inclusion of measures withno national benchmarks. Examined programs weresplit on whether to use one benchmark (e.g., the samenational percentile) or to vary benchmarks by measure.For improvement, programs can use absolute percentage point improvement, gap reduction, or statisticalsignificance to define improvement benchmarks. Eachof these strategies were employed by the examinedprograms. Use of an absolute term for improvementis easy to explain and operationalize but may rewardplans for improvement that results from chance anddoes not represent true improvement. When definingimprovement either through a gap-reduction strategyor a test of statistical significance, states sometimesalso apply a minimum improvement floor to ensurethat these methods do not result in rewards for tiny,meaningless improvement.A final consideration within performance evaluation is whether performance deterioration should beaccounted for. The idea behind this design elementis to ensure plans are not rewarded when quality performance is deteriorating, as this runs contrary to thepurpose of implementing a quality incentive program.Only one state Medicaid program examined accountsfor performance deterioration by financially penalizingplans.California Health Care FoundationPerformance MeasuresThe key decision regarding performance measures isthe selection of which measures should be incentivized in the program. While this consideration is outsideof this report’s scope,1 there are three supplementaryincentive design decisions considered: (1) How manymeasures should be used in the incentive program, (2)should measures be weighted equally, and (3) shouldincentives be tied to plan activities or investments indelivery system reform?Size of the measure set is a key consideration of anyquality incentive program. Smaller measure sets allowplans and their networks to focus improvement effortson a set of high-priority areas for the state but may notbe able to address every area the agency may wantto improve. As the size of the measure set increases,it may signal the importance of broader improvementat the cost of jeopardizing improvement on any givenmeasure. An important consideration, however, is thenumber of different types of providers impacted bythe measures, since plan incentives will often flowdown to providers. Programs ranged widely in the sizeof their measure sets.Another way programs can signal priorities is throughthe weighting of individual measures within the incentive set. It can be difficult to reach a decision on whichmeasures should be weighed higher relative to othermeasures, but programs doing so tend to considerfactors including agency priority focus areas, greatest opportunity for population health impact, greatestvariation between current and target performance,and differential effort or costs required to improvemeasure performance. Examined programs were splitbetween using equal and unequal weights.Finally, researchers found multiple states with incentive programs that reward delivery system reformactivities. The rationale for this approach is that targeted activities and investments to enhance thedelivery system or adopt value-based payment mayhave a longer-term and sustained impact on qualitythat reaches above and beyond a specific set of clinical quality measures. An advantage of this approachwww.chcf.org4
is that tying certain activities to incentive fundingrequires plans to develop infrastructure in areas of priority for the program and its enrollees. This strategycan be used to incentivize activities for which thereare no established quality measures or as a bridge toready plans to adopt existing quality measures onceinfrastructure and performance reporting are in place.Five states and two California programs used thisapproach with incentives for the following activities:advanced primary care model adoption, behavioralhealth integration, promoting integrated health caremodels, including essential community providers inthe network, reducing disparities, value-based payment adoption, access to services for particularpopulations, emergency department utilization, population health management, telehealth innovation,and quality incentives paid to providers.ConclusionMany design decisions are required in the development of a quality incentive methodology. Theprograms examined revealed several possible paths totake in determining the incentive structure, evaluatingperformance, or considering performance measures.California has an opportunity to improve the quality care provided to Medi-Cal enrollees, as well asthe overall Medi-Cal managed care program performance, through the adoption of practices employedin other states.This white paper is a companion piece to PayingMedi-Cal Managed Care Plans for Value: DesignRecommendations for a Quality Incentive Program.IntroductionIn 2020, the California Department of Health CareServices (DHCS) implemented a new set of qualityperformance incentives for Medi-Cal managed careplans (MCPs). Its approach is to implement a ManagedCare Accountability Set (formerly the “ExternalAccountability Set”) and require MCPs to perform atleast as well as 50% of Medicaid plans nationally (upfrom 25%). MCPs that do not meet the benchmark willbe subject to a financial penalty and will be requiredto complete a corrective action plan and qualityimprovement work.DHCS and the California Health Care Foundation(CHCF) jointly sponsored Bailit Health to examinealternative quality incentive methodologies thatDHCS could consider for its use, as external evaluations suggest that DHCS has not been generatingquality improvement through its purchasing activities.2Many state Medicaid programs operate quality incentive programs for contracted managed care plans.These programs link some portion of plan revenueand/or nonrevenue consequences to quality performance. States have pursued this strategy based on acommon belief that explicit incentives linked to performance will motivate plan behavior that will improvevalue for states and the beneficiaries they serve. Thisapproach is sometimes referred to as “value-basedpurchasing.” The application of financial incentivesalone is referred to as “value-based payment.”The purpose of this white paper is to provide information on key quality incentive program design decisionsand choices made by California purchasers and otherstates’ Medicaid programs.MethodologyThe authors examined the approaches taken byCalifornia purchasers and by state Medicaid programsin six other states to incentivize health plan quality performance in order to help inform Medi-Cal’sconsideration of an alternative quality incentive methodology. Specifically, the following programs wereconsidered: California programs California Public Employees’ Retirement System(CalPERS) Covered California (Covered CA)Quality Incentive Program Design Decisions: An Overview of Programs in California and Selected Other Stateswww.chcf.org5
DHCS Medi-Cal Managed Care PlanAccountability Set Sanctions (Medi-Cal MCP) How do managed care plans qualify for theincentive? DHCS Public Hospital Quality ImprovementProgram (DHCS QIP) Should the incentive structure allow for modifications by geography within the state? DHCS Value-Based Payment Incentive(VBP Incentive) Should the incentive structure allow for modifications based on variation in clinical risk amongplans?Other state Medicaid programs Arizona (AZ) Oregon (OR) Michigan (MI) Texas (TX) New York (NY) Washington (WA)What Is the Form of the Financial Incentive?To determine how the quality incentive will be funded,health care agencies must first determine whetherthey will use existing resources, or whether new dollars can be added to fund the incentive.The body of this white paper contains an overview ofkey design decisions and choices made by these programs. Appendix A contains definitions of key terms.Appendix B contains high-level summaries of eachprogram. Appendix C contains a crosswalk of measures used in each program.Key Design DecisionsThere are several key design decisions required todevelop a structured quality incentive program. Thebroad categories of these decisions include:1. Incentive structure2. Performance evaluation3. Performance measures3The white paper reviews key questions within each ofthese categories and discusses advantages and disadvantages of each option.41. Incentive StructureIncentive structurefollowing:considerationsinclude What is the form of the financial incentive? Should nonfinancial incentives be applied?California Health Care FoundationtheBonus. New York structures its incentive program asa bonus program, with funding allocated by the NYDepartment of Finance each year. The bonus payments are structured as performance-based capitationpayments.An advantage of this strategy is that base rates arenot at risk and therefore plans are less financially vulnerable if they fail to achieve the quality benchmarks.The disadvantage of this approach is that a successfulprogram requires significant ongoing financing fromthe state.Withhold. For many states, adding additional dollarsto their program may be infeasible. They thereforeutilize incentive structures that redistribute existingfunds or use a penalty. For example, Washington Statefunds its incentive program by withholding 2% ofplan premiums. Advantages of this approach are thatit is administratively straightforward, plans know inadvance how much of their finances are at stake, andbehavioral economics suggest that negative financialconsequences may result in more significant behavior change than a financial bonus.5 Disadvantages ofthis approach include that the lag time for the planto receive withhold dollars back could create financialstrain, particularly in downcycle years; at-risk base payment may result in fewer overall dollars to the networkdue to uncertainty of what will ultimately be earned; astate’s actuary needs to consider if the withhold performance targets are reasonably achievable; and thiswww.chcf.org6
approach is not viable for programs whose capitationrates are currently at the minimum level for actuarialsoundness.Penalty. Another option is to use a penalty as theMedi-Cal MCP program currently does. Advantagesof this approach are that the potential loss in incomecould induce significant behavior change by gainingplan executive-level engagement, and the penalty canbe structured to apply immediately upon evidence oflow performance. Disadvantages include that planshave uncertainty of financial impacts until penalties areapplied a year or more after the performance perioddue to measurement lag, and the structure may lenditself to plans passing on the penalties directly toprovider payments, potentially impacting providerparticipation and the integrity of the network.A full list of potential options as well as advantagesand disadvantages to each approach can be found inTable 1 (see page 8). Options are categorized as disincentives or incentives, as a financial incentive could becoupled with a disincentive, such as use of a withholdand a quality bonus.Should There Be Nonfinancial Incentives forQuality Performance?In addition to financial incentives, programs mayinclude nonfinancial incentives for quality performance to further encourage plans to improve quality.Common approaches include use of enrollment levers,intermediate sanctions, performance improvementplans and corrective action plans, best practice profiling, and publicizing performance. The advantage ofall these approaches is that they encourage qualityimprovement without the costs associated with financial incentives. All the states studied included severalforms of nonfinancial incentives for managed care planquality performance. The most commonly used nonfinancial incentives were auto-assignment preference,requiring performance improvement or correctiveaction plans, and publicizing performance. The MediCal MCP program utilizes all of the most commonlyused forms of nonfinancial incentives to some degree.Table 2 provides information on nonfinancial incentives (see page 9).Enrollment lever — auto-assignment. Medi-Cal usesthis approach. DHCS uses a subset of its ManagedCare Accountability Set measures, plus a cost indexand a measure of use of essential community providers, to give plans preference in its auto-assignmentalgorithm. An advantage of this strategy is that it indirectly allows high-performance plans to earn moremoney since increasing attributed members increasesthe amount of money allocated to the plan. Attributedmembers tend to be low utilizers of services, so inaddition to increasing allocated members, this strategy directs profitable members to high-performingplans. It also helps ensure that members are receiving care from high-quality providers. While DHCS staffand stakeholder interviews found the auto-assignmentto be a useful incentive to focus plan attention, especially in times when Medi-Cal enrollment is increasing,one study examining California’s auto-assignmentalgorithm found it did not significantly improve quality when compared to county-organized health system(COHS) performance, and it negatively impacted ratesof improvement in other areas of care compared toCOHS.6New York also uses a quality-informed auto-assignment algorithm for new enrollees that do not have aprior link to a provider or health plan. Plans are dividedinto five tiers based on the percentage of pointsearned in the Quality Incentive Program, and onlyplans in tiers 1 to 4 are eligible for auto-assignment.There is no differentiation between tiers 1 to 4 in theauto-assignment algorithm, and tier 5 was defined asa score of less than 36.07% of possible points. As aframe of reference, in many years no plans were intier 5, and in a few previous years only one plan hasbeen in tier 5.Publicizing performance. Washington publisheshealth plan aggregate quality scores by domain in theApple Health (Washington Medicaid) new memberenrollment packet.7 Scores are displayed in a threestar rating format across domains of getting care,keeping kids healthy, keeping women and mothershealthy, preventing and managing illness, ensuringappropriate care, satisfaction of care provided to children, and satisfaction with plan for children.Quality Incentive Program Design Decisions: An Overview of Programs in California and Selected Other Stateswww.chcf.org7
Table 1. What Is the Form of the Financial Incentive?ADVANTAGESDISADVANTAGESPROGRAM USEIncentive or DisincentivePerformancebased capitationrate adjustment.Adjustments to thebase rate receivedby plans based onperformance May be easier to insulate from budget cuts,as any potential gains are built into rates(as opposed to a separate line item in thebudget, for example) Has potential to offset “premium slide” andmay accelerate cost savings and community / social determinants investmentsif plans are not financially penalized byfuture rate declines, particularly if the staterequires reinvestment in those initiatives Relatively easy to administer Plans know in advance the amount ofpotential loss Behavioral economics suggests thata potential loss in income is moreeffective to induce behavior changethan a potential gain In the setting of an incentive, premiums willbe higher than they otherwise might havebeen if performance results in increasingcapitation rates Plan (and possibly provider) opposition toa potential reduction, and minimally adisruption, in revenue given already lowMedi-Cal paymentDisincentiveCapitationwithhold. Aportion ofbase healthplan paymentcontingent uponachievement nt inwhich poor performance results in afinancial fineCA: CalPERS,Covered CAOther states:AZ, MI, OR,TX, and WA Lag time in receiving earned money backmay create financial strains on plans, andpotentially their contracted providers The state’s actuary must consider thewithhold performance targets as“reasonably achievable” by the plansin order for the full portion of thewithhold to be considered as part ofactuarially sound plan rates This approach would not be viable if thecapitation rate including the withhold isat the minimum for actuarial soundness May require more legal counsel involvement to develop policies for assessingpenalties, which may be more complexto administer than a withholdCA:Medi-CalMCP Behavioral economics suggests thata potential loss in income is moreeffective to induce behavior changethan a potential gain Can be structured to be applied immediately upon evidence of poor performanceQuality bonus.Supplementalpayments basedon assessment ofplan performance Relatively easy to administer Plans are not financially vulnerableif they do not achieve qualityperformance targetsRequires significant and sustained statefinancing to reward for excellence andimprovementCA:DHCS VBPIncentive Funding allocated for bonus or incentivepayments may be targeted for cuts orredistribution by states during thebudget process, creating uncertaintyabout sustainabilityOther states:AZ8, OR9,and TX10 (allpartial); NYShared saving.A profit-sharingmodel in whichplan performanceinfluences thepercentage ofprofits it retains Relatively easy to administer If the state requires repayment of excessplan profit at baseline, then the sharedsavings model may ultimately result in thestate collecting less than it otherwisewould if the plan performs well Uncertainty for state and plan on amountthat can be retained/lostIncentiveCalifornia Health Care Foundationwww.chcf.org8
New York annually publishes A Report on the QualityIncentive Program in New York State,11 which allowsusers to see the tiered ranking of plans based on theirquality incentive program scores. Additionally, the NYState Department of Health website has an onlinereport of NY health plans (including Medicaid, commercial HMO/PPO, and specialty plans) that includesa tool to visualize plan performance by each measureand quality domain.12 Advantages of this strategy arethat it creates positive publicity for high-performingplans, and it enables members to make plan selections informed by performance. A disadvantage ofthis strategy is that members may not be aware of oruse available data on variation in plan quality.Table 2. Should There Be Nonfinancial Incentives for Quality Performance?ADVANTAGESEnrollment levers. Quality performanceinfluences enrollment, including newenrollment based on performance, autoassignment preference, extending openenrollment based on performance, excludingnew enrollment based on performance, orfreezing enrollment based on performanceIntermediate sanctions. Civil money penalties, appointment of temporary managementfor a managed care organization (MCO), and/or suspension of payment for beneficiariesenrolled after the effective date of the sanctionPerformance improvement / correctiveaction plan. Requirement to submit a plan toaddress underperformanceBest practice profiling. The provision ofin-depth descriptions of the best practicesused by plans to achieve high-performancerates, possibly taking the form of descriptivetext within a report or on a websitePublicizing performance. The disclosure ofthe performance rates for all plans to interested parties or the public, possibly takingthe form of annual report cards on qualityor performance dashboardsDISADVANTAGESPROGRAM USECA: Medi-CalMCP Allows high-performing plansto obtain more members thanlow-performing plans Increases the number ofmembers who will receive carethrough high-performing plans Ensures members are receivingcare from high-quality providers Helps plans think through astrategy to improve quality,and potentially leads toquality improvement Creates positive publicityfor high-performing plans,and rewards them and theirquality improvement staff Will likely be insufficient CA: Medi-Calto motivate low-perform- MCPing plans to improveOther states:NY, TX, andWA Creates positive publicityfor high-performing plansand “public shaming” forpoor-performing plans Members may not beaware of or use thesedata Other states:MI, NY, and TXDisruptive to membersCA: CalPERSNYIncreases administrativeburden on plansCA: CalPERS,Covered CA,Medi-Cal MCPOther states:AZ, MI, NY,TX, and WAAllows members to makeinformed decisions whenselecting plansQuality Incentive Program Design Decisions: An Overview of Programs in California and Selected Other StatesCA: Medi-CalMCPOther states:AZ, MI, NY,OR, TX, andWAwww.chcf.org9
How Do Managed Care Plans Qualify forthe Incentive?State health care agencies need to determine howplans qualify for quality incentives. Programs candecide to use a “gate,” meaning that specific performance expectations must be met to qualify forthe incentive; a “ladder,” meaning the amount ofan incentive increases or a disincentive decreases asperformance improves; or a combination of the two(Table 3).Medi-Cal MCP uses a gate, imposing financial penalties for each performance measure in the Medi-CalAccountability Set when performance falls below thenational Medicaid 50th percentile. A key advantage ofthis approach is that the program holds plans accountable for minimally acceptable standards. It does not,however, push higher-performing plans to improveand may discourage low performers, who may findthe gate unattainable based on current performance.Texas uses a gate-and-ladder approach, with tieredachievement and improvement targets. It uses aperformance gate for which plans neither receivepoints nor lose points. Its ladder extends “below theground,” with poor and deteriorating performanceresulting in a penalty. For example, in 2020, planscould earn /– 0.375% of their withhold based onachievement or improvement on the quality measureWell-Child Visits in the First 15 Months of Life – Sixor More. Performance above the HEDIS 50th percentile or improvement of two or more percentage pointsresults in positive points earned, increasing as plansmeet the HEIDS 66.67th percentile or improve morethan four percentage points. Performance below thestate mean or deteriorating performance of at leasttwo percentage points result in losses, with increasing penalties for performance below the HEDIS 33rdpercentile or deteriorating performance of greaterthan four percentage points. The gate in this case isthe state mean through HEDIS 50th percentile. Thisapproach is detailed in Table 4 on page 11.An advantage of this approach is that plans are incentivized to continue improving performance, regardlessof their starting point. Texas has centered its ladder onstate mean performance, to assure that positive earnings are only given to those plans meeting minimallyacceptable performance.Oregon uses a modified gate and a ladder. First,plans must meet minimum performance standards for“must-pass” measures, which are a subset of the totalincentive pool measures. If all must-pass measures arenot achieved, the incentive funding for which a planwould be eligible is reduced. This approach differsTable 3. How Do Managed Care Plans Qualify for the Incentive?ADVANTAGESGate. Specific performance expectationsmust be met to qualify for financialincentives.Ladder. The amount of the financialincentive increases/financial disincentivedecreases as performance varies.13Gate and ladder. A combination of theaboveCalifornia Health Care Foundation Establishes minimumquality standardsDISADVANTAGES Does not encourageimprovement beyondthe gate If the minimum standardsare too high, may discourage poor performers Incentivizes continuedquality improvement May not require aminimum standard Requires minimum standards Incentivizes continuedquality improvementMore complex than use ofeither a gate or a ladderPROGRAM USECA: Medi-Cal MCPCA: CalPERS,Covered CAOther states: AZ,MI, NY, OR, TX,and WAwww.chcf.org10
from a true gate in that incentives are still available,but at reduced levels, even if the plan
Incentive Structure Performance Evaluation Performance Measures Conclusion 5 Introduction 5 Methodology 6 Key Design Decisions 1. Incentive Structure 2. Performance Evaluation 3. Performance Measures 22 Conclusion 24 Appendices A. Key Terms B. Program Summaries C. Quality Measures Used in Each Program D. Consumer Experience Survey Use 46 .