Improving Medi-Cal Managed Care Plan Quality

Introduction

Oversight of the quality of care provided by MMC plans is an increasingly important issue for the Legislature to examine, as MMC enrollment and the geographic reach of MMC have grown significantly over the past decade. As of 2014–15, about 70 percent of Medi–Cal beneficiaries (or over 8 million) were enrolled in MMC plans compared to 50 percent (or over 3 million) in 2003–04, and enrollment in MMC is projected to continue growing in 2015–16 to 73 percent of Medi–Cal beneficiaries. The increase in MMC enrollment is partially due to the expansion of MMC to all 58 California counties in November 2013. Prior to this time, MMC plans provided coverage in only 30 counties. (Several other factors contributed to MMC enrollment growth, such as the federal Patient Protection and Affordable Care Act and the economic downturn.)

Regularly evaluating and monitoring the quality of MMC plans is important in holding them accountable for their performance and ensuring that beneficiaries enrolled in MMC receive high quality care. Currently, DHCS has some programs in place aimed at monitoring and improving the quality of MMC plans. The department is also considering the potential to implement new programs aimed at improving the quality of MMC through the upcoming Section 1115 waiver renewal process. (A Section 1115 waiver—as discussed in detail later in this report—permits the state to waive certain federal requirements in order to further the purposes of the Medi–Cal program.) Therefore, this is an optimal time for the Legislature to consider the current quality of MMC plans and the potential for innovative quality improvement programs that may be piloted through the Section 1115 waiver.

In this report, we analyze the quality of MMC plans and current DHCS efforts to improve MMC plan quality, and consider additional steps and alternative approaches, such as P4P, the state could take to monitor plans and stimulate quality improvement in MMC.

Background

Medi–Cal Overview

Federal–State Health Services Program. In California, the federal–state Medicaid program is administered by DHCS as the California Medical Assistance program (Medi–Cal). Medi–Cal is by far the largest state–administered health services program in terms of annual caseload and expenditures. Medi–Cal provides health care services to a wide range of beneficiaries, including low–income children and families, single adults, and seniors and persons with disabilities (SPDs).

Managed Care Is One of Two Medi–Cal Delivery Systems. Medi–Cal provides services through two main delivery systems—fee–for–service (FFS) and managed care. In the FFS system, a health care provider receives an individual payment from DHCS for each medical service delivered to a beneficiary. Beneficiaries in Medi–Cal FFS generally may obtain services from any provider who has agreed to accept Medi–Cal FFS payments. In managed care, DHCS contracts with managed care organizations (MCOs) to provide health care coverage for Medi–Cal beneficiaries through MMC plans. (For example, Anthem Blue Cross is an MCO, and Anthem’s contract with the DHCS to enroll and provide care for Medi–Cal beneficiaries in a specific county is the MCO’s MMC plan.) Managed care enrollees may obtain services from providers who accept payments from the MMC plan, also known as a plan’s “provider network.” The MMC plans are reimbursed on a “capitated” basis with a predetermined amount per person, per month regardless of the number of services an individual receives. The MMC plans provide enrollees with most Medi–Cal covered health care services—including hospital, physician, and pharmacy services—and are responsible for ensuring enrollees are able to access covered health services in a timely manner. In this report, we focus on MMC and do not address the Medi–Cal FFS system.

DHCS currently contracts with 22 MCOs, with some MCOs offering plans in multiple counties. The number and type of managed care plans vary by county. Counties can generally be grouped into four main models of MMC.

• County Organized Health System (COHS). In COHS counties, there is one county–run managed care plan available to beneficiaries. There are 22 counties that operate under the COHS model, covering about 1.9 million beneficiaries as of January 2015.
• Two–Plan. In Two–Plan counties, there are two managed care plans available to beneficiaries. One plan is run by the county and the second plan is run by a commercial health plan. There are 14 Two–Plan counties, covering about 5.7 million beneficiaries as of January 2015.
• Geographic Managed Care (GMC). In GMC counties, there are several commercial health plans available to beneficiaries. There are two GMC counties (San Diego and Sacramento), covering about 960,000 beneficiaries as of January 2015.
• Regional. In the Regional model, there are two commercial health plans available to beneficiaries across 18 counties. There were about 260,000 beneficiaries served under the Regional model as of January 2015.

Imperial and San Benito Counties have managed care plans that do not fit into one of these four models. In Imperial County, there are two commercial health plans available to beneficiaries, and in San Benito County, there is one commercial health plan available to beneficiaries.

Beneficiaries enrolled in both Medicare and Medi–Cal are commonly called dual eligibles. For these beneficiaries, Medi–Cal provides coverage for long–term care and cost sharing not covered by the federal Medicare program. Under a pilot initiative known as the Coordinated Care Initiative (CCI), some dual eligibles in participating counties receive services covered under Medicare and Medi–Cal through a single coordinated managed care delivery system. The CCI includes a quality monitoring and incentive program aimed at ensuring participating dual eligible beneficiaries receive high–quality care. We do not address this quality program in this report because the data available for this analysis were collected prior to the implementation of CCI.

Health Care Quality and Managed Care

What Is Quality Health Care? In order to analyze the quality of health care being provided by MMC, it is first necessary to have a working definition of what quality means in the context of health care. In this report, we base our definition of health care quality on the definition from the Institute of Medicine (IOM)—a branch of the National Academies of Science that has been a leader on work aimed at measuring and improving the quality of health care provided in the United States (U.S.). We define quality as the extent to which health care provided to individuals enrolled in managed care plans: (1) increases the likelihood of individuals’ desired health outcomes and (2) is consistent with recommended care based on current medical knowledge. For example, an annual flu shot increases the likelihood of individuals avoiding the flu—a desired outcome—and the current recommended care is to annually provide a flu shot to individuals over the age of six months. The quality of health care is also interrelated to one’s ability to access health care services, because one cannot receive quality health care if one cannot access health care services. However, an in–depth discussion of access to health care services is beyond the scope of this report.

Quality of care is not necessarily a function of the quantity of health care services provided to patients, but rather is a function of patients’ ability to receive appropriate health care services based on recommended care and patients’ needs and preferences. In some instances, fewer or less costly services may actually be more appropriate and provide a higher level of quality. For example, diagnostic imaging studies, such as an x–ray, are often not recommended for individuals with low back pain because these studies often cannot identify the specific cause of the low back pain and therefore expose the individuals to unnecessary radiation from the x–ray. On the other hand, in some instances, services may be underprovided to patients and additional services may be appropriate. For example, some adults with diabetes do not receive all recommended care, such as an annual eye exam. Therefore, it is important to assess the quality of health care provided to patients across an array of services to get a full picture of health care quality.

Quality Health Care Sometimes Not Provided by U.S. Health Care System. Over a decade ago, IOM released a seminal report on the quality of the U.S. health care system. The report found that “health care today harms too frequently and routinely fails to deliver its potential benefits,” which has led to a large gap between the care we currently experience and the level of health care we could receive. The report pointed to gaps in quality in three areas: (1) overuse of services, such as performing major surgery that is not medically necessary; (2) underuse of services, such as low rates of vaccinations among recommended populations; and (3) misuse of services, such as preventable adverse drug events (an injury resulting from medication use). Since this time, many efforts have been undertaken to understand and improve the quality of health care provided in the U.S., but quality gaps remain. This is possibly due, in part, to the incentive structures of components of the U.S. health care system, such as managed care payment arrangements, as discussed below.

Managed Care Delivery System Has Benefits, Although Payment Structure May Incentivize Underprovision of Services. Managed care is intended to encourage efficient delivery of care leading to a reduction in the overprovision of services that can result from the FFS delivery system. The managed care delivery system can lead to both fiscal accountability on the part of managed care plans and potentially better coordinated care for patients. However, in some cases managed care potentially incentivizes providers and health plans to restrict the provision of services as a result of plans being paid a predetermined capitated rate per enrollee regardless of how many services the enrollee receives and how frequently they receive them. This may in turn jeopardize the quality of care managed care plan enrollees receive. Based on this, payers of managed care plans have found it necessary to hold plans accountable not only financially but also for the quality of care provided. In recognition of this potential issue, many payers, such as Medicare, that contract with managed care plans have implemented programs aimed at incentivizing high–quality care.

In MMC, DHCS is responsible for monitoring MMC plans to ensure plans provide high–quality care to Medi–Cal beneficiaries. Before turning to the state’s current efforts in monitoring MMC quality, we first discuss the role of performance measurement more generally in the health care realm.

The Role of Performance Measurement

Performance Measures Are Used to Assess Health Care Quality and Compare Performance Across Plans. Performance measures are used to assess health care quality. Measures have been developed to assess (1) the health care delivery process (for example, the percent of women aged 50 to 75 who received a mammogram); (2) patient outcomes (for example, disease specific mortality rates); (3) patient experience and satisfaction (for example, patients’ assessments of their ease of obtaining needed prescription drugs); (4) structure (for example, whether a physician’s office has a reminder system to alert patients when they are due for screenings or appointments ); and (5) cost (for example, cost of all care associated with a knee replacement). Performance measures can also be used to compare the health care quality provided across plans and thereby identify high–performing and underperforming plans.

Performance Measurement Faces Some Limitations. Health care performance measures are subject to limitations, and no single performance measure or group of measures is able to entirely capture the overall quality of health care. Nonetheless, performance measures can provide a useful assessment of the quality of health care services provided by a plan. Additionally, the state of performance measurement is evolving as experts develop and refine the best ways to measure quality. For example, there is movement toward understanding the best ways to measure outcomes and, as this work progresses, additional outcome measures are being developed.

Several broadly used measure sets exist for assessing health care quality that allow for comparison across managed care plans. These measure sets include the HEDIS and CAHPS survey.

• HEDIS. HEDIS is a standardized set of performance measures developed by the National Committee for Quality Assurance (NCQA)—a private, not–for–profit organization that conducts work on health care quality—to provide an apples–to–apples comparison across health plans. HEDIS measures assess plan performance on the process of providing health care and in some instances the outcomes of that care. The measures are scored on a scale of 0 percent to 100 percent and assess the percent of patients enrolled in a managed care plan that received a recommended health care service or that met specified health outcome criteria during the past year. On most HEDIS measures, higher scores represent better performance. For example, one HEDIS measure assesses the percent of diabetic patients whose cholesterol level was at or below a recommended level as measured in the prior year. In this case, the more diabetic patients who had appropriate cholesterol levels, the higher the score the plan will receive on the measure. See Figure 1 for additional examples of HEDIS measures.
• CAHPS. The CAHPS survey was developed by the U.S. Agency of Healthcare Research and Quality (AHRQ)—a federal agency within the U.S. Department of Health and Human Services whose mission is to produce evidence to make health care safer, higher quality, more accessible, equitable, and affordable. The survey asks beneficiaries to report on and evaluate their experiences with health care. The CAHPS survey assesses both patient satisfaction and perceptions of access to care. In the context of managed care, CAHPS measures assess the experiences of surveyed managed care plan members in the prior six months. The members are asked about their experiences with getting the care they felt they needed, getting care as quickly as they felt they needed to receive the care, and other perceptions about their experiences with care through their health plan. See Figure 1 for examples of CAHPS measures.

Efforts to Track and Improve MMC Plan Quality

DHCS measures and monitors MMC plan quality and has programs in place that aim to (1) ensure a minimum level of quality among MMC plans and (2) encourage quality improvement.

Measurement of MMC Plan Quality

DHCS Collects Performance Measures to Assess Quality of MMC Plans. The department monitors the quality of care provided by MMC plans through performance measures from HEDIS, CAHPS, and other performance measures developed by the department. (For example, DHCS assesses MMC plans’ utilization of safety net care providers.) DHCS collects HEDIS measures annually and CAHPS survey measures about once every three years. The department also collects the CAHPS survey measures from a representative sample of adults and children enrolled in MMC based on the surveys designed by AHRQ. However, the CAHPS sample is not necessarily sufficient to provide valid data for each individual MMC plan. As a result, this report focuses on HEDIS measures collected by DHCS.

Of the available HEDIS measures developed by NCQA, DHCS selects a subset of measures each year to collect from managed care plans. In the most recent year of data available at the time of this analysis (2012), DHCS used 32 of the 81 HEDIS measures. The department determines which HEDIS measures to use by following the National Quality Forum’s criteria for measure selection, which include factors such as (1) the extent to which the measure targets an aspect of health care where performance is suboptimal or varies across plans, (2) the feasibility of collecting the measure given current data collection capabilities, and (3) the extent to which stakeholders can use the data to improve health care quality. The department also solicits input from experts and stakeholders. The selected measures are intended to capture the diverse population served by Medi–Cal.

Current Programs Aimed at Improving MMC Plan Quality

DHCS Makes MMC Plan Performance Publicly Available. The HEDIS and CAHPS measures are presented in reports posted on DHCS’s website, including annual HEDIS performance reports, triannual CAHPS performance reports, and Medi–Cal monitoring reports (referred to as the Medi–Cal Managed Care Performance Dashboard). A high–level summary of MMC plan performance is also provided to individuals when they enroll in Medi–Cal to help them select an MMC plan. These public reporting efforts are intended to assist beneficiaries in selecting MMC plans with the highest quality and to incentivize improved performance among MMC plans.

DHCS Requires Lower–Performing MMC Plans to Submit Improvement or Corrective Action Plans. The department utilizes HEDIS measures to determine whether MMC plans meet a minimum performance level (MPL)—the prior year’s 25th percentile of Medicaid managed care plan performance nationally on each HEDIS measure. If DHCS determines that an MMC plan did not perform at or above the MPL on any HEDIS measures, the MMC plan is required to submit an improvement plan to DHCS that explains the steps the MMC plan will take to improve care. For plans that do not improve as a result of the improvement plans or have a larger number of quality concerns to address, DHCS has recently begun requiring a Quality of Care Corrective Action Plan (CAP). The MMC plans that perform below the MPL on the same three or more measures for three consecutive years and MMC plans that perform below the MPL on more than 50 percent of the measures in any given year are required to complete a CAP. The CAP requires plans to identify performance improvement milestones and submit monthly documentation to DHCS that provides evidence of meeting the milestones. At the time of this analysis, only one MMC plan was under a CAP to improve its performance on HEDIS measures.

DHCS Assigns Beneficiaries to Better Performing MMC Plans in Two–Plan and GMC Counties. In Two–Plan and GMC counties, plans are rewarded for better performance through a process by which a higher percentage of default enrollments (those beneficiaries who do not select a managed care plan at the time of enrollment) are assigned to plans with better performance (this process is hereafter referred to as the performance–based auto–assignment program). Plan performance is assessed based on the level of performance on each of seven performance measures for the given year, as well as improvement in performance compared to the prior year. The measures for 2014–15 include five HEDIS measures and two measures that assess utilization of safety net care providers.

DHCS Presents Annual Quality Awards to High–Performing Plans. In addition to reporting the performance measurement data for each plan on its website, DHCS presents annual quality awards to high–performing plans based on performance as assessed by HEDIS and CAHPS measures. Achievement awards are presented by DHCS at an annual quality conference to the plans with the highest performance on HEDIS measures, the plans with the highest performance on CAHPS measures, and the plan with the most improved performance on HEDIS measures.

MMC Plan and Provider Quality Incentive Programs May Be Included in Section 1115 Waiver Renewal. DHCS is currently developing a proposal for the upcoming Section 1115 waiver renewal (see nearby box for more information on the Section 1115 wavier). The department is considering including quality incentive programs, such as P4P, targeted at MMC plans or providers in the Section 1115 waiver renewal. Briefly, P4P programs provide financial incentives for meeting specified quality performance standards. (We discuss P4P programs in more detail later in this report.) The department is holding stakeholder meetings to discuss the potential for implementing plan or provider incentive programs within managed care. These discussions are ongoing and DHCS has not yet announced which, if any, quality incentive programs will be included in the Section 1115 waiver renewal.

MMC Plan Quality Measure Performance From 2008 to 2012

Summary of Findings. In this section, we evaluate the quality of MMC plans based on an analysis of 21 HEDIS measures collected from 2008 to 2012. Our main findings include:

• The average performance of MMC plans on HEDIS measures is generally consistent with or better than the national average for Medicaid managed care plans. The average performance of MMC plans, however, is below the national average for commercial managed care plans on many HEDIS measures included in this analysis.
• The average performance among MMC plans on HEDIS measures has generally been static over the past five years despite room for improvement.
• Performance on HEDIS measures varies across MMC plans, indicating that the care experienced by Medi–Cal beneficiaries varies from plan to plan.

Each of these findings is discussed in more detail below.

Our findings are based on analyses of 21 HEDIS performance measures collected from plans in COHS, Two–Plan, and GMC counties from 2008 through 2012 (the most recent year of data available; for more details on the data we used, see the nearby box). See Figure 2 for a list of the 21 measures included in our analysis. We note that focusing on HEDIS measures presents a partial view of MMC plan quality and does not capture other factors such as patient satisfaction and cost.

Categories Used to Compare MMC Plan Performance to National Average Performance. In this analysis, we compared the average performance of MMC plans to the average performance of both Medicaid and commercial managed care plans nationwide. To make these comparisons, we grouped MMC plans into three categories: (1) consistent with national Medicaid/commercial managed care plan average, (2) better than national Medicaid/commercial managed care plan average, and (3) worse than national Medicaid/commercial managed care plan average. The three categories and the criteria used to define each category are explained in Figure 3. We selected these categories for the purposes of illustrating meaningful differences in MMC plan performance compared to Medicaid and commercial managed care plan performance nationwide. As we were not able to identify a commonly accepted approach for determining meaningful differences in performance on HEDIS measures from our review of the academic literature, we developed what we considered to be a reasonable approach.

Average Performance of MMC Plans Generally Consistent With or Better Than Average Performance of Medicaid Managed Care Plans Nationwide . . . As shown in Figure 4, the average performance of MMC plans was consistent with or better than the average performance of Medicaid managed care plans nationwide on the vast majority of HEDIS measures included in this analysis. For example, in 2010, compared to the average performance of all Medicaid managed care plans nationally, the average performance of MMC plans was (1) better than the national Medicaid average on 38 percent of HEDIS measures, (2) consistent with the national Medicaid average on 52 percent of HEDIS measures, and (3) worse than the national Medicaid average on 10 percent of HEDIS measures. The data demonstrate that, for the set of HEDIS measures considered in this analysis, California’s MMC plans typically performed at least as well, on average, as Medicaid plans nationally, with only a few measures that are below the national average.

. . . But MMC Plans Often Perform Below Commercial Managed Care Plans. As shown in Figure 5, the average performance of MMC plans is below the average performance of commercial managed care plans nationally on many measures. (See the box below for further discussion of theories explaining this difference.) For example, in 2010, the average performance of MMC plans was (1) better than the national commercial average on 29 percent of HEDIS measures, (2) consistent with the national commercial average on 14 percent of HEDIS measures, and (3) worse than the national commercial average on 57 percent of HEDIS measures. This suggests that there is room for performance improvement on average among MMC plans. Further, as the best performing MMC plans on each measure performed at or above the commercial managed care plan average on all but one HEDIS measure, it appears that the higher level of performance attained by commercial managed care plans is attainable by at least some MMC plans.

Average Performance of MMC Plans Largely Static From 2008 to 2012. The average performance of MMC plans from 2008 through 2012 has been largely static on the HEDIS measures included in this analysis and on most measures has not trended toward better or worse performance. From 2008 to 2012, average performance on 71 percent of HEDIS measures changed by less than 3 percentage points, average performance on 19 percent of HEDIS measures improved by more than 3 percentage points, and average performance on 10 percent of HEDIS measures declined by more than 3 percentage points. We note that our analysis looked at a limited period of time and it may take longer for trends in average performance to become evident.

Substantial Variation in Performance Across MMC Plans. Among California’s MMC plans, there is substantial variation in performance on the 21 HEDIS measures considered in this analysis, as shown in Figure 6. For example, in 2010, the average difference in performance between the best and worst performing plans was 36 percentage points on HEDIS measures included in this analysis. This variation in HEDIS performance translates into wide variation in receipt of some recommended health care services among Medi–Cal beneficiaries. For example, in 2012, 76 percent of women enrolled in the best performing MMC plan had a postpartum care visit during the recommended time frame after giving birth compared to only 29 percent of women enrolled in the worst performing MMC plan.

LAO Assessment of Current Programs Aimed at Improving MMC Plan Quality

In this section, we assess the effectiveness of MMC quality monitoring and improvement programs currently administered by DHCS.

Many MMC Plans Perform Below MPL on at Least One Measure Despite Requirements Otherwise. From 2008 through 2012, between 54 percent and 93 percent of MMC plans in a given year performed below the MPL (the 25th percentile of Medicaid plans nationally from the prior year) on at least one HEDIS measure collected by DHCS. Therefore, plans generally did not meet the requirement to perform at or above the MPL on all HEDIS measures between 2008 and 2012. Intuitively, the improvement plans DHCS requires for plans performing below the MPL could over time lead to better performance, but we are not able to evaluate this due to data constraints. (We note this analysis does not capture the impact of DHCS’s recently instituted policy of requiring CAPs for certain MMC plans that perform below the MPL.)

Assigning Enrollees to Plans Based on Performance Did Not Improve Quality of Care After Two Years. In Two–Plan and GMC counties, DHCS uses the performance–based auto–assignment program to assign a higher percentage of default enrollment to plans with better performance. Researchers have analyzed whether this performance–based auto–assignment process actually incentivized improvement. They did this by comparing performance over time in Two–Plan and GMC counties relative to performance in COHS counties, where the performance–based auto–assignment program is not implemented. The analysis indicated that performance among plans in Two–Plans and GMC counties had not improved more than performance among plans in COHS counties. This suggests that the auto–assignment incentive did not result in improved quality of care among MMC plans in Two–Plan and GMC counties.

Impact of Public Reporting and Quality Awards Is Unclear. To the best of our knowledge, no evaluations of the impact of reporting the performance of MMC plans on HEDIS and CAHPS measures or the impact of giving annual quality awards to high–performing plans have been conducted. Research on the impact of public reporting in the context of Medicaid, Medicare, and commercial health care has been mixed in terms of (1) whether providing enrollees with information on plan performance influences plan selection and (2) whether publicly reporting plan performance stimulates quality improvement among the plans. Two evaluations of the impact of providing CAHPS information to Medicaid beneficiaries in New Jersey and Iowa found that the information did not influence beneficiaries choice of managed care plan. The New Jersey study found that this was likely due to many beneficiaries not reading the information. Only those beneficiaries in the New Jersey study who read the CAHPS information were more likely to choose a higher quality plan than beneficiaries in a comparison group who did not receive the CAHPS information. Researchers theorize that the success of public reporting is largely dependent on the design and implementation of these efforts. For example, plan performance “report cards” that are difficult for beneficiaries to read and understand are not likely to impact the beneficiaries’ enrollment decisions.

Managed Care P4P Programs Have Potential to Improve Quality

Our analysis of HEDIS data indicates that the quality of MMC plans can be improved. The DHCS has implemented several strategies to monitor and incentivize improvement among MMC plans. However, these existing strategies could be modified based on experience to date, and in a later Section of this report we provide recommendations for modifying the existing MMC quality improvement programs in order to make them more effective.

In this section, we focus on an alternative quality improvement strategy—P4P. While California’s current performance–based auto–assignment program is a P4P program, this program only provides a noncash financial incentive and does not include cash–based financial incentives that have traditionally been provided in other managed care P4P programs.

What Is P4P?

P4P Programs Provide Performance Incentives Based on Quality. In 2001, the IOM concluded that there is a need for improved quality of health care based on “abundant evidence that serious and extensive quality problems exist throughout the U.S. health care system.” In response to these quality concerns, P4P programs (also known as Value–Based Purchasing programs) have been implemented. The P4P programs assess hospital, provider, or health plan performance based on a set of performance measures and traditionally have provided cash–based financial incentives for performance improvement and/or maintaining high performance. Some P4P programs also provide noncash financial incentives based on performance, such as auto–assignment of enrollees. While the academic literature shows potential for success of P4P programs in improving the quality of health care, not all P4P programs have been successful. Researchers have identified key features of P4P programs that are associated with the success of a P4P program. We discuss these best practices for P4P programs in detail later.

P4P Programs in Managed Care

Many Payers Have Implemented P4P Programs in Managed Care to Incentivize Improved Quality. P4P programs have been implemented in managed care to incentivize quality improvement and to counterbalance a potential incentive under managed care capitated payment arrangements for plans to reduce access to services so as to increase profitability. For example, Centers for Medicare and Medicaid Services (CMS) has implemented a P4P program for its Medicare managed care plans. In California, eight commercial managed care plans have joined together with the Integrated Healthcare Association (IHA)—a nonprofit association that conducts work on health care quality and affordability in California—to form a P4P program which provides financial incentives based on performance on HEDIS, CAHPS, and other program–specific performance measures.

Many States Have Implemented P4P Programs in Medicaid Managed Care. In addition to California, at least 19 other states have implemented some type of P4P program in Medicaid managed care. P4P incentives implemented in other states include financial incentives in the form of bonus payments, payment withholds, or auto–assignment of enrollees. While other states have used enrollment incentives, California is the only state that has implemented a performance–based auto–assignment program without also implementing cash–based financial incentives. (As discussed above, DHCS is considering implementing additional P4P programs for MMC plans and providers which may include cash–based financial incentives.)

At Least 16 MMC Plans Have Implemented Provider P4P Programs. A recent survey of MMC plans conducted by IHA found that 16 of 20 plans surveyed (2 plans had not yet been surveyed at the time of this report) were implementing some type of P4P program with the providers with whom they contract, such as physician groups. To our knowledge, no evaluations of these P4P programs have been published, although individual plans may have evaluated their P4P programs internally. The structure of these programs is likely to differ across plans in terms of measures selected and incentives provided. As discussed further below, this may be problematic for providers who contract with multiple MMC plans and therefore face different P4P program incentives.

P4P Program Best Practices

Based on experience with P4P programs in health care, experts on P4P have recommended a number of best practices to implement when using P4P to encourage quality improvement. We discuss these best practices in detail below. This Section is largely informed by two academic articles on best practices in P4P programs. We also conducted an additional literature review on evaluations and the theory of P4P programs to further inform this section.

Stakeholders Should Be Engaged Throughout Design and Implementation of P4P Program. When designing a P4P program, the input of stakeholders (such as plans, providers, and consumer advocates) is important in informing the design and implementation of the P4P program and in increasing the likelihood of stakeholder buy–in. If stakeholders are able to contribute, they are more likely to feel they have a stake in the program’s outcomes and this may ultimately make the program more successful. Engaging stakeholders also raises awareness of the P4P program which is critical to the P4P program’s success. Stakeholders need to be aware of what the program’s goals are in order to be able to work towards meeting those goals.

Measures Are a Critical Element of P4P Programs and Should Be Chosen Carefully. The measure set for a P4P program informs our understanding of health plan quality and provides information on areas that need improvement. There are several best practices that inform the selection of measures for a P4P program.

• P4P Programs Should Include a Broad Set of Measures. If the goal of a P4P program is to encourage broad–based quality improvement rather than to focus improvement on a narrow area, then the P4P program should be based on a set of measures that cover a wide range of health care services. Use of a limited number of measures may encourage improvement only on the health care services covered by those measures, possibly to the detriment of other areas of health care. In contrast, a broader set of measures would do a better job of capturing overall health care quality and encourage improvement across a range of health care services.
• All Measures Should Have Room for Improvement. Measures selected for a P4P program should have room for improvement among health plans. If measures are selected on which health plans are already performing well, the value of the P4P program is reduced because there is not room for the quality improvement the P4P program aims to incentivize. When measures are selected in areas where plans have room for improvement, this allows health plans to demonstrate meaningful changes in performance.
• Measure Set Should Generally Be Consistent but Reviewed and Updated Over Time. A measure set that is generally consistent over time has the benefit of allowing health plans to implement quality improvement initiatives and understand whether they are improving performance over time. Nonetheless, measure sets need to be reviewed and revised periodically to ensure that the measures continue to reflect current medical knowledge and still have room for improvement. Once the majority of plans are performing well on a given measure, the measure should continue to be tracked to ensure that performance remains high across plans, but the measure should no longer be used to determine incentives. Other measures should be added over time to continue to ensure that a wide range of performance is assessed by the P4P program. Changes should also be made to the measure set as new measures emerge and more information becomes available on the best ways to measure quality.
• Measures Should Be Aligned With Those Used by Other Payers and P4P Programs. Many health care providers are under contract with several payers. For example, an individual physician may treat MMC enrollees, Medicare managed care enrollees, and commercial managed care enrollees. To address this issue, payers should coordinate to track and incentivize performance using the same set of measures to the extent possible. This sends a clear message to health plans and providers regarding those areas that they should focus on in the quality improvement initiatives. For example, in the context of MMC, measures for the families and children population could be aligned with those measures used in P4P programs implemented in the commercial market, and measures for SPD populations could be aligned with those measures used by CMS for Medicare managed care plans. Alignment of measures also allows for comparison across managed care plans and populations.

Plan Performance Should Be Assessed and Incentivized Based on Improvement. In addition to determining the level of a managed care plan’s performance on measures in a given year, plans should also be assessed on whether their performance has improved over time. This would more likely encourage low–performing plans to work towards higher attainment scores than basing performance incentives solely on annual levels of performance. In the absence of improvement measures, low–performing plans may determine that it is not worthwhile to work toward better performance because modest increases in performance on attainment measures may not achieve the threshold for performance incentives. Measuring improvement also allows for identification of managed care plans that are improving the quality of care provided over time.

Incentives Should Be Large or Important Enough to Create a Response From Plans. The incentives used in P4P programs (whether cash–based or enrollment incentives) need to be sufficiently large or important to generate a response from health plans. If health plans only see incentives as minimally impacting their business, they are unlikely to invest resources into quality improvement, thus limiting the improvement the P4P program aims to achieve. In order to determine the appropriate level of incentives, stakeholder input should be considered to understand what level of incentive is viewed as important to the stakeholders. However, other considerations—such as the cost associated with implementing the incentives—also need to be weighed.

Performance Targets Should Be Clearly Communicated. In order to help plans improve performance and meet goals for performance incentives, the performance targets should be clearly communicated to managed care plans. Evaluations of P4P programs have found that greater performance improvements are attained when plans are given advanced notice of performance targets and receive clear explanations of what targets need to be met (in terms of levels of performance and improvement in performance) in order to receive performance incentives.

Plans Should Be Provided With Technical Assistance and Information on Successful Efforts of Other Plans. In order to increase the likelihood of success of the P4P program at improving the performance of health plans, health plans should be provided with technical assistance. The assistance should both help the health plans to understand the structure of the P4P program and help the plans as they develop and implement quality improvement initiatives. Further, when a given health plan implements a quality improvement initiative that successfully improves the health plan’s performance, that success should be shared with other health plans. The sharing of best practices will allow all plans to better understand which approaches to quality improvement work well.

P4P Programs Should Be Evaluated. P4P programs need to be evaluated over time to determine the extent to which improvements in quality have occurred. To the extent possible, P4P programs should be modified based on the successes and failures identified in the program evaluation. Further, research indicates that P4P programs may have the potential to worsen existing disparities in care (for example, racial disparities) because providers may have an easier time improving performance for certain groups. Based on this, disparities in care should be assessed prior to implementing a P4P program and should be tracked on an ongoing basis to ensure that the P4P program is not worsening disparities.

LAO Recommendations

We offer several recommendations, including modifications to existing strategies implemented by DHCS aimed at monitoring and improving the quality of care provided by MMC plans, as well as recommendations for a new P4P program for MMC plans. Our first recommendation is applicable to the extent that the administration proposes P4P programs through its Section 1115 waiver renewal proposal. If the administration does not propose a P4P program specific to MMC with cash–based incentives, our second recommendation regarding a pilot P4P program is applicable. Our remaining recommendations relate to modifications to existing strategies implemented by DHCS and are applicable regardless of whether the administration proposes a new MMC P4P program.

Consider the Extent to Which P4P Programs Proposed in Section 1115 Waiver Renewal—if Any—Align With Best Practices. If DHCS proposes P4P programs in the Section 1115 waiver renewal, such as new program approaches using cash–based incentives, it will be important for the Legislature to consider whether these programs align with the best practices discussed above. In such a situation, we recommend that the Legislature require DHCS to report at budget or policy hearings on the design of the P4P program prior to the program’s implementation. The department should explain how the P4P program aligns with the best practices outlined above and how expert and stakeholder input were incorporated into the program’s design. The more proposed P4P programs align with best practices, the more likely they are to be successful at improving the quality of care. While we focus on MMC in this report, the best practices outlined in this report apply more broadly to P4P programs implemented in managed care or in other areas of the Medi–Cal program such as hospital care.

Direct Implementation of P4P Pilot Program for MMC Plans (if No P4P Program With Cash–Based Incentives Is Proposed in Section 1115 Waiver Renewal). If DHCS does not propose a P4P program with cash–based financial incentives for MMC plans in the Section 1115 waiver renewal, we recommend the Legislature direct the implementation of such a P4P program. We recommend beginning with a pilot P4P program that could be scaled up after evaluating its effectiveness at improving quality among MMC plans. We recommend the Legislature enact legislation directing that implementation of the pilot P4P program incorporate the following components.

• Pilot P4P Program With Cash–Based Financial Incentives and Based on Best Practices. The pilot P4P program should implement the best practices discussed above, including (1) use of a broad set of measures that can be used to compare performance of MMC plans to other managed care plans, (2) use of performance incentives based on improvement in performance over time, (3) implementation of performance targets that are easy to understand, and (4) communication of the performance targets to MMC plans in advance of performance measurement. In terms of structuring the cash–based financial incentives, P4P programs implemented in other managed care settings have taken several approaches, including: (1) annual performance–based bonus payments above the capitated rate plans received, (2) withholds of a percentage of the capitated rate which are paid out based on achievement of performance targets, or (3) withholds of annual capitated rate increases which are paid out based on achievement of performance targets. Bonus payments require additional resources, whereas withholds do not. The state would need to seek guidance from CMS regarding which financial approaches may be implemented in MMC and would need to seek the necessary approvals from CMS.
• Include MMC Plans From Diverse Set of Counties. To understand the impact of P4P across plans of various sizes and plans that serve varied populations, plans should be chosen from both rural and urban areas. Likewise, plans should be chosen that operate in different types of MMC counties.
• Consider Expert and Stakeholder Input in Developing and Implementing Pilot P4P Program. The design of a pilot P4P program is important for its ultimate effectiveness at improving performance among MMC plans. We recommend the use of an expert panel to advise on the design of the P4P program to ensure that the most up–to–date knowledge on P4P can be incorporated into the program. We also recommend the use of a stakeholder panel to solicit input from MMC plans, advocates, and Medi–Cal beneficiaries to ensure buy–in from all stakeholders prior to the implementation of a pilot P4P program.
• Require DHCS to Report on How Pilot P4P Program Aligns With Best Practices. We recommend that the Legislature require DHCS to report at budget or policy hearings on the design of the P4P program prior to the program’s implementation. The DHCS should explain how the P4P program aligns with each of the best practices outlined above and explain how expert and stakeholder input were incorporated into the program’s design.
• Evaluate the Impact of the Pilot P4P Program. We recommend the state contract with a research organization to evaluate the effectiveness of the pilot P4P program, including assessing stakeholder feedback. The research organization should be consulted in the development of the pilot to ensure that it is structured to allow for a meaningful evaluation. An evaluation of the program’s success would allow the program to be modified such that an effective P4P program model can be developed through the pilot process before consideration of scaling up the program to all MMC plans. The evaluation should address (1) impact of the program on improving MMC plan performance, (2) MMC plans’ feedback on the program, and (3) potential modifications to the program that may improve its effectiveness.

Require Changes to the Performance–Based Auto–Assignment Program to Align With P4P Best Practices. The performance–based auto–assignment program implemented by DHCS aligns with some, but not all, of the best practices for P4P discussed above. We recommend the Legislature enact legislation requiring the following changes be made to the performance–based auto–assignment program to better align the program with P4P best practices and therefore improve the potential for the program’s success.

• Expand Measure Set. We recommend the Legislature require DHCS to expand the measure set used in the performance–based auto–assignment program. While DHCS tracks many HEDIS performance measures for managed care plans, only seven performance measures are used in the performance–based auto–assignment program. This may create a distorted incentive system wherein managed care plans focus on performing well on this small number of measures while letting performance lag in other areas. The DHCS could implement a broader set of measures using the HEDIS measures that they are already tracking annually.
• Provide Advance Notice of Measures. The DHCS has indicated that plans are notified of the measures included in the performance–based auto–assignment program for the current year at the midpoint of the year. The rationale for delaying notification is to encourage plans to focus on all areas of quality rather than only the seven measures included in the performance–based auto–assignment program. The best practices described above indicate that it would be preferable to use a broader set of measures in the performance–based auto–assignment program and to give plans sufficient notice of those measures in advance. Evaluations of P4P programs have found that improvement is more likely when plans are aware of the measures and goals of the P4P program.
• Evaluate Revamped Performance Based Auto–Assignment Program. After changes are made to better align the performance–based auto–assignment program with best practices and a sufficient period of program implementation with the new design has occurred, the program should be evaluated to determine its impact on MMC plan performance. To the extent that the program is found to not have an impact on plan performance, the Legislature could consider requiring DHCS to discontinue this program and invest the resources in alternative, more effective quality improvement strategies.

Direct DHCS to Expand CAHPS Collection. The Legislature should direct DHCS to expand its use of the CAHPS survey to capture patient satisfaction data annually rather than every three years. In addition, DHCS should collect a sufficient sample of CAHPS surveys from enrollees in each MMC plan to allow for the calculation of valid CAHPS scores for each MMC plan. This would allow CAHPS survey results to be used in P4P and other quality improvement programs targeted at MMC plans. While we acknowledge there will be additional costs associated with collecting the data annually, we find this information would be valuable in understanding MMC plan quality. Accordingly, we recommend that DHCS report on the cost of collecting the CAHPS survey annually.

Require DHCS to Report on Success of New CAP Policy Targeting MMC Plans Performing Below the MPL. We find that despite DHCS’s minimum performance requirement for MMC plans, some plans continue to perform below the MPL on HEDIS measures. While DHCS has recently implemented a new CAP policy to target these low–performing plans, it will be important to see whether this has an impact on plan performance. We recommend the Legislature enact legislation requiring DHCS to report on the effectiveness of this policy at improving performance of low–performing MMC plans to levels above the prior year’s national Medicaid 25th percentile.

Conclusion

Overall, we find that MMC plans are performing at least as well as Medicaid managed care plans nationally on the HEDIS measures considered in this analysis. There is, however, wide variation in performance across MMC plans and room for improved performance. The state’s Section 1115 waiver renewal, which the state is currently negotiating with the federal government, provides a unique opportunity for the state to further its efforts to improve the quality of MMC plans by obtaining federal funds to support new performance improvement programs, including P4P. We provide a description of best practices for implementing P4P programs to guide the state’s efforts in this area.