What Are the Incentives for Medicare Prescription Drug Plans to Consider Long-Term Outcomes and Cost?

BACKGROUND: Medicare Prescription Drug Plans (PDPs) do not have incentives to consider long-term outcomes and costs associated with both medical and pharmacy benefits (LTOCMP) when making formulary decisions. OBJECTIVE: To identify existing quality measures, payment models, and public reporting tools for PDPs that are related to formulary decision making and that could be used as potential incentives for PDPs to consider LTOCMP. METHODS: PubMed, Google, and quality measure databases, as well as Center for Medicare and Medicaid Innovation and Centers for Medicare & Medicaid Services (CMS) websites, were searched for appropriate quality measures, payment models, and public reporting tools. RESULTS: Few quality measures and other mechanisms exist that are related to formulary decision making and have the potential to be incentives for PDPs to consider LTOCMP. Only 3 such tools were identified: (1) Medicare Part D star ratings quality measures in use by CMS, (2) the Medicare Plan Finder website, and (3) URAC PBM Accreditation standards and measures. Furthermore, the majority of the quality initiatives identified were only indirectly related to motivating PDPs to consider LTOCMP. CONCLUSIONS: Efforts are needed to develop mechanisms to align PDP incentives with LTOCMP.

Q uality initiatives in health care have gained a great deal of attention. The Centers for Medicare & Medicaid Services (CMS) have developed value-based payment models and goals, 1,2 emphasized quality measurement, 3,4 and created new health care delivery payment models, such as accountable care organizations. 5,6 All of these efforts align with the National Quality Strategy's 3 aims: better health, better care, and lower costs. 7 However, gaps still exist, and there is a need for continuous learning and improvement. 8 One major gap is a flawed health care payment system and the need to restructure financial incentives, specifically transitioning incentives from the current system that pays for quantity to one that rewards quality. 8 This restructuring can be done by applying various alternative payment models, such as adding incentives for care coordination on top of fee-for-service payments or paying clinicians a single capitated payment for all care provided to a patient. 8 Another gap, however, is the lack of evidence to support which alternative payment strategy is optimal. 8 In Medicare, Prescription Drug Plans (PDPs) cover pharmacy benefits through Part D, while Medicare covers medical benefits through Parts A and B. 9 This is different from Medicare Advantage plans, which cover both medical and pharmacy benefits. Additionally, Medicare PDPs do not retain enrollees for long periods of time; approximately 50% of enrollees are not in their original PDPs 3 years after they enroll. 10 Enrollees switch plans in order to save money and often because there are more choices in alternative plans. 10 Although Medicare PDPs should care about the 50% of enrollees who remain after 3 years, they are not encouraged to emphasize long-term outcomes. Thus, 2 factors may influence Medicare PDPs to consider short-term costs over long-term outcomes: (1) the fragmentation of pharmacy and medical benefit management and (2) the short-time horizon considered when making drug formulary decisions.
Fragmentation of pharmacy and medical benefits and consideration of only short-time horizons can create negative or reverse incentives for patient outcomes. For example, the $1,000-per-pill Sovaldi (sofosbuvir) is highly effective in curing chronic viral hepatitis C, which prevents downstream end-stage liver disease and hepatocellular carcinoma, potentially leading to long-term savings in medical costs. [11][12][13][14][15][16] Yet, a Medicare PDP focused on short-term clinical outcomes and costs related only to the pharmacy benefit is not motivated to cover such an expensive drug. 13 • Approximately 50% of Medicare Prescription Drug Plan (PDP) enrollees are not in their original PDPs 3 years later. • A health plan that does not have many long-term enrollees is less likely to pay for expensive health care services, which may not affect enrollees until much later. • The use of quality measures, value-based payment models, and other quality tactics in health care is growing.

What is already known about this subject
• This review provides an overview of what does and does not exist as incentives for Medicare PDPs to consider long-term outcomes and costs associated with medical and pharmacy benefits when making formulary decisions. • Developers of quality measures, value-based payment models, and other quality tactics can use this information to improve upon current and future quality tools.
when there is potential for harm from the provision of a service exceeding possible benefit; and (3) underuse, or the failure to provide a service that would have produced a favorable outcome for a patient. 19 Payment models were searched for on the Center for Medicare and Medicaid Innovation (CMMI) website. Other public standards and reporting tools were also searched for on the CMS website.
One reviewer searched all of the above databases from March 2015 to April 2015. Quality measures, payment models, and public reporting tools were included using the following criteria: • Directly related to the formulary decision-making process.
Here, "formulary" was specifically mentioned as part of the quality measure, payment model, or public reporting tool.
• Indirectly related to the formulary decision-making process.
Here, "formulary" was not directly stated as part of the quality measure, payment model, or public reporting tool; however, one could reasonably assume that the quality measure could affect the formulary decision-making process.
Quality measures, payment models, and other mechanisms were excluded if not relevant to Medicare PDPs.

Quality Measures
No quality measures directly related to the formulary decisionmaking process were found that were developed specifically for Medicare PDPs. However, 7 quality measures that are indirectly related to the formulary decision-making process were identified ( Table 1). Six of these measures are Medicare Part D star ratings measures, and 1 of these measures is a URAC PBM Accreditation measure.
Of the 7 quality measures, 6 are related to underuse of medications. Of these, 4 are related to adherence to medications for diabetes, hypertension, or high cholesterol. The last of the 7 quality measures, "High-Risk Medication," is related to the misuse of medications. This measure reports the percentage of enrollees who received a medication with a high risk of serious side effects when alternatives were available.
The Medicare Part D star ratings measures were also reported with findings from 2015 (Table 1). In the first measure, enrollees reported that on average it is easy to use their drug plans to get their prescriptions 88% of the time in 2-star plans to 91% of the time in 5-star plans. These percentages are crude estimates converted from combined scores originating from multiple survey questions asked on a 1 (never) to 4 (always) Likert scale. 20,21

Payment Models and Other Public Standards and Tools
No payment models that are directly or indirectly related to the formulary decision-making process were found for Medicare PDPs.
It can be argued that consumers should be able to choose plans to meet their needs and not be required to pay additional costs for expensive drugs that they will not need. However, systemwide failure occurs if Medicare PDPs do not have sufficient incentive to consider long-term outcomes and costs related to both the medical and pharmacy benefit (LTOCMP). Currently, a 12-week regimen of sofosbuvir is rarely covered by Medicare PDPs without costing the patient at least $6,000 in out-ofpocket payments, which may deter patients from treatment. 17 If the system provided more incentives for Medicare PDPs to consider LTOCMP, enrollees would have more Medicare PDP options covering sofosbuvir treatment at a lower out-of-pocket cost, and more enrollees would be appropriately treated, which would result in positive ramifications for LTOCMP.
The objective of this review was to identify existing quality measures, payment models, and other public reporting mechanisms for Medicare PDPs that are related to formulary decision making and that could be used as potential incentives for Medicare PDPs to consider LTOCMP.

■■ Methods
PubMed and Google were used to search for quality measures, payment models, and public reporting tools by using combinations of the following search terms: "Medicare Part D" [MeSH Term], "quality improvement" [MeSH Term], "quality measures," "quality indicators," and "formulary." Using the search term "formulary," quality measures were searched for in the following measure databases and collections:  31 Getting needed prescription drugs Underuse Percentage of the best possible score the plan earned on how easy it is for members to get the prescription drugs they need using the plan.
This case-mix adjusted measure is based on 3 CAHPS survey questions: 1. In the last 6 months, how often was it easy to use your prescription drug plan to get the medicines your doctor prescribed? 2. In the last 6 months, how often was it easy to use your prescription drug plan to fill a prescription at your local pharmacy? 3. In the last 6 months, how often was it easy to use your prescription drug plan to fill a prescription by mail?
The score is calculated by converting the mean of the distribution of responses to a scale from 0 to 100. 20

Quality Measures Directly or Indirectly Related to Formulary Decision Making That Could be Incentives for Medicare PDPs to Consider LTOCMP
The URAC PBM Accreditation measure was the only mechanism found that is directly related to the formulary decision-making process ( Table 2). Two standards contained in this measure, "P&T/Formulary Development" and "Economic Formulary Considerations," are used to ensure that a PBM is making formulary decisions based on clinically appropriate, safe, and cost-effective drug therapy and that cost considerations come after efficacy and safety considerations. Two mechanisms that are indirectly related to the formulary decision-making process were also found: the Medicare Part D star ratings and Medicare Plan Finder website ( Table 2). The Medicare Plan Finder website (https://www.medicare.gov/ find-a-plan/questions/home.aspx) compares PDPs with one another and allows prospective enrollees to estimate their total drug costs through each of the plans. This website also allows prospective enrollees to see formulary tiers and restrictions for their medications.

■■ Discussion
There are few available quality measures, standards, and tools directly related to the formulary decision-making process that can be used as incentives for Medicare PDPs to consider LTOCMP. Of the few quality mechanisms, only a URAC accreditation standard specifically addressed formularies per se. Yet, not all PBMs that PDPs contract with have URAC accreditation. Moreover, PBMs often create client-specific formularies and are not accredited for these client-specific processes.
Indirectly, Medicare Part D star ratings quality measures may encourage PDPs to consider LTOCMP when making formulary decisions, but these measures are limited in scope. For instance, the adherence measures are helpful because they encourage PDPs to make greater use of chronic medications to promote better long-term health outcomes and total costs, despite potentially increasing short-term pharmacy costs. However, these measures are limited to a subset of drug classes for 3 chronic conditions: diabetes, hypertension, and high cholesterol. Furthermore, whether a patient receives access to appropriate medication when considering LTOCMP is only measured in 1 specific scenario with diabetic patients.
There are many other chronic conditions, and whether PDPs are considering LTOCMP when making formulary decisions for these conditions is not assessed with the limited existing quality measures. Moreover, what may be more meaningful are adherence and access quality measures that span multiple conditions. For example, one such measure could be a composite measure of overall adherence to medications for all chronic conditions, instead of individual adherence scores for only 3 chronic conditions. While CMS cross-cutting measures apply to multiple settings, [22][23][24] these measures would apply to multiple conditions.
Considering the Donabedian framework of structure, process, and outcome quality measures, 25 the quality measures found in this review were all process measures. Although outcome measures are often most desired, they are not feasible

Name Description
Directly or Indirectly Related to Formulary Decision Making URAC PBM Accreditation standards 33,34 1. P&T/Formulary Development: PBM has a process to promote clinically appropriate, safe, and cost-effective drug therapy.
2. Economic Formulary Considerations: formulary decisions based on cost factors only after safety, efficacy, and therapeutic need have been established.

Directly
Medicare Part D star ratings 35 Positive incentives: 1. Product expansion: low star ratings by a single PDP can inhibit the PDP sponsor from expanding into new service areas or product offerings. CMS may reject plan applications to expand service areas or product offerings based on the Past Performance Methodology, which reviews 11 performance categories, one of which is the star ratings.

2.
Marketing: a "high performing icon" is placed next to 5-star PDP contracts on the Medicare Plan Finder website.

Negative incentives:
3. Marketing: A "low performing icon" is placed next to plans that have received fewer than 3 stars during the past 3 consecutive years on the Medicare Plan Finder website.   26 This would make it possible to hold Medicare PDPs accountable for how their services affect the outcomes of their enrollees, potentially through the creation of outcome-oriented quality measures. In addition to the use of quality measures, how quality measures are displayed, such as on the Medicare Plan Finder website, can also be an incentive for Medicare PDPs to consider LTOCMP. On the that website, patients input their ZIP code, current type of Medicare plan, and medication and pharmacy information and are provided with a list of plan options. Each plan option provides monthly premiums; deductible and drug copay/coinsurance; estimated annual drug costs; health benefits (if applicable); drug coverage/drug restrictions; and other programs, such as medication therapy management programs, estimated annual health and drug costs, and the overall star rating for the plan. 17 Patients can then choose the best plan for themselves based on expected costs and benefits in the next year, as well as quality indicators such as the star rating measures from the previous year. Unfortunately, star rating measures from years prior to the previous year are not currently provided on the Medicare Plan Finder website. Therefore, one improvement would be to include star rating measures, as well as costs, benefits, and specific quality measures, over time on the Medicare Plan Finder website. By making this longer-term information readily available to the public, plans would be encouraged to maintain high quality indicators over a longer time period, which may provide additional reason for plans to consider LTOCMP for their enrollees.

Limitations
There are limitations in using quality measures as incentives for Medicare PDPs to consider LTOCMP. The unintended consequences include burden to health plans, providers, and patients, as well as inaccuracies when trying to link long-term medical outcomes to Medicare PDP management of the pharmacy benefit. Considering these limitations, another way to encourage Medicare PDPs to consider LTOCMP, although again perhaps difficult to implement, would be to create multiyear health plan contracts that tie the pharmacy benefit to the medical benefit. By directly putting the health plan at financial risk for an enrollee's pharmacy and medical use for more than 1 year, the health plan will have to consider LTOCMP.
While this review has focused on Medicare PDPs, other plans, such as Medicare Advantage Prescription Drug (MAPD) plans and many commercial prescription plans, already combine medical and prescription benefits as part of the same health plan. Thus, these plans need to consider outcomes related to both medical and pharmacy benefits, although not necessarily for the long term. Interestingly, MAPD plan enrollment has increased from 5.3 million (13% of Medicare enrollees) in 2004 to 16.8 million (31% of Medicare enrolless) in 2015, and the number of PDPs has declined, from 1,875 in 2007 to an all-time low of 1,001 in 2015. 27,28 This may be because premiums have been decreasing for MAPD plans and increasing for PDPs. 28,29 Future research should explore incentives for MAPD plans and commercial plans to consider LTOCMP, as well as the effects of formulary decision making when it only considers short-term outcomes related to pharmacy benefits.

■■ Conclusions
Few quality measures and other mechanisms exist that can act as incentives for PDPs to consider LTOCMP when making formulary decisions. The Medicare Part D star ratings quality measures and their use by CMS have potential as incentives for PDPs to consider LTOCMP, although adherence measures for more than 3 chronic conditions are needed. The Medicare Plan Finder website can also be an incentive for PDPs to consider LTOCMP; adding quality measures that matter to patients, as well as displaying longer-term information on the website, can create such incentives. Also, the URAC PBM Accreditation standards and measures provide incentives for PBMs that contract with PDPs to consider medical outcomes and costs, although not necessarily for the long term.
There is a need for the development of more incentives for PDPs to consider LTOCMP when making formulary decisions regarding coverage; tier placement and cost sharing; and use of prior authorization, quantity limits, and step therapy. Successfully creating such incentives could lead to better health, better care, and lower costs in the long term.