Currently Viewing:
The American Journal of Managed Care July 2014
Managed Care Patients' Preferences, Physician Recommendations, and Colon Cancer Screening
Sarah Hawley, PhD, MPH; Sarah Lillie, PhD; Greg Cooper, MD; and Jennifer Elston Lafata, PhD
Individual Treatment Effects: Implications for Research, Clinical Practice, and Policy
Jennifer S. Graff, PharmD; Thaddeus Grasela, PharmD, PhD; David O. Meltzer, MD, PhD; and Robert W. Dubois, MD, PhD
Simple Errors in Interpretation and Publication Can Be Costly
Dwight Barry, PhD; Lindsey R. Hass; MPH, Paul Y. Takahashi, MD; Nilay D. Shah, PhD; Robert J. Stroebel, MD; Matthew E. Bernard, MD; Dawn M. Finnie, MPA; and James M. Naessens, ScD
Success of Automated Algorithmic Scheduling in an Outpatient Setting
Patrick R. Cronin, MA; and Alexa Boer Kimball, MD, MPH
Currently Reading
Drug Plan Design Incentives Among Medicare Prescription Drug Plans
Haiden A. Huskamp, PhD; Nancy L. Keating, MD, MPH; Jesse B. Dalton, MA; Michael E. Chernew, PhD; and Joseph P. Newhouse, PhD
Structural Capabilities in Small and Medium-Sized Patient-Centered Medical Homes
Shehnaz Alidina, MPH; Eric C. Schneider, MD, MSc; Sara J. Singer, MBA, PhD; and Meredith B. Rosenthal, PhD
The Underuse of Carotid Interventions in Veterans With Symptomatic Carotid Stenosis
Salomeh Keyhani; Eric Cheng; Susan Ofner; Linda Williams, and Dawn Bravata
Characteristics of Older Adult Physical Activity Program Users
Dori E. Rosenberg, PhD, MPH; Lou Grothaus, MA; and David Arterburn, MD, MPH
Evaluating a Hepatitis C Quality Gap: Missed Opportunities for HCV-Related Care
Yang Liu, MD; Renee H. Lawrence, PhD; Yngve Falck-Ytter, MD; Brook Watts, MD; and Amy A. Hirsch, PharmD
Shifting Cardiovascular Care to Nurses Results in Structured Chronic Care
Elvira Nouwens, MSc; Jan van Lieshout, MD, PhD; Pieter van den Hombergh, MD, PhD; Miranda Laurant, PhD; and Michel Wensing, PhD

Drug Plan Design Incentives Among Medicare Prescription Drug Plans

Haiden A. Huskamp, PhD; Nancy L. Keating, MD, MPH; Jesse B. Dalton, MA; Michael E. Chernew, PhD; and Joseph P. Newhouse, PhD
Medicare-Advantage Prescription drug plans (MA-PDs) and standalone PDPs appear to respond to different incentives for plan design.
Brand name drugs without generic equivalents. For brand drugs without generic equivalents, there were no statistically significant differences between MA-PDs and PDPs in either formulary coverage or PA/step therapy (weighted by drug market share, to down-weight drugs that are used rarely), with the exception of statins; MAPDs were significantly more likely than PDPs to require PA/step therapy for statins without generic equivalents (Table 1). However, compared with PDPs, a much larger proportion of MA-PDs covered all brand name medications without generic equivalents for 4 of 6 classes (ARBs, bisphosphonates, statins and antidepressants) (Figure.) Note that the denominator includes all drugs regardless of market share, unlike Table 1, in which the calculations are weighted by market share). For example, in the Figure, a third (33.5%) of MA-PDs covered all brand ARBs with no generic equivalents while only 5.0% of PDPs did. The Figure does not include antipsychotics (a protected class) or Alzheimer’s medications (not protected) because the proportion of both MA-PDs and PDPs covering all brands with no generics for each class was greater than 90%, allowing limited opportunity to detect coverage differences. The 3 Alzheimer’s medications in this category included the Exelon™ (rivastigmine) patch, which may provide a better delivery route than oral medications for some Alzheimer’s patients, as well as Namenda (memantine) tablets and solution, which have a specific indication for moderate to severe Alzheimer’s, have fewer side effects than the cholinesterase inhibitors, and may be disease-modifying.5,6

Average copayment requirements per 30-day prescription were higher for MA-PDs relative to PDPs for 3 classes (ARBs, bisphosphonates, and statins) and substantially lower for MA-PDs vs PDPs for antipsychotics (Table 2). There was no statistically significant difference in average copayment requirements between MA-PDs and PDPs for Alzheimer’s medications or antidepressants.

Combination drugs. Results were mixed for the brand combination drugs (Table 1). MA-PDs were more likely to cover the bisphosphonate combination (alendronate plus vitamin D) and the antidepressant/antipsychotic combination (olanzapine plus fluoxetine) relative to PDPs. MAPDs were more likely to require PA/step therapy for the statin and bisphosphonate combinations and less likely to require PA/step therapy for the antidepressant/antipsychotic combination than PDPs. Average copayment requirements per 30-day prescription were significantly higher for MA-PDs than PDPs for ARB combination medications and significantly lower for MA-PDs than PDPs for the antidepressant/antipsychotic combination; there were no statistically significant differences in average co-payment requirements for the other 4 classes (Table 2).

F-tests for each of the 3 outcomes (formulary coverage, PA/step therapy, and cost sharing) rejected the null hypothesis that all effects were zero.


Across 6 medication classes used commonly by Medicare beneficiaries, we documented differences in average plan design for MA-PDs and PDPs, with MA-PDs generally offering more generous coverage, and in some cases, substantially more generous coverage. MA-PDs are more likely to cover and less likely to require PA or step therapy for brand name drugs with generic alternatives than PDPs. MA-PDs also often require a lower copayment for these drugs on average relative to PDPs. For brands without generics, we generally found no statistically significant differences in average rates of coverage and PA/ step therapy, but MA-PDs were more likely to cover all brands without generics in a given therapeutic class. Interestingly, average copayment requirements per 30-day prescription for brands without generics were sometimes higher for MA-PDs than PDPs, suggesting that PDPs may be more likely to cover only a subset of brands without generic equivalents but require lower cost sharing for those they do cover relative to MA-PDs. The fact that fewer statistically significant differences between MA-PDs and PDPs were found for the 2 protected classes than for the 4 non-protected classes we studied suggests that the 2 types of plans may be acting in a similar way with respect to the formulary protections afforded to these classes.

There are several possible explanations for why MA-PDs might cover brands with generic equivalents more generously than PDPs. One is that MA-PDs may be less concerned than PDPs about generous coverage of brand name drugs with generic equivalents due to the greater influence over clinical decision making that may result from care management processes and provider network arrangements used by MA plans/ MA-PDs relative to PDPs, which have no direct relationship with clinicians that provide services to individuals covered by traditional Medicare. For example, if a MA plan provides financial incentives or decision support that encourages generic prescribing, or the plan contracts only with clinicians with high generic prescribing rates, network clinicians may rarely prescribe brand drugs even if the drugs are covered.

A second possible explanation for the more generous coverage of brands with generic equivalents by MA-PDs could be that demand for a PDP is more sensitive to a coverage restriction for a particular drug than demand for a given MA-PD. A Medicare beneficiary who decides to join an MA plan with an MA-PD presumably selects a plan by evaluating features of both the MA plan’s nondrug coverage and its MA-PD drug coverage. As a result, if a PDP concerned about adverse selection wishes to discourage enrollment by individuals with a chronic condition associated with high medication costs, imposing coverage restrictions for medications in that class may be more likely to affect PDP enrollment of beneficiaries who use those medications than a similar restriction would affect enrollment for an MA plan and its MA-PD. Hsu and colleagues found that the risk adjustment method used for the Medicare Part D benefit, prescription drug hierarchical condition categories (RxHCCs), tended to overpredict 2006 costs for beneficiaries with low actual costs and to underpredict 2006 costs for those with high actual costs.9 If this pattern had continued since this early research was conducted, MA-PDs might have been more likely to be “overpaid” for relatively healthier enrollees, and thus perhaps better positioned to offer more generous coverage. However, CMS attempted to correct this issue in 2011 by increasing the low income subsidy (LIS) rates and decreasing the non-LIS rates. Based on the findings of Hsu and colleagues, this change may have resulted in an “overcorrection.”9

A third possible explanation, although perhaps a less likely one, is that MA-PDs could have a greater incentive than PDPs to offer generous coverage of brand name drugs without generic equivalents if the plan believes that a subset of patients will only adhere to a medication regimen by taking brand name medications and that greater medication adherence in the class can lead to lower total healthcare spending (all else being equal). Enrollees who prefer a brand formulation can apply for a formulary exception allowing them to obtain the brand version at the generic price; no data are available on the extent to which formulary exceptions are granted in this situation, however. Jung, McBean, and Kim found some support for this hypothesis for statin medications.7 They documented statistically higher rates of adherence to statin therapy for MA-PD enrollees relative to PDP enrollees, although they noted that the magnitude of the difference was very small and unlikely to be clinically meaningful. However, Erten and colleagues found similar overall use of guideline- recommended diabetes medications in MA-PDs and PDPs.8 We observed more generous coverage and less cost sharing for brands with generic alternatives in MA-PDs versus PDPs for ARBs, bisphosphonates and statins, as well as lower cost sharing for antipsychotics and antidepressant/ antipsychotic combination brand drugs, all examples where one might expect a greater likelihood of offset onto nondrug expenditures resulting from medication adherence. However, we also observed better coverage for Alzheimer’s medications, a class where an offset seems less likely.

Some of the observed differences in generosity for MA-PDs and PDPs result from the fact that MA plans with premium bids below the county benchmark receive rebates from the federal government that can be used to lower their monthly Part D premium, enhance their drug benefits, or lower drug cost sharing. However, these MA plans can instead choose to use these rebates to lower their Part B premiums, enhance nondrug benefits, or lower nondrug cost sharing, so more generous drug coverage for MA-PDs is not a given.

There are several limitations to our analysis. First, we are unable to observe all factors influencing plan design decisions for PDPs and MA-PDs. We are only able to make inferences based on observed plan design differences. Second, one might expect to see different patterns of plan design for classes that have not experienced generic entry relative to classes with at least one generic option. We were unable to examine this issue because nearly all of the most common classes utilized by Medicare beneficiaries now have at least one generic alternative. Third, our analysis may underestimate differences in MA-PD and PDP coverage to the extent that a subset of MA-PDs may have the same formulary as a PDP (and vice versa) due to common ownership. Finally, drug market share may differ in privately insured populations included in the MarketScan database and Medicare Part D plans, although many of the largest PDPs are sponsored by large pharmacy benefit managers and insurers that use the same formulary and utilization management tools for both their commercial and Part D plans.


We found modest confirmatory evidence suggesting that PDPs and MA-PDs respond to the different incentives around plan design that they face as a result of the scope of their benefits: only outpatient prescription drugs in the case of PDPs and all covered drug and nondrug healthcare services in the case of MA plans offering an MA-PD. MA-PDs are generally more likely to cover and less likely to require PA or step therapy for brand name drugs with generic alternatives than PDPs, and MA-PDs often have lower copayment requirements for these drugs. For brands without generics, we generally found no differences in average rates of coverage and PA/step therapy, but MA-PDs were more likely to cover all brands without generics in a given therapeutic class. Future work is needed to understand the factors that influence Medicare drug benefit design decisions among Part D plans, the impact of those decisions on outcomes, and how Medicare beneficiaries select from among MA-PD and PDP options.

Author affiliations: From From Harvard Medical School, Department of Health Care Policy, Boston, MA, (HH, NK, JD, MC, JN); Brigham and Women’s Hospital, Department of Medicine, Boston, MA, (NK); Harvard University, John F. Kennedy School of Government, Cambridge, MA, and Harvard University, School of Public Health, Boston, MA (JN).

Funding source: This research was funded by the National Institute on Aging (#PO1 AG032952).

Author disclosures: The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship information: Concept and design (HH, NL, MC, JN); acquisition of data (HH, NL, MC, JN); analysis and interpretation of data (HH, NK, JD, MC, JN); drafting of the manuscript (HH, JD); critical revision of the manuscript for important intellectual content (HH, NK, JD, MC, JN ); statistical analysis (HH, NK, JD, MC, JN); provision of study materials or patients (HH, NK, JD, MC, JN); obtaining funding (HH, NK, MC, JN); administrative, technical, or logistic support (JD); supervision (HH).

Address correspondence to: Haiden A. Huskamp, Ph.D., Harvard Medical School, Department of Health Care Policy, 180 Longwood Avenue, Boston, MA 02115.
1. Medicare Payment Advisory Commission. A databook: health care spending and the Medicare program: June 2012. Washington, DC: Medicare Payment Advisory Commission; 2012:155-181.. http://www

2. Congressional Budget Office. Offsetting effects of prescription drug use on Medicare’s spending for medical services, November 2012. MedicalOffsets-11-29-12.pdf

3. Polinski JM, Mohr PE, Johnson L. Impact of Medicare Part D on access to and cost sharing for specialty biologic medications for beneficiaries with rheumatoid arthritis. Arthritis Rheum. 2009;61(6):745-54.

4. Frakt AB, Pizer SD. A first look at the new Medicare prescription drug plans. Health Aff (Millwood). 2006;25(4):W252-261.

5. Kornhuber J, Weller M, Schoppmeyer K, Riederer P. Amantadine and memantine are NMDA receptor antagonists with neuroprotective properties. J Neural Transm Suppl. 1994;43:91-104.

6. McShane R, Areosa Sastre A, Minakaran N. Memantine for dementia. Cochrane Database Syst Rev. 2006;(2):CD003154.

7. Jung K, McBean AM, Kim JA. Comparison of statin adherence among beneficiaries in MA-PD plans versus PDPs. J Manag Care Pharm. 2012; 18(2):106-115.

8. Erten MZ, Stuart B, Davidoff AJ, Shoemaker JS, Bryant-Comstock L, Shenolikar R. How does drug treatment for diabetes compare between Medicare Advantage prescription drug plans (MAPDs) and stand-alone prescription drug plans (PDPs)? Health Serv Res. 2013;48(3):1057-1075. 9. Hsu J, Huang J, Fung V, et al. Distributing $800 billion: an early assessment of Medicare Part D risk adjustment. Health Aff (Millwood). 2009; 28(1):215-225.
Copyright AJMC 2006-2019 Clinical Care Targeted Communications Group, LLC. All Rights Reserved.
Welcome the the new and improved, the premier managed market network. Tell us about yourself so that we can serve you better.
Sign Up