University of Southern California Dissertations and Theses

Thesis on Medicare Part D

(USC Thesis Other)

Thesis on Medicare Part D

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Thesis on Medicare Part D
by
Yifan Xu

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the Requirements for the Degree
DOCTOR OF PHILOSOPHY
Health Economics

December 2019

ii

Table of Contents
Acknowledgements ...................................................................................................................................... iv
Chapter 1: Introduction ............................................................................................................................... 1
References ................................................................................................................................................. 4
Chapter 2: Variation in Generic Dispensing Rates in Medicare Part D ........................................................ 5
Abstract ..................................................................................................................................................... 5
Introduction ............................................................................................................................................... 6
Methods .................................................................................................................................................... 7
Data ....................................................................................................................................................... 8
Formulary Design ................................................................................................................................. 8
Multivariate Analyses ........................................................................................................................... 9
Results ..................................................................................................................................................... 10
Discussion ............................................................................................................................................... 14
Conclusions ............................................................................................................................................. 17
References ............................................................................................................................................... 19
Tables and figures ................................................................................................................................... 22
Appendix ................................................................................................................................................. 26
Chapter 2 : Long-Term Persistence of Prescription Drug Spending in Medicare Part D ........................... 33
Introduction ............................................................................................................................................. 33
Relevant work ......................................................................................................................................... 34
Methods .................................................................................................................................................. 35
Data and sample .................................................................................................................................. 35
Prescription drug spending.................................................................................................................. 36
Results ..................................................................................................................................................... 37
Distribution of prescription drug spending over time (figure 1-3)...................................................... 37
Persistence in prescription drug spending (figure 4, 5, table 1) .......................................................... 37
Cumulative mortality by spending groups (figure 6) .......................................................................... 38
Discussion ............................................................................................................................................... 39
Conclusions ............................................................................................................................................. 41
References ............................................................................................................................................... 41
Tables and figures ................................................................................................................................... 42
Chapter 4: Enrollment Growth in the Part D Enhanced Plans .................................................................... 48
iii

Background ............................................................................................................................................. 48
Methods .................................................................................................................................................. 50
Data and sample .................................................................................................................................. 50
Enrollment ........................................................................................................................................... 51
Premium .............................................................................................................................................. 51
Beneficiary characteristics and switch behavior ................................................................................. 52
Results ..................................................................................................................................................... 52
Overall trend of enrollment and premium in the enhanced Part D plans ............................................ 52
United Health Group ........................................................................................................................... 53
Humana ............................................................................................................................................... 54
Discussion ............................................................................................................................................... 55
Future work ............................................................................................................................................. 57
References ............................................................................................................................................... 58
Tables and figures ................................................................................................................................... 59
Chapter 5: Summary and Future Research Directions ................................................................................ 69

iv

Acknowledgements
This dissertation would not have been possible without the tremendous support and intellectual
guidance from my advisor, Geoffrey Joyce. I am deeply grateful for his dedication and inspiration to his
students. His comments and insightful ideas never failed to lead me to promising directions and helped
me reflect on my research. Our conversations have always motivated me deeply to pursue excellence in
my research. He is also a great mentor and friend, to whom I always feel comfortable about expressing
my thoughts and feelings.
I would like to extend my sincerest gratitude to my thesis committee members for their interest in
my research and insightful comments along the way, which also challenged and enriched my ideas and
understanding. I would like to thank Professor John Romley, for his excellent teaching for econometrics;
Professor Erin Trish, who inspired me for new ideas and provided thoughtful guidance; Professor Dana
Goldman, who encouraged me to always believe in myself and strive for the highest standards and
perfection.
Furthermore, I would like to pay special thanks to Professor Jeffrey McCombs, who offered me
the admission to this wonderful program in the first place. During the years at USC, Dr McCombs has
helped me and all his students without any reservation, be it getting funding and fellowships, internships
and job opportunities, networking, dinner events and tickets to football games! I would also like to thank
Patricia St Clair, our best resource for all data and programing related issues. She has worked diligently to
support our work.
This long journey would not have been possible without the support and friendship of my peers at
USC, and the memories we had together will be cherished forever. I would like to thank Wendy Cheng,
Ning Ning, Jianhui Xu and Laura Henkhaus, for all the insightful conversations, genuine support and the
good times.
v

Last but not least, I owe my deepest gratitude to my parents, Jialian and Deming, and my
husband, Libin, who have supported me unconditionally and always believed in me throughout the
journey. They are my daily inspiration, and without their unwavering love, understanding, and sacrifices,
I would not be where I am today. Finally, I want to thank my dearest son, Justin, who has brought me
endless joy and laughter, and it is because of him, I become a stronger mother, thinker, and individual.

1

Chapter 1: Introduction
Spending for prescription drug is high and continues to increase in the United States, with a total
expenditure of $344 billion in 2018.(IQVIA, 2019) Medicare, the largest single purchaser for all health
care service, has spent nearly $100 billion on its outpatient prescription drug program, Part D, in
2018.(Office of Budget (OB), 2017) The growing expenditure is attributable to a combination of factors
including the introduction of new drugs, increased volume of drug use, and rising prices.(Gellad et al.,
2012) In spite of the concerns of its financial sustainability, Part D has been a success in many respects. It
has improved the access to prescription drug for millions of elderly beneficiaries, who are the biggest
users of medications. Now nearly 90 percent of the prescriptions filled are generic drugs, which have
saved the program hundred billions of dollars in the past years,(Association for Accessible Medicine,
2017) and lowered the cost-sharing for the beneficiaries. Due to the competitive nature of Part D market,
beneficiaries are entitled to choose a plan that may best meet their medical and financial needs, out from
many available plan options. In addition, beneficiaries are allowed to switch plans during the annual
open-enrollment period if they were no longer satisfied with their current plans. Overall, the competition
among the private Part D insurers keep the plan premiums low and stable, with an average monthly
premium of $30 for many years,(MedPAC) and the premiums are project to decline in the coming year.
Nevertheless, there are several underlying issues with the current prescription drug system. The
one that gets most attention is the high prices paid for the prescription drugs. While the clinical
effectiveness of some innovative medication may justify its high price, many expensive drugs may not
yield an equivalent value as their alternatives that have lower prices. Many studies have shown that other
high-income countries can achieve both lower list and net prices for almost all prescription drugs while
regulate the use of certain drugs that provide limited value compared to its lower-cost competitors. Thus,
it raises question about the value of our spending and affordability issues.(Kang et al., 2019) Another
critical issue lies in the benefit design of Part D, where there is a coverage gap for the beneficiary to pay
2

full price of their drugs after initial coverage, and a catastrophic phase where they pay 5% of the drug list
prices, but no hard cap on beneficiary’s out-of-pocket spending. The design, combined with the
increasing prescription drug spending, has distorted incentives for health plans to manage drug spending
while shifting costs to Medicare and imposing heavy financial burdens on those patients with the highest
spending on drugs.(Lieberman, 2019) The absence of limit on out-of-pocket spending in Part D
undermines a vital protection of insurance that is the norm in commercial insurance, leaving a small
percentage of beneficiaries exposed to unlimited out-of-pocket spending for prescription drugs. In
addition, beneficiaries pay for the list drug price rather than the net price after rebates and discounts.
There has been a number of policy proposals aiming to reconstruct the benefit design under Part
D. In the Trump Administration’s Fiscal Year 2020 budget, it proposes strategies targeted at increasing
competition, encouraging better negotiation, incentivizing lower list prices, and lowering out-of-pocket
costs for beneficiaries.(Executive Office of the President and Office of Management and Budget, 2019)
More specifically, it proposes to “provide beneficiaries with more predictable annual drug expenses
through the creation of a new out-of-pocket spending cap”. In July 2019, the bipartisan bill to lower
prescription drug price, the Prescription Drug Pricing Reduction Act (PDPRA), has passed by the Senate
Finance Committee. PDPRA seeks to address many issues with the Medicare Part D, focusing on
lowering drug price, bolstering competition, and improving benefit designs.(Lieberman, 2019)
The studies presented in this dissertation utilize real-world data to investigate three important
topics on Part D and discuss various policy-relevant issues that are less researched before, concerning
prescription drug spending for both plans and beneficiaries, and strategic formulary designs and premium
setting of plans. The data used for all three studies are from a 20 percent random sample of the
administrative data of Medicare Part D claims, through a data use agreement with the National Bureau of
Economic Research, and plans’ contract and enrollment information from Center of Medicare and
Medicaid Services. The data contain rich information on prescription drug claim, beneficiary
characteristics, plan characteristics and its formulary design.
3

Our first study examines the variation in medication use among Part D plans. While on average
the utilization of generic drugs has increased substantially over time and reached nearly 90 percent, it
masks considerable variation across plans, even within a therapeutic class. Such variation presents
significant potential savings for beneficiaries, plan sponsors and the federal government. We use
Medicare Part D claims from 2006-2014 to examine the variation in generic and brand use across plans
and how patient, plan and area characteristics affect generic use within a therapeutic class.
While part of the high spending can be saved by using more cost-effective drugs (i.e., generics),
beneficiaries with the highest cost face additional financial burden in the absence of a cap on the out-of-
pocket spending. Few prior work studied the long-term persistence of prescription drug spending among
the elderly population and the availability of the eight-year longitudinal claims data allow us to evaluate
the degree of long-term spending persistence among Part D enrollees. Results from such study add solid
empirical support for the policy change of adding limits on out-of-pocket spending in Medicare Part D.
Lastly, we examine the recent enrollment growth in Part D plans offering enhanced benefit
package (“Enhanced alternative”). These plans, especially stand-alone prescription drug plans, have been
attracting more enrollees over time by offering enhanced benefit such as reduced cost-sharing, some
coverage in the donut hole and an expanded drug formulary. While the coverage for plans with standard
benefit has become considerably more generous over time, it is interesting to understand why the more
expensive enhanced plans are becoming popular, and what type of beneficiaries are enrolled in these
plans. Findings from these analyses would provide important implications for federal spending and the
ongoing integrity of the Part D competitive bidding structure.

4

References
Executive Office of the President, and Office of Management and Budget (2019). A budget for a better
America: fiscal year 2020 budget of the U.S. government.
Gellad, W.F., Donohue, J.M., Zhao, X., Zhang, Y., and Banthin, J.S. (2012). The Financial Burden From
Prescription Drugs Has Declined Recently For The Nonelderly, Although It’s Still High For Many.
Health Aff. Proj. Hope 31, 408–416.
Kang, S.-Y., DiStefano, M.J., Socal, M.P., and Anderson, G.F. (2019). Using External Reference Pricing
In Medicare Part D To Reduce Drug Price Differentials With Other Countries. Health Aff. (Millwood) 38,
804–811.
Lieberman, L.A., Paul Ginsburg, and Steven M. (2019). Understanding the bipartisan Senate Finance
prescription drug reform package.
MedPAC The Medicare Prescription Drug Program (Part D): Status Report.
Office of Budget (OB), A.S. for F.R. (ASFR) (2017). FY 2018 Budget in Brief - CMS - Medicare.
2017 Generic Drug Access and Savings in the U.S. Report | Association for Accessible Medicines.

5

Chapter 2: Variation in Generic Dispensing Rates in Medicare Part
D

Abstract
The use of generics in Medicare Part D drives cost savings for plan sponsors, beneficiaries and the federal
government. However, there is considerable variation in generic use across plans, even within a
therapeutic class. In this paper, we use Medicare Part D claims from 2006-2014 to examine the variation
in generic and brand use across plans and how patient, plan and area characteristics affect generic use
within a therapeutic class. We find that while generic use has increased markedly over time in Part D,
substantial variation across plans persists in several common therapeutic classes. Beneficiary
characteristics such as gender and health status are associated with generic use, as are plan characteristics
such as plan type (PDP or Medicare Advantage), premium and parent company. We find circumstantial
evidence that in certain classes rebates may play a role in influencing brand use, although the exact
relationship is unknowable given rebates are proprietary.

6

Introduction
The growing gap between a drug’s list price and what manufacturers actually receive post-
discounts or rebates (i.e. the net price) is at the at the center of an increasingly public debate over rising
drug costs. Pharmaceutical manufacturers argue that net drug prices have increased modestly over the
past 5 to 7 years, and that the trend towards higher list prices with commensurately larger rebates is
driven by PBMs and other entities in the supply chain whose revenues are often tied to the list price of a
drug.
Health insurers and pharmacy benefit managers (PBMs) negotiate rebates or discounts from drug
manufacturers in return for preferred formulary placement. They pass on some fraction of the rebates to
the plan, retaining the rest as part of the “spread” or the difference between what the PBM reimburses the
pharmacy for a drug and what it bills the health plan. The key issue is that each transaction in the supply
chain is proprietary, including the size of the rebate and the portion that is passed on to the plan. This
opaque arrangement can create an incentive for a PBM to favor high-cost, high rebate drugs over lower
cost products.
Unlike commercial markets, Part D plans are required to disclose rebates and other discounts to
CMS, which are used to calculate premiums and actuarial value. Full disclosure of rebates should
mitigate plan incentives to favor high cost, high rebate drugs over lower cost products. However, recent
work demonstrates that the incentives in the commercial market to sometimes favor high-cost, high rebate
drugs also exist in Medicare Part D, leading to some brand name drugs costing beneficiaries less than
generic versions.
1

Several other structural features of Part D may also decrease beneficiaries’ incentives to use
generics. First, about 12 million (28%) Part D beneficiaries receive some form of low-income subsidy.
2

Protecting low-income beneficiaries from some or all cost sharing makes them less price sensitive and
thus largely unaffected by a formulary that favors high cost, high rebate drugs. Secondly, high list price
7

drugs move beneficiaries into the catastrophic phase of the benefit sooner, where plan liability is low and
federal liability is high.
3
Once a beneficiary reaches the catastrophic phase, the federal government pays
80% of the drug costs, the plan pays 15% and non-subsidized beneficiaries pay 5%. However, these
percentages are based on the negotiated list price of the drug, not the net price after rebates and other
discounts. The Medicare Payment Advisory Commission (MedPAC) has previously pointed out that
beneficiaries reaching the coverage gap is the fastest growing portion of the program’s cost.
4

Finally, the ACA and the Balance Budget Amendment of 2018 restructured the Part D benefit to
reduce beneficiary cost-sharing in the coverage gap.
5,6
The primary change required manufacturers to
provide discounts on brand drugs purchased in the coverage gap, with the discount reaching 70% of the
list price in 2019. While this discount is valuable to beneficiaries, it counts towards their out-of-pocket
spending, pushing more beneficiaries into the catastrophic phase faster, where plan and patient liability is
lowest.
7

The net effect of these factors on PBM and plan incentives is uncertain. Given the unavailability
of rebate data, an indirect way to assess PBM and plan incentives is to examine the variation in
medication use across plans. For example, while generic drug use in Medicare Part D is generally high,
off-patent brands such as Lipitor (atoravastatin), Nexium (esomeprazole) and Neurontin/Lyrica
(gabapentin) remained top-selling medications despite widely available generic versions or over-the-
counter substitutes. Favoring high cost brands over lower cost generics makes little sense in a
competitive market, and thus may reflect incentives for plans and plan sponsors to steer beneficiaries to
heavily rebated drugs despite higher net costs to Medicare. We use Medicare Part D claims from 2006-
2014 to examine differences in brand and generic drug use within a therapeutic class, and the extent to
which beneficiary, plan and area characteristics explain the variation.
Methods
8

Data
We used Part D drug claims from January 2006 through December 2014 for a 20% random
sample of Medicare beneficiaries enrolled in either stand-alone prescription drug plans (PDP) or
Medicare Advantage plans offering drug coverage (MAPD). We did not analyze employer retiree plans
for this analysis given the lack of information about those plans. The pharmacy claims capture all key
elements related to prescription drug events including drug name, National Drug Code (NDC), dosage,
quantity, date of service, and payments made by the beneficiary, plan, and other third-party coverage. We
aggregated individual NDCs into therapeutic classes based on First Databank (FDB) definitions, focusing
on 10 of the most commonly used classes in Part D: antiasthmatics, antihistamines, antidiabetics,
antineoplastics, anticoagulants, antihypertensives, antihyperlipidemics, anticonvulsants, antidepressants
and antiulcerants.
We linked Part D claims to the plan characteristics file, which contains detailed information on
each plan’s formulary, benefit design, and other utilization management (UM) policies. This allows us to
control for plan-level premiums, cost-sharing requirements, and formulary restrictions such as prior
authorization that can affect whether a brand or generic drug is dispensed.
Formulary Design
Most Part D plans use tiered formularies to steer beneficiaries to preferred medications. In 2015,
the vast majority of Part D enrollees were in plans that used five cost-sharing tiers: preferred and non-
preferred tiers for generic drugs, preferred and non-preferred tiers for brand drugs, and a tier for high-cost
specialty drugs, with average copayments of $1/$4/$38/$80/29%.
8
Plans may also combine formularies
with other rules— such as step therapy and prior authorization—to manage beneficiaries’ use of
prescription drugs. Restrictive formularies that list fewer drugs in each therapeutic class will generate
higher rebates because they are better able to steer enrollees to a particular manufacturer’s drug and away
from the drug’s competitors. Plans can also encourage enrollees to switch from a brand-name drug to the
9

generic form of a different drug that is in the same therapeutic class through various methods, commonly
referred to as “therapeutic substitution”.
We relied on First Databank (FDB) definitions to classify drugs as either single-source brand
(SS), multisource brand (MS), or generic (G). Single-source brands have no generic substitutes and so
prescriptions written for SS drugs limit Part D plans’ opportunities for generic substitution. By contrast,
multi-source brands have generic equivalents with the same active ingredient(s), and given the financial
incentives to prescribe a generic over its brand equivalent, almost all prescriptions for multisource brand
drugs are dispensed as generics.
9
As a result, any observed variation in generic drug utilization rates
across plans is essentially variation in single-source drug prescribing.
The rate at which single-source drugs are prescribed depends on a number of factors, which can
vary across plans. For example, plans vary in their provider networks, geography, and patient demand for
newer or highly advertised drugs. For some conditions, single-source drugs are the only available
treatment. Therefore, the health status and preferences of beneficiaries may lead to differences across
plans in the rates of single-source drug use. Similarly, subsidized beneficiaries are largely insulated from
cost-sharing and as a result, have lower rates of generic use.
4

Multivariate Analyses
Our key outcome is the generic dispensing rate, defined as the percent of generic fills out of all
fills in a therapeutic class per year, weighted by the days supplied. We additionally examine the rate of
single source drug use within a class, which is the inverse of the generic utilization rate minus the
multisource use rate.
We regressed a set of beneficiary and plan characteristics in a logit model to assess the odds that a
given claim was for a generic drug, within a therapeutic class. The unit of analysis is the prescription drug
claim. We adjusted for several beneficiary characteristics that come from the Master Beneficiary
10

Summary and Enrollment files, including age, gender, race/ethnicity, beneficiary status (dual, low-income
subsidy, non-LIS) and health status, as measured by HCC scores.
We also controlled for a detailed set of plan characteristics using publicly available data from the
Centers for Medicare & Medicaid Services. We included each plan’s monthly premium and whether it
provides some coverage in the donut hole. Managed plans tend to have more control over physician
prescribing behavior and typically make greater use of utilization management tools, so we include a
binary indicator for a Medicare Advantage plan (MAPD). Analyses from 2010 onward also include the
percentage of drugs in each plan that require prior authorization or step therapy. Given that three firms --
UnitedHealth, Humana, and CVS Health -- account for over half of all Part D enrollees and the ten largest
sponsors account for nearly 90 percent of enrollment, we include binary indicators for each of the ten
largest plan sponsors to assess variation in generic use across parent companies.
2
Finally, we included
state, year and drug class fixed-effects, and used StataCorp Stata/MP 14.1 for all analyses. Significance
levels were determined using clustered standard errors at the plan level. Our primary regression results
report the years 2010-2014 as the indicator for whether the drug was subject to prior authorization
requests was available for these years. We provide the results for the full set of years (2006-2014) without
the prior authorization indicator in the online appendix.
Results
Following trends in the commercial market, the generic dispensing rate (GDR) in Part D reached
88 percent for MAPDs and 85 percent for PDPs in 2014, up from 63 and 53 percent (respectively) in
2006 (Exhibit 1). Generic use remained higher in Medicare Advantage plans compared with standalone
PDPs, but the difference narrowed over time. Generic drugs cost much less on average than their branded
counterpart, so the high rate of generic prescribing resulted in billions of dollars of savings to the
Medicare program.
11

Exhibit 2 shows the distribution of generic dispensing rates in 2014 at the plan-class level,
separately for PDPs and MAPDs. The ends of the boxes show the interquartile range of GDRs (25
th
and
75
th
percentiles), and the whiskers represent the 10
th
and 90
th
percentiles of Part D plans. What is
immediately evident in Exhibit 2 is the variation in generic use across drug classes, which is largely due
to different product cycles. Classes such as antiasthmatics were brand dominant in 2014, with average
generic dispensing rates of 20% to 25%. By contrast, drugs used to treat high blood pressure, cholesterol
and depression predominantly generic. Plan-level variation in generic use tends to be lower in generic
dominant classes such as antidepressants and higher in classes with over-the-counter substitutes such as
antihistamines and antiulcerants.
More importantly, Exhibit 2 highlights variation in generic use within a therapeutic class. For
example, in the treatment of type-2 diabetes, 10 percent of PDPs have generic prescribing rates of 50% or
less, while 1 in 10 plans have GDRs over 70%. Given that each plan faces the same set of treatment
options, large differences in medication use within a class may be due in part to non-clinical factors.
High generic/low brand use within a class could reflect less generous coverage or a more restrictive
formulary that requires prior authorization or step therapy before covering high cost brand drugs.
Alternatively, low generic/high brand use could reflect more generous brand coverage or differential
strategies by plans and PBMs that favor high cost, high rebate drugs.
To better understand how high and low GDR plans differed in their medication use, Exhibit 3
shows class-level market shares within selected classes among Part D plans in the top and bottom
quartiles of generic dispensing in 2014. In each of the four classes, plans in the bottom quartile of GDR
have significantly higher use of a single brand drug relative to the top-quartile GDR plans. For example,
esomeprazole (Nexium) was the 2nd most widely used antiulcerant in low GDR plans with an average
market share of 17.4%, but rarely filled in the highest GDR plans. Similarly, olopatadine, a branded
allergy medication, was the most widely used antihistamine in low GDR plans with a market share of
37.0% in 2014, yet almost never used in the highest GDR plans.
12

A possible explanation for these differences is that some plans are getting better rebates for a
single-source drug and therefore place them on a lower tier, with lower patient cost-sharing to encourage
use. We cannot know for certain whether this is the reason since rebates are proprietary, but we do see
this pattern occurring in some classes. For example, more than 6 out of 10 high GDR plans do not cover
Nexium, while a similar fraction of low GDR plans consider it a “preferred brand” (tier 3). Similarly,
brand antihistamines Pataday and Patanol (olopatadine) are commonly excluded from coverage in high
GDR plans and tier 3 products in most low GDR plans.
In the case of cholesterol lowering agents, higher use of a SS statin (Crestor) may reflect more
generous coverage, as recent evidence suggests that rosuvastatin (Crestor) is more effective in reducing
cardiovascular risk than other (generic) statins.
10
However, this information was not widely known in
2014 and thus is less likely to be driving the differences observed in Exhibit 3.
Thus far, our results have been descriptive and have not controlled for a range of beneficiary,
plan, and area characteristics that are likely to affect a physician’s or beneficiary’s choice of medication.
Exhibit 4 displays the odds ratios of filling a generic prescription from class-level regressions that control
for patient, plan and area characteristics (full results are in the Appendix). Low-income beneficiaries
eligible for both Medicare and Medicaid (dual eligible) and those receiving low-income subsidies were 19
and 24 percent less likely to receive a generic compared to non-subsidized beneficiaries, conditional on
receiving a drug in the class.
Plan-level factors also affect the mix of brand and generic drugs. Beneficiaries enrolled in plans
with higher premiums were less likely to fill a generic drug, presumably because coverage of brand drugs
was more generous in “enhanced”, higher premiums plans. Those enrolled in standalone drug plans
(PDPs) were 11 percent less likely to fill a generic prescription than similar beneficiaries in MAPDs,
although these differences moderated over time.
13

While there is a large body of evidence that patient out-of-pocket costs affect prescription drug
use and the choice of drug, formulary restrictions such as prior authorization did not affect generic
dispensing rates in our sample after controlling for other plan and beneficiary characteristics.
As we noted above, the Part D market is highly concentrated, with three firms—UnitedHealth, Humana,
and CVS Health—accounting for over half of all Part D enrollment in 2018 and nine firms having 88% of
the market.
2
Some of these plans also operate PBMs. We hypothesized that these and other organizational
features may play role in influencing generic usage. For example, managed care organizations typically
exert more control over provider’s prescribing patterns and thus tend to have greater formulary
compliance than independent practice associations (IPAs) or preferred provider organizations (PPOs).
Accordingly, we found that beneficiaries in Kaiser Permanente, a staff model HMO, were 76
percent more likely to fill a generic than beneficiaries in other plans, and all of Kaiser’s plans are MAPD.
They were followed closely by Wellcare, a large insurer predominantly operating in government markets
(Medicaid, Medicare Advantage and Part D). On the other end, generic dispensing rates were lowest in
plans offered by Express Scripts, CVS Caremark and United Healthcare, who operate as both PBMs and
Part D plans. For example, beneficiaries in UnitedHealth and Express Scripts plans were 16 and 13
percent (respectively) less likely to fill generics than other beneficiaries.
Beyond beneficiary and plan characteristics, prior work has shown that Medicare spending on
pharmaceuticals varies widely across areas of the country such as hospital referral regions, even after
adjustment for demographic characteristics, insurance coverage, and individual health status.
11
Across the
country, average GDRs across the 10 classes range from 0.82 to 0.88 (Exhibit A2 in the Appendix). The
states with the lowest predicted GDRs are New York, Texas and Louisiana, while the highest are
Massachusetts, the upper mid-west and the northwest.

14

Discussion
Despite high overall use of generic drugs in Part D, generic utilization rates within a therapeutic
class vary substantially across plans. Some of the variation is due to the mix of beneficiaries enrolled in a
plan. Dual eligibles and those receiving low-income subsidies face little or no cost-sharing and thus are
less price sensitive to high cost drugs. Plan features also matter. Generic dispensing rates are higher in
MAPDs compared to PDPs, although the gap is narrowing over time. Generic rates are also lower in
plans with higher premiums that typically have lower cost-sharing (reduced deductibles or copayments)
or offer some coverage in the donut hole that make higher cost drugs more accessible.
Nonetheless, significant variation persists after controlling for beneficiary and plan
characteristics, raising concerns that the program’s design and institutional features may incentivize
PBMs and health plans to favor high cost, high rebate drugs over lower cost alternatives. A widely cited
example of this is omeprazole (Nexium), which was the highest cost drug in Medicare Part D in 2013 at
$2.5 billion, despite widespread availability of generic substitutes and over-the-counter products,
including a “cousin” molecule, Prilosec. We found that plans in the lowest quartile of GDR have larger
shares of a singular brand-name medication than high GDR plans, which may not even cover drugs like
Nexium.
Why, in a competitive market, would Part D plans favor a higher cost drug? One possible
explanation is the growth in rebates and the increasing gap between list and net drug prices.
12
Rebates are
discounts off the list price, typically paid by manufacturers to pharmacy benefit managers and plans in
exchange for preferred formulary placement. Rebates are most common for high-cost brand drugs in
competitive therapeutic classes, where there are multiple choices of similar products and discounts can
exceed 50% of the list price. Within Medicare Part D, rebates have more than doubled as a share of total
spending since the program’s inception – from 9.6 percent in 2007 to 21.8 percent in 2017.
13

15

While rebates lower the plan’s overall costs, a majority of drugs in PDPs now require coinsurance
rather than a copayment (but not in MAPDs). This too incentivizes plans to prefer high price, high rebate
drugs, where the beneficiary bears a large fraction of the cost at the point of sale. While there are limits
on plan profits, plans may find it strategically better to offer lower premiums and higher cost-sharing at
the point of sale, which is what the rebate game fosters. In addition, high list prices accelerate
beneficiaries’ progression through the benefit phases, shifting a larger share of drug spending to the
catastrophic phase where plan liability is low. In fact, after accounting for beneficiary OOP costs,
mandated manufacturer discounts in the coverage gap, and the federal reinsurance program, Part D plans
pay a modest portion of total prescription drug costs. Further, in some cases, a very high rebate can
exceed the Part D plan’s liability for the drug. This creates an incentive in some cases to prefer a high-
cost, high-rebate drug over a lower priced alternative despite an increase in costs to beneficiaries and the
federal government.
14

Unlike commercial plans, a large fraction of Part D beneficiaries receives subsidies that reduce or
eliminate out-of-pocket costs. Plans with a large proportion of subsidized beneficiaries may design
formularies that favor brands with high rebates over lower cost brands or generics, as these beneficiaries
are insulated from the impact of coinsurance on brand-name drugs. A recent analysis examined the 57
unique drug formularies offered across 750 PDP plans in 2016 to determine how often branded products
were given more favorable formulary placement than generic products, where favorable placement was
defined as being on a lower cost-sharing tier or having fewer restrictions such as prior authorization, step
therapy, or quantity limits.
15
The authors found that 41 of the 57 formularies placed at least 1 branded
product in a lower cost-sharing tier than its generic product, and all formularies had at least 1 multisource
drug covered without a generic product.
Stand-alone PDPs manage prescription drug coverage only, while MAPD plans are at risk for
both medical and drug expenditures. These differences can lead the two types of plans to design different
formularies. Recent work suggest that MAPDs provide more generous coverage of essential drugs,
16

particularly those for which increased adherence reduces medical expenditures, e.g. antihypertensive and
antidiabetic therapies, which is consistent with our findings of modestly higher generic dispensing rates in
MAPDs.
16,17

Less expected was the degree of variation in generic rates across the largest Part D sponsors, with
markedly lower generic utilization among firms operating as both a PBM and drug plan, specifically
UnitedHealthcare, CVS and Express Scripts. The Part D marketplace is dynamic and continues to
undergo consolidation, particularly since 2014, our last year of data for this study. In the interim years,
several large mergers have been completed or proposed between Anthem and WellPoint (both BCBS
affiliates),
18
CVS Health and Aetna,
19
Express Scripts and Cigna,
20
Centene and WellCare.
21
While
Anthem and WellPoint merged in 2004, WellPoint did not begin operating as Anthem until 2014).
18

Torchmark sold its Part D plans to Silverscript in 2016, which is a subsidiary of CVS.
22
By 2018, three
insurers—UnitedHealthcare, Humana and CVS Health—held approximately two-thirds of the stand-alone
Part D market.
2
Consolidation at the parent level may allow the largest insurers to negotiate better
discounts with manufacturers, but the pendulum of leverage may have swung too far towards the largest
PBMs and health plans. As companies continue to consolidate in the Part D space, the potential for
rebates to drive utilization become greater.
Our analyses are limited by the fact that 2014 is the most recent year of claims data available to
us. Thus, to the extent that that trends in generic use have changed in more recent years, we are unable to
capture these dynamics in our analyses. There can be important clinical differences among drugs
approved to treat the same condition. Some drugs in the class may be more effective or have fewer side-
effects, at least for some patients, and sold at different strengths and/or with dosing regimens, and
physicians may have different experiences that affect their prescribing behavior. These unmeasured
factors could bias our results if they differ systematically differ across plans, but there is no evidence to
suggest it is a concern. Finally, we find suggestive evidence that rebates may be playing a role in
discouraging generic utilization given the lack of rebate data. The Trump administration released a
17

proposal this year to apply rebates at the point of sale that the administration abandoned in July.
23,24
While
the impacts of this proposal on beneficiaries, plans and the federal government were uncertain, this policy
would reduce the incentive to place branded drugs on lower cost sharing tiers thus increasing generic use
rates.
Greater use of generic medicines is one way to constrain growth in health care spending.
Nonetheless, it is important to be realistic about what can be achieved. Despite some widely publicized
examples of excess profiteering in generic markets, most of the growth in drug spending will continue to
be driven by new medicines. For some treatments, like certain cancer immunotherapies, the complex
manufacturing process means that the scope for off-patent products is still limited. Yet there are
opportunities for significant cost savings from generics. One option is for CMS to prohibit Part D plans
from giving branded products more favorable formulary placement than generic products. An alternative
is to change the incentives in Part D, as intended by the proposal by the Department of Health and Human
Services to ban rebates in Medicare Part D. Another option is to implement an out-of-pocket cap for
prescription drug expenditures, akin to the proposed bill by Senator Wyden that would eliminate
beneficiary cost-sharing in the catastrophic phase.
25

Conclusions
We find that overall generic utilization is increasing in Part D, but there are persistent variations
across plans. We find that beneficiary and plan characteristics, state and parent organization all are factors
contributing to differences in the mix of brand and generic drugs within a therapeutic class.
Variation in generic use does have real consequences for plans, beneficiaries and taxpayers.
Physician prescribing habits and beneficiaries’ perceptions of the quality of generic drugs are harder to
tackle for policy makers, but plan decisions and benefit design are easier. Focusing on changing the
underlying benefit design to reduce incentives to use high-cost drugs or encouraging the use of more
18

tightly managed formularies that restrict beneficiaries’ access to newer expensive brand medications
within a therapeutic class could increase generic use.
Like medical spending, drug spending is variable and thus warrants greater attention to better
understand optimal prescribing and the role of incentives for plans and intermediaries in the supply chain
to promote one type of a drug over another. We find evidence that rebates may be playing a role in
discouraging generic use in particular therapeutic classes, but because there is no public information on
the extent of these rebates we cannot know for sure. This lack of knowledge makes it difficult to pinpoint
an exact cause where a policy solution could be crafted.

19

References
1. Dusetzina SB, Jazowski S, Cole A, Nguyen J. Sending The Wrong Price Signal: Why Do Some
Brand-Name Drugs Cost Medicare Beneficiaries Less Than Generics? Health Affair. 2019;38(7):1188-94.
2. Cubanski J, Damico A, T. N. Medicare Part D in 2018: The Latest on Enrollment, Premiums, and
Cost Sharing. Kaiser Family Foundation, 2019 5/17/19. Report No.
3. MedPAC. Payment Basic: Part D Payment System. Medicare Payment Advisory Commission,
2019 http://medpac.gov/docs/default-source/payment-
basics/medpac_payment_basics_18_partd_final_sec.pdf?sfvrsn=0. Report No.
4. MedPAC. March 2018 Report. Chapter 14: The Medicare prescription drug program (Part D):
Status report. Medicare Payment Advisory Commission, 2018 March 2018. Report No.
5. H.R.1892 - Bipartisan Budget Act of 2018, 115-123 (2018).
6. Cubanksi J. Summary of Recent and Proposed Changes to Medicare Prescription Drug Coverage
and Reimbursement. Kaiser Family Foundation; 2018; Accessed 10/22/19, from:
https://www.kff.org/medicare/issue-brief/summary-of-recent-and-proposed-changes-to-medicare-
prescription-drug-coverage-and-reimbursement/.
7. Cubanski J NT, Damico A. . Closing the Medicare Part D Coverage Gap: Trends, Recent
Changes, and What’s Ahead. . Kaiser Family Foundation, 2018.
8. Hoadley J, Cubanksi J, T N. Medicare Part D at Ten Years: The 2015 Marketplace and Key
Trends, 2006–2015. Kaiser Family Foundation.
9. CBO. Effects of Using Generic Drugs on Medicare’s Prescription Drug Spending. Congressional
Budget Office, 2010 Contract No.: 21800.
10. Kaneko K, Saito H, Sasaki T, Sugawara S, Akasaka M, Kanaya T, Kubota I. Rosuvastatin
prevents aortic arch plaque progression and improves prognosis in ischemic stroke patients. Neurol Res.
2017;39(2):133-41.
20

11. Zhang Y, Baicker K, Newhouse JP. Geographic variation in Medicare drug spending. N Engl J
Med. 2010;363(5):405-9. PMCID: PMC3364516.
12. GAO. Use of Pharmacy Benefit Managers and Efforts to Manage Drug Expenditures and
Utilization. United States Government Accountability Office, 2019 Contract No.: GAO-19-498.
13. Medicare Board of Trustees. 2019 Annual Report of the Boards of Trustees of The Federal
Hospital Insurance and Federal Supplementary Medical Insurance Trust Funds. 2019.
14. Gabriela Dieguez, Maggie Alston, Tomicki S. A primer on prescription drug rebates: Insights into
why rebates are a target for reducing prices. Milliman; 2019; Accessed 10/22/19, from:
http://www.milliman.com/insight/2018/A-primer-on-prescription-drug-rebates-Insights-into-why-rebates-
are-a-target-for-reducing-prices/.
15. Socal MP, Bai G, Anderson GF. Favorable Formulary Placement of Branded Drugs in Medicare
Prescription Drug Plans When Generics Are Available. JAMA Intern Med. 2019;179(6):832-3. PMCID:
PMC6547248.
16. Lavetti K, Simon K. Strategic Formulary Design in Medicare Part D Plans. Am Econ J Econ
Policy. 2018;10(3):154-92. PMCID: PMC6335045.
17. Starc A, Town R. Internalizing Behavioral Externalities: Benefit Integration In Health Insurance.
NBER Working Paper. 2015;Working Paper 21783.
18. Anthem. Company History. 2019.
19. Health C. CVS Health Completes Acquisition of Aetna. 2018.
20. Cigna. Cigna completes combination with Express Scripts. 2018.
21. Livingston S. Centene-WellCare deal nabs shareholder approval. Modern Healthcare. 2019
6/24/19.
22. SEC. Torchmark Corporation: Form 8-k. United States Securities And Exchange Commission;
2016; Accessed 7/23/19, from:
https://www.sec.gov/Archives/edgar/data/320335/000032033516000088/a8-kpartdsale.htm.
21

23. CMS. Proposed Safe Harbor Regulation. Centers for Medicare & Medicaid Services; 2018;
Accessed 7/23/19, from: https://www.cms.gov/Research-Statistics-Data-and-
Systems/Research/ActuarialStudies/Downloads/ProposedSafeHarborRegulationImpact.pdf.
24. Thomas K, Goodnough A. Trump’s Efforts to Rein In Drug Prices Face Setbacks. New York
Times. 2019 7/11/2019.
25. Wyden R. Wyden Proposes Cap on Prescription Drug Costs. 2016 updated 4/27/2016 from:
https://www.finance.senate.gov/ranking-members-news/wyden-proposes-cap-on-prescription-drug-costs.

22

Tables and figures
Exhibit 1: Generic/Multisource/Single Source dispensing rates by plan type (MA-PD and PDP plans),
2006-2014

Notes: GDR = generic dispensing rate; SS = single source; PDP = stand-alone prescription drug plans;
MA = Medicare Advantage prescription drug plans. This figure includes all drugs dispensed.

0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2006 2007 2008 2009 2010 2011 2012 2013 2014
Percent of Fills
MA PDP
Generic
Multi-Source
Single Source
23

Exhibit 2: Class-specific GDR distribution by MAPD and PDP plans 2014

Notes: GDR = generic dispensing rate; SS = single source; PDP = stand-alone prescription drug plans;
MA = Medicare Advantage prescription drug plans. This figure is limited to plans with more than 200
enrolled beneficiaries.

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Antiasthmatics
Anticoagulants
Anticonvulsants
Antidepressants
Antidiabetics
Antihistamines
Antihyperlipidemics
Antihypertensives
Antineoplastics
Antiulcerants
MAPD PDP
24

Exhibit 3: Market share for top active ingredients comparing plans in the top and bottom quartile of GDR,
by selected classes
Plans in top-quartile of GDR Plans in bottom quartile of GDR
Class Generic name
Market
share (%)
Modal tier for
selected SS
Generic name
Market
share (%)
Modal tier for
selected SS
Antihistamine Hydroxyzine (G ) 33.5 Olopatadine (SS ) 39 Tier 3

Promethazine (G) 33.3

Hydroxyzine (G) 23.7

Loratadine (G) 9.7

Levocetirizine (G) 11.6

Levocetirizine (G) 8.5

Promethazine (G) 11.4

Cetirizine (G) 4.7

Azelastine (G) 4.3

Olopatadine (SS) 3.8
Tier 3 OR Not
covered
Cyproheptadine 2.1

Azelastine (G) 1.2

Loratadine (G) 1.2
Antiulcerants Omeprazole(G) 52.3 Omeprazole (G) 42.1

Pantoprazole (G) 17.7

Esomeprazole (SS ) 17.4 Tier 3

Ranitidine (G) 15.1

Pantoprazole (G) 13.9

Famotidine (G) 7

Ranitidine(G) 10.7

Lansoprazole (G/MS) 3.6

Famotidine(G) 6.3

Esomeprazole (SS) 0.88 Not covered Lansoprazole (G/MS) 2.1
Anticonvulsants Gabapentin 46.8

Gabapentin 35.2

Clonazepam 21.9

Clonazepam 15.6

Lamotrigine 4.7

Pregabalin (SS) 10.1 Tier 3

Levetiracetam 4.4

Divalproex 8.7

Divalproex 5.6

Levetiracetam 6.1

Pregabalin (SS) 3.4 Tier 4 Lamotrigine 5.9
Antihyperlipide
mics
Simvastatin (G) 38.3

Atorvastatin (G) 28.8

Atorvastatin (G) 29.6

Simvastatin (G) 26.5

Pravastatin (G) 11.4

Pravastatin (G) 12.3

Lovastatin (SS) 8.8

Rosuvastatin (SS) 11 Tier 3

Rosuvastatin (SS) 1.8 Tier 3 Ezetimibe (SS) 5.3

Ezetimibe (SS) 1.7 Lovastatin (SS) 3.8
Notes: G = generic; MS = multisource; SS = single source. This figure is limited to plans with more than
200 beneficiaries.
25

Exhibit 5: Selected regression results

Notes: The parent company designations are as of 2014. BCBS Affiliates includes Anthem, Wellpoint (which were
separate companies at the time) and any parent company selling Blue Cross and/or Blue Shield plans.

-2%
-7%
-1%
-19%
-24%
-3%
-15%
-20%
-24%
-11%
5%
-5%
-3%
2%
-2%
-15%
4%
-16%
1%
2%
-13%
-9%
76%
67%
-7%
5%
-40% -20% 0% 20% 40% 60% 80% 100%
Age
White
LIS
Premium, $21-30
Premium, $40+
Gap coverage, Tier 1
Percent of fills with PA
Number of drug classes, 5-7
Humana
Aetna
Express Scripts
Kaiser
BCBS affiliate
< Less Likely | More Likely >
26

Appendix
A1: Full regression results
Variables 2010-2014 2006-2014
Age, odds ratio 0.9762*** 0.9871***
[95% confidence interval] [0.974 - 0.978] [0.986 - 0.989]
Age, squared 1.0002*** 1.0001***

[1.000 - 1.000] [1.000 - 1.000]
Female 0.9267*** 0.9040***

[0.921 - 0.932] [0.899 - 0.909]
White 0.9863** 0.9935

[0.974 - 0.999] [0.982 - 1.005]
Dual 0.8124*** 0.8504***

[0.790 - 0.836] [0.832 - 0.870]
LIS 0.7583*** 0.7979***

[0.744 - 0.772] [0.785 - 0.811]
AK 1.0799 1.1001*

[0.967 - 1.206] [0.994 - 1.218]
AL 1.2131*** 1.2016***

[1.108 - 1.329] [1.102 - 1.310]
AR 1.1736*** 1.1578***

[1.071 - 1.286] [1.068 - 1.255]
AZ 1.2376*** 1.3103***

[1.120 - 1.368] [1.190 - 1.443]
CA ref ref

CO 1.1936*** 1.2354***

[1.098 - 1.297] [1.146 - 1.332]
CT 0.9162* 0.9231*

[0.833 - 1.007] [0.849 - 1.004]
DC 1.0163 1.0135

[0.904 - 1.142] [0.919 - 1.118]
DE 0.9475 0.9594
27

[0.851 - 1.055] [0.865 - 1.063]
FL 1.0365 1.0562

[0.938 - 1.145] [0.956 - 1.167]
GA 1.0810* 1.0992**

[0.987 - 1.184] [1.012 - 1.194]
HI 1.0009 0.9472

[0.868 - 1.154] [0.837 - 1.072]
IA 1.3604*** 1.3250***

[1.242 - 1.490] [1.212 - 1.448]
ID 1.3023*** 1.2065***

[1.186 - 1.430] [1.108 - 1.314]
IL 1.2181*** 1.2311***

[1.110 - 1.337] [1.135 - 1.336]
IN 1.1167** 1.1216***

[1.019 - 1.224] [1.033 - 1.218]
KS 1.1916*** 1.1840***

[1.086 - 1.307] [1.083 - 1.295]
KY 1.2236*** 1.2478***

[1.115 - 1.342] [1.150 - 1.354]
LA 0.9112* 0.9256*

[0.828 - 1.003] [0.850 - 1.008]
MA 1.5847*** 1.5094***

[1.445 - 1.738] [1.393 - 1.635]
MD 1.0507 1.0339

[0.957 - 1.154] [0.952 - 1.123]
ME 1.1543** 1.2183***

[1.018 - 1.309] [1.096 - 1.355]
MI 1.3610*** 1.3205***

[1.243 - 1.490] [1.213 - 1.437]
MN 1.3781*** 1.2897***

[1.263 - 1.504] [1.181 - 1.408]
MO 1.2086*** 1.2653***

[1.104 - 1.324] [1.168 - 1.370]
28

MS 1.1096** 1.1239***

[1.012 - 1.216] [1.035 - 1.221]
MT 1.2457*** 1.2420***

[1.137 - 1.365] [1.146 - 1.346]
NC 1.1067** 1.0870**

[1.008 - 1.215] [1.001 - 1.181]
ND 1.1656*** 1.1985***

[1.055 - 1.287] [1.087 - 1.322]
NE 1.2210*** 1.1914***

[1.120 - 1.331] [1.098 - 1.292]
NH 1.1964*** 1.2170***

[1.090 - 1.313] [1.121 - 1.321]
NJ 0.8142*** 0.7966***

[0.735 - 0.902] [0.732 - 0.866]
NM 1.2812*** 1.3299***

[1.171 - 1.402] [1.228 - 1.440]
NV 1.3526*** 1.3573***

[1.187 - 1.541] [1.208 - 1.526]
NY 0.8721*** 0.8754***

[0.791 - 0.962] [0.805 - 0.952]
OH 1.1730*** 1.1546***

[1.070 - 1.286] [1.067 - 1.250]
OK 1.1452*** 1.1522***

[1.047 - 1.252] [1.064 - 1.247]
OR 1.3827*** 1.3079***

[1.250 - 1.529] [1.194 - 1.433]
PA 1.1085** 1.0950**

[1.003 - 1.225] [1.011 - 1.186]
RI 1.3772*** 1.3202***

[1.260 - 1.506] [1.220 - 1.429]
SC 0.9953 1.0014

[0.909 - 1.090] [0.924 - 1.086]
SD 1.2839*** 1.2527***
29

[1.168 - 1.412] [1.150 - 1.365]
TN 1.2214*** 1.2536***

[1.119 - 1.333] [1.154 - 1.362]
TX 0.9233* 0.9500

[0.843 - 1.011] [0.877 - 1.029]
UT 1.2210*** 1.1767***

[1.122 - 1.329] [1.091 - 1.269]
VA 1.1653*** 1.1864***

[1.066 - 1.274] [1.097 - 1.283]
VT 1.3159*** 1.2458***

[1.195 - 1.449] [1.139 - 1.363]
WA 1.4135*** 1.3897***

[1.288 - 1.552] [1.281 - 1.508]
WI 1.3465*** 1.3309***

[1.225 - 1.480] [1.227 - 1.444]
WV 1.0654 1.0847*

[0.966 - 1.175] [0.996 - 1.182]
WY 1.2346*** 1.2148***

[1.123 - 1.357] [1.121 - 1.316]
Enrollment over 1000 0.9747*** 0.9896

[0.956 - 0.994] [0.971 - 1.008]
Premium <$20 ref ref

Premium 21-30 0.8509*** 0.8966***

[0.826 - 0.876] [0.873 - 0.921]
Premium 31-40 0.7965*** 0.8427***

[0.773 - 0.820] [0.819 - 0.867]
Premium >40 0.7611*** 0.8260***

[0.731 - 0.792] [0.797 - 0.856]
Prescription Drug Plan 0.8899*** 0.8290***

[0.864 - 0.917] [0.805 - 0.853]
Gap coverage, Tier 1 1.0521*** 1.0588***

[1.017 - 1.089] [1.029 - 1.089]
Gap coverage, Tier 2 0.9543** 0.9894
30

[0.918 - 0.992] [0.954 - 1.026]
Humana 1.0447*** 0.9116***

[1.011 - 1.079] [0.877 - 0.947]
UnitedHealth 0.8415*** 0.8447***

[0.807 - 0.877] [0.815 - 0.875]
Aetna 1.0086 0.9901

[0.967 - 1.052] [0.948 - 1.034]
Cigna 1.0208 1.0180

[0.985 - 1.058] [0.986 - 1.051]
Express Scripts 0.8666** 0.8839**

[0.775 - 0.969] [0.801 - 0.975]
CVS 0.9085*** 0.9228***

[0.862 - 0.957] [0.885 - 0.962]
Kaiser 1.7561*** 1.9425***

[1.611 - 1.914] [1.810 - 2.085]
Wellcare 1.6706*** 1.6784***

[1.540 - 1.812] [1.579 - 1.784]
BCBS 0.9274*** 0.8951***

[0.891 - 0.965] [0.865 - 0.927]
Torchmark 1.0505 0.9758

[0.966 - 1.143] [0.912 - 1.044]
2006

ref

2007

1.1552***

[1.135 - 1.175]
2008

1.4276***

[1.399 - 1.457]
2009

1.6962***

[1.657 - 1.736]
2010 ref 2.0246***

[1.974 - 2.077]
2011 1.1511*** 2.3150***

[1.133 - 1.170] [2.252 - 2.380]
2012 1.6411*** 3.2592***
31

[1.609 - 1.673] [3.166 - 3.355]
2013 2.3380*** 4.5755***

[2.265 - 2.413] [4.422 - 4.734]
2014 2.7864*** 5.3724***

[2.698 - 2.878] [5.197 - 5.554]
Antiasthmatics ref ref

Anticoagulants 10.7166*** 7.4520***

[10.303 - 11.146] [7.166 - 7.750]
Anticonvulsants 42.4618*** 20.2320***

[40.242 - 44.804] [19.300 - 21.209]
Antidepressants 43.6316*** 25.8011***

[41.461 - 45.915] [24.628 - 27.030]
Antidiabetics 8.1589*** 7.5205***

[7.774 - 8.563] [7.254 - 7.797]
Antihistamines 36.8340*** 29.0036***

[31.999 - 42.399] [26.689 - 31.519]
Antihyperlipidemics 16.8092*** 9.4363***

[15.835 - 17.844] [8.956 - 9.942]
Antihypertensives 32.1475*** 18.9799***

[30.050 - 34.391] [18.036 - 19.974]
Antineoplastics 27.8976*** 15.0869***

[26.400 - 29.480] [14.458 - 15.743]
Antiulcerants 27.8074*** 15.1506***

[25.424 - 30.414] [14.162 - 16.208]
Percent of drugs with Prior
Authorization/Step Therapy 0.9734

[0.852 - 1.112]

Number of classes beneficiary is
taking (1-3) ref ref

Number of classes (4-5) 1.0219*** 1.0352***

[1.013 - 1.031] [1.028 - 1.042]
Number of classes (6-7) 0.9771*** 1.0061

[0.965 - 0.990] [0.996 - 1.016]
32

Number of classes (7+) 0.8460*** 0.8955***

[0.831 - 0.861] [0.883 - 0.908]

A2: Predicted Generic Dispensing Ratio by State, 2014

33

Chapter 2 : Long-Term Persistence of Prescription Drug Spending
in Medicare Part D

Introduction
It is well documented that the distribution of healthcare expenditure is highly skewed, particularly
in the U.S., with a small fraction of people accounting for a disproportionate share of the total spending in
any given year. For example, several studies have found that 5 percent of people were responsible for
nearly half of total medical care spending in the year. (Berk and Monheit, 2001, NIHCM, 2017) However
there is considerable heterogeneity even among the highest spenders.(Claxton et al 2019) Some have high
spending due to an acute illness, resulting in in a hospitalization or expensive medical treatment, while
others require frequent outpatient visits that may require use of expensive medications. While high
medical spending for an acute illness tends not to persist, high spenders receiving treating for a chronic
illness are more likely to face high medical bills in subsequent years.
Prior work examining persistence in drug spending finds greater stability over time, particularly
among the heaviest users.(Wrobel et al., 2003) However, few studies look at the persistence of
prescription drug spending over extended periods of time, particularly among the elderly who make up a
disproportionate share of prescription drug use in any year. Understanding the degree of persistence in
prescription drug spending among the elderly is highly pertinent given the growth in high cost specialty
drugs to t treat chronic conditions such as multiple sclerosis, rheumatoid arthritis, and diabetes. In
addition, most of the elderly are getting their prescription drug coverage through Medicare Part D, where
there is no cap on beneficiary out-of-pocket spending. High drug costs and persistent use will result in
some beneficiaries facing high out-of-pocket spending every year.
34

In this study, we examine the persistence of prescription drug spending among the elderly for up
to 7 years and the out-of-pocket burden associated with it.
Relevant work
Several studies have documented the concentration and persistence of medical spending among
the elderly, using Medicare fee-for-service claims. Garber et al (1998) found considerable turnover
among people in the top 5th percentile of the spending distribution. Only 17.9% of those in the top 5
percent of spending in a year remained in the highest spending group the next year, and 10.8% remained
two years later.(Garber et al., 1997) The lack of persistence in medical spending was attributable to the
acute nature of illness, in which most beneficiaries recover or decease. Accordingly, the study found
markedly higher mortality rates in succeeding years for beneficiaries in the top 5
th
percentile of spending.
By contrast, there is greater persistence in prescription drug spending compared to medical
services. Coulson and Stuart (1992) examined the persistence in pharmaceutical spending among the
elderly in the Pennsylvania Pharmaceutical Assistance Contract for the Elderly (PACE) program. They
found substantial persistence in prescription drug spending at both ends of the distribution, concluding it
can take a substantial length of time to show regression to the mean.(Coulson and Stuart, 1992)
More recent relevant work by Hirth et al (2016) examined the long-term persistence in health care
spending among a cohort of privately insured in the U.S using claims data from Truven Market Scan.
They found that while prescription drug spending was less concentrated (e.g. top 5% spenders account for
40% of the total drug spending) than medical spending (e.g., top 5% accounts for 55% of the total), the
correlation over time was substantially higher for drug spending.(Hirth et al., 2016) For example, the
correlation of drug spending correlation by the fifth year (0.57) was still two-thirds as large as it was at
one year (0.85), while the correlation in medical spending was 0.13 and 0.31 at five and one years,
respectively.
35

Other studies examined the persistence in prescription drug spending have found that prior year’s
spending was more predictive of future spending, even after controlling for health, medication use and
other characteristics.(Banthin and Miller, 2006; Wrobel et al., 2003) For example, by using two cohort
years’ data from Medical Expenditure Panel Survey, Banthin and Miller have found that about half of the
variation in prescription drug expenditures can be predicted using information from the prior year, when
including therapeutic categories based on drug utilization.
Methods
Data and sample
We use a nationally representative 20 percent random sample of Medicare beneficiaries enrolled
in Part D (include both stand-alone prescription drug plans and Medicare-Advantage plans with
prescription drug benefits), through a data use agreement with the Centers for Medicare and Medicaid
Services. The primary data for the study is the Prescription Drug Event (PDE) files from 2006-2014. PDE
files contain detailed information of each claim, including date, total cost and out-of-pocket cost, NDC of
the drug, and encrypted beneficiary identifier and plan identifier. The beneficiary identifier allows the
claims to be linked to the Master Beneficiary Summary File, which contain information of demographics,
social-economic status, and mortality. Plan identifier links the PDE files to the plan characteristic files.
Claims include all self-administered prescription drug covered under Medicare Part D. Any drug requires
out-patient visit are not included, and are under Medicare Part B.
Our sample includes Medicare beneficiaries who have Part D claims during the study period and
live in the continental states in the U.S. We exclude anyone who have employer-sponsored drug benefit
coverage at the time, as the benefit design of these plans are different from the general Part D plans and
there is very little information about the plans.
In each year, the sample includes over 7 million beneficiaries with a drug claim(s) and more than
200 million claims per year. We follow more than 2 million beneficiaries for the entire eight-year period
36

from 2007 to 2014. The availability of eight recent years of fully adjudicated claims of millions of
enrollees allows us to substantially improve on previous studies in terms of timeliness, length of follow-
up, and sample sizes.
Prescription drug spending
We calculate each beneficiary’s annual total drug spending by aggregating the total cost
1
of all
outpatient prescription drug claims in the year. Individuals are categorized into different spending groups
based on their percentile of the distribution. Following Garber et al (1998), we define the low spending
group as those in the bottom half (0-50%) of the spending distribution, middle spenders between the 50
th

to 95% percentiles, and the high spending group in the top 5% of the drug spending distribution in the
year. All dollars are adjusted for inflation using Consumer Price Index for prescription drugs.
We show the distribution of prescription drug spending both cross-sectionally and over time. For
spending concentration, we calculate the proportion of overall drug expenditures attributable to the
spending groups in selected years (2007, 2010 and 2014). To measure persistence, we group beneficiaries
into the three spending groups (low, medium, and high) based on their 2007 spending level and follow
their drug spending over time. We then show the dynamics of same beneficiary moving between spending
groups over time by computing the transition probabilities in each year, again using 2007 as baseline. We
compute trends in drug spending for a fixed cohort of survivors, as well as for a changing sample of
beneficiaries that allows for mortality. We track both total spending and beneficiary out-of-pocket
spending for the different cohorts.
As our methods have been broadly followed those in the Garber et al
2
, we repeat the persistence
analysis by using 2010 as the baseline and present the results as a sensitivity analysis.

1
The totalcst variable from the data is our primary variable for calculating spending, it is the sum of both plan paid
amount and beneficiary’s out-of-pocket amount.
2
In their study, they follow beneficiaries’ medical spending in the previous and following years.
37

Results
Distribution of prescription drug spending over time (figure 1-3)
Figure 1 shows total Part D spending over time
3
. Despite increased concern over rising drug
prices, the growth in prescription drug spending has been relatively modest in Part D from 2006 to 2013
after adjusting for inflation. Drug spending increased dramatically in 2014, primarily due to the
introduction and rapid dissemination of high-cost specialty drugs, most notably novel therapies to treat
Hepatitis C.
Figures 2 and 3 show the distribution of total spending and average annual spending by percentile
groups by year. Beneficiaries in the top 5 percent of the spending distribution in 2007 spent $17,162 on
average, rising to $26,614 by 2014. The increase is particularly concentrated among the top 2 percent of
spenders, where average annual spending increased from $21,791 in 2007 to $38,226 in 2014, a 150%
cumulative growth. By contrast, spending among the other groups was relatively flat after adjusting for
inflation. In addition to the increase in average spending among the top 5% beneficiaries, the total
spending attributable to this group of beneficiaries have grown over time. In 2007, the top 5% people
constitute 31.4% of all spending in our sample. By 2014, nearly half (48.3%) of the total prescription drug
spending are concentrated among the 5% users. The top 2% alone, has consumed a third of the total
spending.
Persistence in prescription drug spending (figure 4, 5, table 1)
To better understand the long-term persistence of drug spending, we show the average total
annual spending over time, by using 2007 spending groups
4
as the baseline and follow the cohort over a
7-year period. We compute the average spending for both survivors and decedents in the following years,
and survivors who stayed in our sample throughout the study period (a sample of 2.1 million

3
The overall spending has been scaled up to a 100% data.
4
The beneficiaries are defined as having low (0-50
th
), middle (51-95
th
) and high (95
th
up) spending based on the
distribution of annual spending in 2007.
38

beneficiaries). In 2007, the average annual spending on prescription drug is $14,273 for beneficiaries in
the high spending group (i.e., the top 5%), 21 times as large for those in the lower half of the distribution
($682). Total spending of the high spenders declines slightly over time, except for 2014. In 2014, those
who remain in the top 5% of spending still expend 8 times as much as beneficiaries in the lowest
spending group. We also find increasing drug spending by those in the low spending group, rising from
$682 in 2007 to $1,606 in 2014. The increase is likely due to beneficiaries aging and developing chronic
conditions over this period. Restricting the sample to those who survived over the 7-year period changed
the level, but not the trend in spending over time (Figure 4(b)). By using 2010 as baseline, the pattern is
similar in terms of persistently high spending for the top 5% spenders in the following years. These
people also have high spending in previous years (Figure 7).
Figure 5 shows the persistence in out-of-pocket spending since 2007. Similar to the pattern for
total spending, beneficiaries who have high drug expenditures in 2007 have continuously high out-of-
pocket spending in subsequent years, and the out-of-pocket spending for the low spending group has been
increasing and doubled over the 7-year period.
Table 1 shows transition probabilities of moving from one spending category to another based on
the 2007 spending level. Low, middle and high spending groups in each subsequent year is based on the
spending distribution in that year. We restrict the sample to beneficiaries who survived till 2014 for this
analysis and examine mortality separately. The key finding from Table 1 is the high degree of persistence
for each group. For example, 78.3% of those in the top 5% of spending in 2007 remain in top spending
group in the following year, and nearly half remain after 7 years. The same degree of persistence is found
among the middle and low total spenders in 2007.
Cumulative mortality by spending groups (figure 6)
Figure 6 shows cumulative mortality rates from 2008 to 2014 for the three spending groups
defined using 2007 spending. While there is a positive correlation between drug spending and mortality,
the association is modest. Differences in cumulative mortality between high and middle spending groups
39

vary from 1% to 2%, suggesting that high prescription drug spending is not a reliable predictor of
mortality.
Discussion
In this study, we examine both the concentration and the long-term persistence of prescription
drug spending among the elderly population covered by Medicare Part D. First, we see substantial
concentration in prescription drug spending, like other studies have found and a similar pattern to the
distribution of medical spending. People at the right extreme of the distribution spend a disproportionate
share of the total prescription drug expenditure. In addition, we find the spending distribution has become
more skewed over time, with the top 2% spenders account for over a third of the total spending, up from
20% when Part D started. The stronger concentration in Part D spending is due to the price hike of
existing pharmaceutical products and the growth in novel therapies with high price tag. In 2014, Medicare
Part D has spent $4.5 billion on Hepatitis C virus drugs, compare to $283 million in 2013. This explains
the jump in total spending and average spending by the top percentile group we see in 2014.
Secondly, in contrast to medical spending, we observe substantial persistence in prescription drug
spending in the Medicare population. In particular, nearly half of the people who are once among the high
spending group (i.e., top 5%) remain as a top 5% spending after 7 years. The high degree of spending
persistence raises concerns on the out-of-pocket cost burden on the elderly population. In addition, we see
mortality merely has effect on mitigating the persistent high spending, due to the chronic nature of illness.
To our knowledge, we are the first to use the Medicare Part D data to study the long-term
persistence among elderly population in the recent decade. Several other studies have demonstrated the
strong persistence in prescription drug spending, but most of them relied on older and short-term data.
The one recent study describing long-term prescription drug spending by Hirth (2016) focused on people
under 65 years old and this population are generally healthier than our sample. Focusing on elderly
population has several policy implications. First, the elderly population make up 15% of the U.S
40

population but consume 35% of all prescription drug. Second, many of the novel therapies aim to treat
chronic conditions, such as rheumatoid arthritis, diabetes, high cholesterol, neurologic disorders, and the
elderly are at a much higher risk to develop these conditions or at a more advanced stage of the disease.
Third, most of the elderly are getting the prescription drug coverage from Medicare, where there is no cap
on limiting out-of-pocket spending.
Our finds of substantial degree of persistence in prescription drug spending means that the same
people are experiencing high prescription drug costs year after year. As costs grow, particularly for
specialty drugs and novel therapies, patients can face high cost-sharing burdens and are at higher risk of
reaching the catastrophic phase, where they need to pay 5% of the gross, non-discounted price of all
prescription drugs for the remainder of the year. A recent study by Trish et al (2018) find that compared
to 2007, there has been considerable increase in the number and proportion, as well as the magnitude of
spending of non-low-income subsidy beneficiaries reaching catastrophic phase in 2015.(Trish et al., 2018)
There are several limitations of the study that need to be noted. First, the DUA with CMS during
the period of this project only allows us to examine data up to 2014. With data from more recent years, it
could allow us to examine long-term persistence by using a baseline cohort from a more recent year,
where people may have a different level of spending due to the rapid increase in drug cost since 2014.
However, given other studies and our finding on the strong persistence in drug spending, it is unlikely that
the degree of persistence would be different. Secondly, our data doesn’t include any prescription drug that
are covered under Medicare Part B. Thus we haven’t accounted for any high-cost specialty drugs that may
require administration in an inpatient setting, such as drugs for cancer, rheumatoid arthritis, multiple
sclerosis. Since many of these drugs require long-term or lifetime usage, we may have underestimated the
level of spending per beneficiary, while the persistence pattern is not likely to be different.

41

Conclusions
In this study, we add new evidence of the long-term persistence in prescription drug spending
among the U.S elderly population. We document increasing concentration and strong spending
persistence on prescription drug over a 7-year. With the rapid increases in spending on high-cost specialty
drugs in recent years, it means that beneficiaries who need these drugs would continuously face high and
growing out-of-pocket spending on prescriptions despite having insurance. Thus adding a limit on out-of-
pocket to Medicare’s prescription drug benefit will provide important financial protection for
beneficiaries with high drug spending.
References
Banthin, J.S., and Miller, G.E. (2006). Persistence in Medicare Prescription Drug Expenditures by
Treatment Class. p.
Berk, M.L., and Monheit, A.C. (2001). The Concentration of Health Care Expenditures, Revisited. Health
Aff. (Millwood) 20, 9–18.
Coulson, N.E., and Stuart, B. (1992). Persistence in the use of pharmaceuticals by the elderly: Evidence
from annual claims. J. Health Econ. 11, 315–328.
Garber, A.M., MaCurdy, T.E., and McClellan, M.B. (1997). Persistence of Medicare Expenditures
Among Elderly Beneficiaries (National Bureau of Economic Research).
Hirth, R.A., Calónico, S., Gibson, T.B., Levy, H., Smith, J., and Das, A. (2016). Long-Term Health
Spending Persistence among the Privately Insured in the US. Fisc. Stud. 37, 749–783.
Trish, E., Xu, J., and Joyce, G. (2018). Growing Number O
f Unsubsidized Part D Beneficiaries With Catastrophic Spending Suggests Need For An Out-Of-Pocket
Cap. Health Aff. (Millwood) 37, 1048–1056.
Wrobel, M.V., Doshi, J., Stuart, B.C., and Briesacher, B. (2003). Predictability of Prescription Drug
Expenditures for Medicare Beneficiaries. Health Care Financ. Rev. 25, 37–46.
NIHCM - The Concentration of U.S. Health Care Spending.
A look at people who have persistently high spending on health care - Peterson-Kaiser Health System
Tracker.

42

Tables and figures
Figure 1: Overall Part D spending over time ($2014)

Figure 2: Average spending by percentile groups ($2014)

$-
$20
$40
$60
$80
$100
$120
2006 2007 2008 2009 2010 2011 2012 2013 2014
Overall Part D Spending
Billions
0
5
10
15
20
25
30
35
40
45
2007 2010 2014
Thousands
<50% 50-80% 80-90% 90-95% 95-98% >98%
43

Figure 3: Share of spending by percentile groups

Figure 4(a): Average total spending over time for beneficiaries in low/middle/high spending groups of
2007 (include both survivors and decedents, $2007)

0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2007 2010 2014
<50% 50-80% 80-90% 90-95% 95-98% >98%
$-
$2,000
$4,000
$6,000
$8,000
$10,000
$12,000
$14,000
$16,000
2007 2008 2009 2010 2011 2012 2013 2014
Low Middle High
44

Figure 4(b): Average total spending over time for beneficiaries in low/middle/high spending groups of
2007 (include survivors only, $2007)

Figure 5(a) Average out-of-pocket spending over time for beneficiaries in low/middle/high spending
groups of 2007 (include both survivors and decedents, $2007)

$-
$2,000
$4,000
$6,000
$8,000
$10,000
$12,000
$14,000
$16,000
2007 2008 2009 2010 2011 2012 2013 2014
Low Middle High
$-
$500
$1,000
$1,500
$2,000
$2,500
$3,000
$3,500
$4,000
$4,500
$5,000
2007 2008 2009 2010 2011 2012 2013 2014
Low Middle High
45

Figure 5(b) Average total spending over time for beneficiaries in low/middle/high spending groups of
2007 (include survivors only, $2007)

$-
$500
$1,000
$1,500
$2,000
$2,500
$3,000
$3,500
$4,000
$4,500
$5,000
2007 2008 2009 2010 2011 2012 2013 2014
Low Middle High
46

Table 1: Proportion of beneficiaries moving between spending groups over time (%)
2007 2008 2009 2010 2011 2012 2013 2014

L M H L M H L M H L M H L M H L M H L M H
L 81.6 18.1 0.3 75.4 24.1 0.5 71.2 28.1 0.6 67.6 31.6 0.8 64.2 34.8 1.1 61.2 37.5 1.3 58.2 40.1 1.7
M 14.1 82.6 3.3 18.5 77.3 4.2 20.8 74.3 4.9 22.3 72.2 5.5 23.3 70.6 6.1 24.5 68.9 6.6 24.9 68.1 7.0
H 0.9 20.8 78.3 2.0 30.8 67.3 2.9 35.3 61.9 3.6 37.8 58.6 4.1 42.5 53.5 4.5 45.9 49.6 5.1 48.1 46.8
L-low (0-50
th
); M-middle (51-95
th
); H-high (95-100
th
)

47

Figure 6: Cumulative mortality of 2007 spending groups

Figure 7: Average total spending in prior years and following years for beneficiaries in low/middle/high
spending groups of 2010 (include survivors only, $2010)

0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
2008 2009 2010 2011 2012 2013 2014
Low Middle High
$-
$2
$4
$6
$8
$10
$12
$14
$16
$18
$20
2007 2008 2009 2010 2011 2012 2013 2014
Thousands
Low (0-50%) Middle (51-95%) High (96-100%)
48

Chapter 4: Enrollment Growth in the Part D Enhanced Plans
Background
Medicare Part D is a voluntary outpatient prescription drug benefit for Medicare beneficiaries,
provided through private insurers either as a stand-alone plan, for those enrolled in the traditional
Medicare, or as part of the benefits provided through a Medicare Advantage Plan. Private insurers
compete for beneficiaries based on price (premiums and cost-sharing) and quality (formulary coverage),
but must first submit premium bids to CMS reflecting the estimated cost and profit for delivering the
standard benefit to an enrollee with average health. The standard Part D benefit design includes four
phases: a deductible, initial coverage period (typically 25% coinsurance), coverage gap and catastrophic
phase. Based on the growth rate of per capita spending in Part D, the thresholds of the standard benefit
parameters change annually and have increased every year since 2006, with the exception of 2014.
1

However, most plans no longer follow the standard benefit design. Rather, they offer alternative designs
that are either actuarially equivalent to the standard benefit or a more generous “enhanced’ plan that
carries a higher premium. Actuarially equivalent plans (AE) have the same deductible as the standard
benefit, but different cost-sharing requirements, typically relying on copayments rather than coinsurance.
Basic alternative (BA) plans offer a reduced or zero deductible and different cost-sharing requirements
than the standard benefit, but like AE plans, have the same actuarial value as the standard benefit. In
contrast to the three basic plans, Part D sponsors can also offer “enhanced plans” (EA) that provide more
generous coverage than the standard benefit at a higher cost (premium) to beneficiaries, most notably,
offering some degree of coverage in the donut hole.
While EA plans are more generous than any of the “basic” plans, they vary along a number of
dimensions. Enhanced plans typically charge a higher premium in exchange for additional benefits, such
as a low or zero deductible, reduced cost-sharing, some degree of coverage in the donut hole, and/or an
49

expanded formulary that covers drugs excluded from the standard benefit. In submitting bids to CMS,
enhanced plans must report separate premiums for the basic and supplemental portions of their bid. The
basic portion reflects the value of the basic coverage and is set to be same as the premium for plans with
basic coverage in the same service area, while the supplemental premium is for the “enhanced” coverage
offered beyond the basic design.
For stand-alone PDPs, a sponsor is not permitted to offer a plan that provides enhanced
alternative coverage in a particular service area unless it also offers a plan that provides only basic
prescription drug coverage in that same area. This requirement ensures that PDP sponsors offer at least
one option for Part D coverage for a premium at the cost of basic prescription drug coverage.(CMS
Medicare Prescription Drug Benefit Manual, Chapter 5) By contrast, Medicare Advantage plans (MA-
PD) can offer enhanced alternative coverage without offering basic coverage, as long as at least one plan
with zero Part D supplemental premium (after application of Part A/B rebate dollars) is offered in the
service area. Given the lack of restriction, MA-PD plans tend to be enhanced plans and historically there
are more beneficiaries enrolled in the enhanced MA-PD plans than in the stand-alone PDPs, due to the
lower premium.
In 2007, 32% of Part D beneficiaries enroll in either an enhanced MAPD or stand-alone PDP. By
2018, that fraction has increased to nearly 60%
e
. While much of this growth is attributable to the
expansion of MA-PD and employer plans (EGWPs) – which tend to be enhanced plans – the share of
enrollees in enhanced stand-alone PDPs has also grown over time. Among stand-alone PDPs, 4 out of 10
beneficiaries are now enrolled in an enhanced plan compared to just 2 out of 10 in 2007. Further, our
preliminary analysis finds there is an acute increase of enrollment in the stand-alone PDPs in 2013 from
2012, with nearly 4 million beneficiaries enrolled in the enhanced plans. Interestingly, the jump in the

e
Calculated using Medicare Advantage/Part D Contract and Enrollment Data from CMS.
(https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-
Reports/MCRAdvPartDEnrolData/index.html)
50

enrollment in enhanced plans is accompanied by a decrease of enrollment in basic plans. So far very little
is known about the causes and consequence of this growth in enrollment in enhanced plans as a share of
the overall market and as a share stand-alone PDP enrollment.
The primary objective for this study is to document the growth in enrollment in enhanced plans
and possible reasons why. We examine the likely causes of the growth in enhanced Part D plans by
analyzing enrollment and premium data from the Center of Medicare and Medicaid Services (CMS). We
investigate the types of beneficiaries newly enrolling and switching into enhanced plans,and the range of
plan sponsors driving this growth.
Methods
Data and sample
The primary data source for the study is the publicly available Medicare Advantage/Part D
Contract and Enrollment Data from CMS. These files are available from 2006, up to 2019 and are
updated monthly. In order to capture complete information in a given year, we use data from 2007 to
2018 and for each year, we use the monthly report of December. The enrollment files contain information
on contract and plan id, plan type, organization type, employer group health plan indicator, plan name,
and parent organization. In addition, these files report premium, cost-sharing information, and other plan-
level characteristics. In our analysis, we group all defined standard, basic alternative, and actuarial
equivalent plans as basic plans, as comparison to enhanced alternative plans.
As the CMS public data doesn’t contain any individual-level information, we supplement the
analysis by using data from the 20% random sample of Medicare beneficiaries enrolled in Part D, through
a Data Use Agreement with the National Bureau of Economic Research. The data include administrative
data of Part D claims, plan characteristics files, and Master Beneficiary Summary Files (MBSF). The plan
characteristics files contain the same information as the CMS public files, and the two sources of files can
51

be linked by the unique plan identifier (i.e., contract id and plan id). The MBSFs include beneficiary’s
monthly enrollment information, and detailed sociodemographic (e.g., date of birth, date of death, gender,
race) and socioeconomic information (e.g., zip code, dual-eligible status, whether receiving low-income-
subsidies). There is also a Chronic Conditions segment of MBSF where each beneficiary is flagged for
the presence of specific chronic condition(s). This information allows us to examine health status of the
beneficiaries.
We exclude all employer group health plans and any plans having an unknown region
assignment. Unless noted otherwise, our primary sample is restricted to the stand-alone PDPs given our
primary interest in the enrollment shift in these plans. All analyses are performed using StataCorp
Stata/SE 15.0.
Enrollment
To show the overall enrollment growth in enhanced plans over time, we count the number of
beneficiaries enrolled in those plans, and its share to the entire Part D market. We aggregate the number
of enrollees across all regions under each plan to compute the plan level enrollment in each year. As the
enrollment in enhanced stand-alone PDPs has also grown over time, we describe the trend of enrollment
in enhanced stand-alone PDPs and its share of overall stand-alone PDPs over time. To assess the cause of
an acute enrollment increase in the enhanced stand-alone PDPs, we identify plans that have the largest
enrollment and thus causing the overall enrollment growth in the enhanced stand-alone PDPs in certain
years.
Premium
The premium analysis is restricted to stand-alone PDPs as the reported premiums for MA-PDs
contain both medical and prescription drug coverage and thus don’t reflect the true premiums for the Part
D portion. We first show the overall trend of premiums change in both plans with basic coverage and
enhanced benefits, then examine the trend of premium change over time for the selected plans that had
52

large enrollment change. For each plan, we calculate the average premium weighted by enrollment across
all regions in a given year. Premiums are reported as basic, supplemental and total for the enhanced plans,
and are not adjusted for inflation.
Beneficiary characteristics and switch behavior
Upon identifying plans that switched from basic to enhanced, we examine the characteristics of
the beneficiaries who are involuntarily switched to an enhanced plan due to their plan’s organizational
change. We summarize the basic demographic information (age, gender, race), and percentage of
beneficiaries who are dual eligible and receiving low-income-subsidies (LIS) in the immediate year prior
to the change.
We calculate the beneficiary retention rate of the plans that switched to enhanced over time. The
stayers and switchers are defined as voluntary by adjusting for any lost-to-follow-up and mortality.
Results
Overall trend of enrollment and premium in the enhanced Part D plans
Across the entire non-employer sponsored Part D market, enrollment in enhanced plans (either
MA-PD or stand-alone PDP) has grown substantially over time (Figure 1). Number of beneficiaries
enrolled in the enhanced Part D plans has increased from 6.7 million in 2007 to 21.3 million in 2018, with
the percentage share increasing from 32.6% to 59.2% during the past decade. The growth rate has been
relatively stable over time, except for the year of 2013, where there is an acute 20% increase from the
previous year.
As we mentioned earlier, the growth is largely attributable to the expansion of MA-PDs. In 2018,
there are over 2,500 non-EGWP MA-PDs, up from 1,395 in 2007, and more than 80% of the plans offer
enhanced benefit. However, the share of enrollment in the stand-alone PDPs has also grown over time
53

(Figure 2). In 2018, 40% of the beneficiaries enrolled in stand-alone PDPs are in an enhanced plan, versus
20% in 2007. What’s striking from Figure 2, though, is a sharp increase from 2012 to 2013 with a
doubled enrollment in the enhanced plans. We then further examine what caused this acute increase.
Average premiums have increased modestly over time (Figure 3). In 2007, average monthly
premium for basic stand-alone PDPs are $26.7 and it has gone up to $39.1 in 2018. Average monthly
premium for enhanced plans are $20 higher than the basic plans, on average, with a larger difference
($30) between 2011 and 2015. However, basic premiums of enhanced plans have been consistently higher
than the basic plans and the difference has been diverging since 2011 (except for 2016).
United Health Group
By examining the plans and its parent companies that contribute to the enrollment increase in
2013, we find it is a consequence of an organizational change by the United Health Group, who switched
30 basic plans to enhanced plans in 2013, carrying over millions of beneficiaries. Figure 4 show the
overall enrollment in all basic and enhanced plans of UHG plans over time.
During 2007 to 2018, there are 304 plans of UHG. Some plans (N=36) are available for the entire
period, while others are only available for certain years. All of the 36 plans switched from basic to
enhanced in 2013 or later, with the 2013 change carrying over the most beneficiaries. To further examine
this change, we describe the enrollment and premium change among the switched plans. Figure 5 shows
the trend of enrollment and premium (basic, supplemental and total) of the 30 plans over time and each
line represents one plan. Upon switching to an enhanced plan in 2013, the average increase in the
supplemental premiums is $4.4. However the total premiums of these plans remain unchanged ($40.96 on
average in 2013, and $40.99 in 2012), due to the decreased basic premiums. Starting in 2014, the total
premiums of these plans begin to increase and by 2018, the average total premiums have reached $84,
with a sharp $15 increase in the supplemental premiums from 2017. With this rapid increase in the
premiums, enrollment has only gone down modestly. In Table 1, we compare the characteristics of
54

beneficiaries enrolled in plans that are later switched (in 2013 or in later years) versus not in 2012,
looking at what (if anything) would have encouraged UHG to switch these plans as opposed to others.
Among the plans that are switched, beneficiaries are slightly older, more likely to be female and a
minority (i.e., more black and Hispanic). While there is a higher share of people with dual-eligible and
receiving low-income-subsidies among the plans that are switched in later years, it doesn’t seem
surprising since these plans are basic in 2012 and the subsidized beneficiaries are automatically signed up
for plans with basic coverage.
After observing the premiums increasing over time for the plans became enhanced, we examine if
beneficiaries chose to stay with their plans by using the enrollment information from MBSFs
f
. Results
from Table 2 show the retention rate in each of the following year (after 2012) and cumulatively over
time. We have calculated both unadjusted and adjusted retention rate by controlling for mortality and lost
to follow-up. Overall, an average of 90% of the beneficiaries who survived in our sample stay with the
same plan in each year during 2013 and 2016, and 67.3% of them are still in these plans four years later.
Among beneficiaries who switched to other plans, less than 20% (12.3% to 19.3%) chose a different
UHG plans offering basic coverage, while few chose a different UHG enhanced plan (although there’s
few option for a different enhanced plan outside these 30 plans). The others chose to enroll in a MA-PD
or a non-UHG plan.
Humana
In order to see if the strategy of switching basic plans into enhanced plans and raising premiums
over time was a single case of UHG, or a common play in the insurance market, we examine another
major company, Humana, and carry out the same analyses. We find Humana has done a similar switch
while on a smaller scale. In 2011, 19 basic plans with 285 thousand beneficiaries are switched to

f
The MBSFs are available up to 2016; and it’s a 20% random sample. Thus the numbers in this table reflect the 20%
sample, while the rates are unaffected.
55

enhanced plans and the average total premiums doesn’t increase at the time of switching (average $40 for
both 2010 and 2011). Over time, the average total premiums have reached $78 in 2018, while other
enhanced plans only have an average total premium of $48.
g
Among these plans, retention of the
beneficiaries are still substantial, with 87%-94% of those survived chose to stay with the plan, and over
half (52.9%) remained in the plans 6 years after the switch. Detailed results of enrollment and premium
change, comparisons of beneficiary characteristics among plans switched and non-switched, and
breakdown of the churn in the follow-up years can be found in Table 3,4, and Figure 6,7.
Discussion
Enrollment in enhanced Part D plans has expanded over time, with a substantial growth from
enrollment in the enhanced stand-alone PDPs. We find that instead of beneficiaries voluntarily choose to
enroll thus leading to such growth, the increase mainly comes from organizational changes of certain
major insurers. In 2013, United Health Group switched 30 stand-alone PDPs, carrying a total of 3.5
million enrollees, from basic plan to enhanced plan. Humana has done a same switch for 19 of its plans in
2011, with less beneficiaries. We observe a similar strategy in premium setting for these plans became
enhanced under both UHG and Humana. Upon switching to enhanced plan and thus an increase in the
supplemental premium, the plans on average had a small reduction in the basic premiums, which together
kept the total premium unchanged. However in later years, the total premiums of these plans began to
increase and almost doubled by 2018. What differs between such strategic movement of UHG and
Humana is the number of beneficiaries enrolled in those plans. 3.5 million beneficiaries that were enrolled

g
35 enhanced plans that are available since 2014 with 2.4 million beneficiaries have an average total premium of
$21 in 2018. Another 15 enhanced plans that are available from 2007 (switched to basic in 2010 only; it’s unclear if
it’s a coding error in the data) with 403 thousand beneficiaries have an average total premium of $76.6 in 2018.

56

in the UHG basic plans have been involuntarily enrolled in the enhanced plans, which directly caused the
acute enrollment increase in the enhanced Part D market in 2013.
Given the substantial premium increase, one might expect the beneficiaries to re-enroll to a
different plan with lower premium by taking advantage of the annual Medicare open enrollment period.
However in our sample, only 10% of the beneficiaries in these plans chose to switch in every year and
over half stay with their plans 6 years after the plan switch. This is consistent with the findings from the
literature on consumer inertia in the Medicare Part D program. Hoadley et al. (2013) has documented a
13% switching rate during the annual enrollment period of Medicare Part D using data from 2006 through
2010.
2
They find that even though large premium increases are associated with higher switching rates,
most enrollees with relatively large premium increases, such as $10 or more per month, do not switch
plans. In addition, they find that enrollees in PDPs that increased deductibles or dropped coverage of
brand-name drugs in the coverage gap are more likely to switch out of these plans than enrollees in PDPs
that keep their benefit design. This aligns well with our findings that the majority of beneficiaries choose
to stay with their enhanced plans, which generally have lower or zero deductible, and offer additional
coverage in the coverage gap, despite the substantial premium increase.
While our work hasn’t assessed the spending between beneficiaries who switched and stayed,
existing work has suggested that consumers are more likely to pay attention to plan choice if
overspending
h
in prior year is more salient and if their old plan gets worse, for instance due to premium
increases.
3
Beneficiaries who actively switch Part D plans had lower overspending than stayers, and those
forced to switch plans due to plan termination. Insurers, on the other hand, profit from consumer inertia.
Ho et al (2015) estimate a model of consumer plan choice with inattentive consumers and show that high
observed premiums are consistent with insurers profiting from consumer inertia.
4
They estimate an

h
Overspending in the Heiss et al (2010) study is defined as the difference between the consumer’s total costs
(include premiums, deductibles, and copayments) induced by the plan the consumer has chosen and that of the least-
cost alternative (for given prescription drug use)
57

average of $1,050 saving over three years per person if all consumers were attentive; and $1.3 billion
savings for the government.
Our finding of higher basic premiums of enhanced plans than premiums of basic plans is also
interesting. As enhanced plans charge higher premium through the supplemental premium, which reflects
the additional or enhanced benefit package, the basic premium still reflects the standard benefit or its
actuarially equivalent benefit package. Although such difference is not warranted, the current structure of
Part D’s risk corridors may provide plan sponsors financial incentives to submit higher bid.
Determination of associated federal subsidies for Part D plans is only based on the basic premium
component of plan bids and Medicare subsidizes the plan sponsors 74.5% of the expected cost of basic
benefits. Therefore if plan sponsors bid higher (for its basic premium) than it would actually have cost on
the benefit spending (other than catastrophic benefits), plan sponsors can keep most of the difference as
profits.
5
Thus our findings provide important insight regarding a currently under-appreciated aspect of the
Part D market that has important implications for federal spending and the ongoing integrity of the Part D
competitive bidding structure.
Future work
It is a general understanding that enhanced plans charge a high premium than the basic plans by
offering more generous cost-sharing benefits, such as lower or zero deductible, higher coinsurance for
certain services, additional coverage for the donut hole, etc. However, there is no systematic evidence
regarding the attributes on which enhanced plans differ from basic plans, but understanding these features
is important for assessing their potential value to beneficiaries. Additionally, understanding how enhanced
benefits have evolved over time is key to understanding the dynamics at play, particularly since the basic
Part D benefit design has become considerably more generous as the doughnut hole has been filled in
since 2011.
58

References
Oct 12 P, 2018. An Overview of the Medicare Part D Prescription Drug Benefit. Henry J Kais Fam
Found. October 2018. https://www.kff.org/medicare/fact-sheet/an-overview-of-the-medicare-part-d-
prescription-drug-benefit/. Accessed October 16, 2019.
Hoadley J, Hargrave E, Summer L, 2013. To Switch or Not to Switch: Are Medicare Beneficiaries
Switching Drug Plans To Save Money? Henry J Kais Fam Found. October 2013.
https://www.kff.org/medicare/issue-brief/to-switch-or-not-to-switch-are-medicare-beneficiaries-
switching-drug-plans-to-save-money/. Accessed September 18, 2019.
Heiss F, McFadden D, Winter J, Wuppermann A, Zhou B. Inattention and Switching Costs as Sources of
Inertia in Medicare Part D. Cambridge, MA: National Bureau of Economic Research; 2016.
doi:10.3386/w22765
Ho K, Hogan J, Morton FS. The Impact of Consumer Inattention on Insurer Pricing in the Medicare Part
D Program. National Bureau of Economic Research; 2015. doi:10.3386/w21028
MedPAC. The Medicare Prescription Drug Program (Part D): Status Report.
http://www.medpac.gov/docs/default-source/reports/mar18_medpac_ch14_sec.pdf. Accessed October 19,
2019.

59

Tables and figures
Figure 1: Enrollment in enhanced plans, by year

Figure 2: Enrollment in enhanced stand-alone PDPs, by year

0%
10%
20%
30%
40%
50%
60%
70%
0
5
10
15
20
25
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Millions
N %
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
0
1
2
3
4
5
6
7
8
9
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Millions
N %
60

Figure 3: Enrollment-Weighted Premiums of Basic and Enhanced PDPs, 2007- 2018

26.7
28.6
33.4
36.1
36.5
34.3
36
34.5
33.7
42.9
36.4
39.1
29.1
33.4
39.3
41.8
49
45.5
49.5
47.7 47.8
45.2
48.9
47.8
45.2
49.1
55.5
56.1
77.4
74.8
72.6 72.4
67.9
63.9
66.4
64.3
0
10
20
30
40
50
60
70
80
90
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Average PDP Beneficiary Monthly Premium ($)
Enhanced PDPs
(Total)
Basic PDPs
Enhanced PDPs
(Basic)
61

Figure 4: Enrollment in basic and enhanced stand-alone PDPs over time (UHG)

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Basic 4562469 3855229 3861187 4330419 4562878 4007657 1001057 1079019 1337757 1559689 1505343 1416302
Enhanced 92225 114057 333914 102692 102782 107738 3515517 3843709 3566076 3226892 3233226 3136737
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Millions
Basic Enhanced
62

Figure 5: Basic and supplemental premiums of UHG plans that switched from basic to enhanced, over
time

63

Table 1: Summary of beneficiary characteristics in UHG plans that switched from basic to enhanced in
2013, always enhanced, and switched from basic to enhanced beyond 2013

Plans switched
in 2013
Non-Switched
(Enhanced in 2012
and beyond)
Non-Switched
(Basic in 2012-2013
and enhanced in later
years)
Age 73 71.9 71.9
Female 61.4% 57.9% 60.4%
LIS 4.5% 1.9% 3.1%
Dual 15.9% 6.4% 29.9%
White 85.8% 89.5% 84.1%
Black 7.1% 4.8% 3.7%
Hispanic 3.5% 2.6% 5.8%
N of plans 30 32 6
Enrollment 3,500,000 113,000 494,000

64

Table 2: Percentage of beneficiaries of UHG plans that stayed or switched after their basic plan changed
to an enhanced plan in 2013

2012 2013 2014 2015 2016
Stayed 704,009 613,275 536,462 461,344 394,056
Died in t 29,663 26,608 23,154 21,152 18,782
Lost to FU in t+1 1,012 15,410 1,284 293

Retention in t+1

91.1% 93.9% 90.1% 89.6%
Stayed%(unadjusted)

87.1% 76.2% 65.5% 56.0%
Stayed%(adjusted)

91.1% 85.0% 76.0% 67.3%
Mortality 4.2% 4.3% 4.3% 4.6% 4.8%
Cumulative mortality 4.2% 8.0% 11.3% 14.3% 17.0%
Went to different plan

60,059 34,795 50,680 45,843
Died

11.2% 4.0% 14.3% 13.8%
Went to MAPD

29.9% 32.9% 22.0% 17.9%
Other UHG enhance pdp

0.8% 2.5% 0.9% 0.9%
UHG basic pdp

13.7% 19.3% 12.3% 12.5%
other enhance pdp

24.1% 19.6% 21.7% 12.3%
other basic pdp

15.7% 25.7% 23.2% 38.6%
Other

15.7% 0.0% 19.9% 17.9%

65

Figure 6: Enrollment in Humana’s basic and enhanced PDPs, by year

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Basic 2132773 1579956 836541 1053967 858230 1397819 1738782 1736351 1690685 1788904 1871933 1694172
Enhanced 1324453 1545934 1241227 667997 1469924 1436132 1358411 1960332 2543004 2882255 3162388 3183520
0
0.5
1
1.5
2
2.5
3
3.5
Millions
Basic Enhanced
66

Figure 7: Premiums and enrollment in Humana plans that switched from basic to enhanced

67

Table 3: Characteristics of beneficiaries that remained in Humana basic plans or switched to enhanced in
2011

Switched Non-Switched
Age 74.5 73.8
Female 65.2% 61.7%
LIS 8.7% 4.3%
Dual 18.4% 9.9%
White 89.5% 88.5%
Black 6% 4.9%
Hispanic 2.1% 3.8%
Enrollment 285,000 1,400,000

68

Table 4: Percentage of beneficiaries of Humana plans that stayed or switched after their basic plan
changed to an enhanced plan in 2011

2010 2011 2012 2013 2014 2015 2016
Stayed 57,209 46,708 40,553 35,635 31,019 26,761 22,556
Died (among the stayed) 2,559 2,285 2,035 1,926 1,637 1,536 1,324
Lost to FU in t+1 1,133 735 19 673 18 16

Retention in t+1

87% 93% 93% 94% 91% 89%
Stayed%(adjusted)

87.3% 80.3% 73.6% 67.7% 60.6% 52.9%
Stayed% (unadjusted)

81.6% 70.9% 62.3% 54.2% 46.8% 39.4%
Cumulative mortality 4.5% 8.5% 12.0% 15.4% 18.3% 20.9% 23.3%
Went to different plan

6,809 3,135 2,864 2,017 2,596 2,653
Mortality

3.9% 5.3% 15.7% 5.8% 19.2% 17.4%
Went to MAPD

24.8% 33.1% 26.3% 26.2% 21.8% 17.0%
Other Humana basic pdp 31.0% 15.9% 8.1% 11.1% 7.4% 10.7%
Other Humana enhanced pdp

1.4% 2.8% 2.6% 22.8% 15.9% 12.7%
Other enhanced pdp

3.3% 12.8% 21.5% 6.5% 11.7% 6.4%
Other basic pdp

38.6% 33.9% 22.5% 33.4% 23.0% 35.5%
Other

- - - - 20.1% 17.5%

69

Chapter 5: Summary and Future Research Directions
In this dissertation, we investigate three topics that are highly relevant yet less researched under
the current policy discussion of reforming Medicare. The rapidly increasing healthcare spending,
particularly on prescription drug, is a pressing issue and deserves comprehensive and in-depth analysis.
We utilize empirical data from Part D beneficiaries and plan sponsors to evaluate the variation in
medication use, long-term spending persistence of prescription drugs and the recent trend of increasing
enrollment in plans offering enhanced benefits.
We find considerable variation in use of generics among plans despite the increasing overall
generic dispensing rate in Part D. For each of the commonly used therapeutic class we select in our
analysis, we observe 8% - 40% difference in using generics. We also find that beneficiary and plan
characteristics, state and parent organization all are factors contributing to differences in the mix of brand
and generic drugs within a therapeutic class. Variation in generic use does have real financial
consequences for plans, beneficiaries and taxpayers. For example, we estimate about $500 to $800
million saving for United Health Group if they would have increased the generic use to the average in our
sample. In addition, by examining tier placement for selected brand name drug, we find suggestive
evidence that manufacturer rebates may be influencing plans’ formulary design. Focusing on changing the
underlying benefit design to reduce incentives to use high-cost drugs or encouraging the use of more
tightly managed formularies that restrict beneficiaries’ access to newer expensive brand medications
within a therapeutic class could increase generic use. Future work can assess if plans with more generous
coverage for brand name or specialty drugs are associated with better health outcomes of its enrollees.
In our second paper, we observe substantial degree of persistence in prescription drug spending
among Part D enrollees. More than half of the beneficiaries with highest spending remain the spending
level after 7 years. The persistently high spending, combined with the trend of increasing cost on
prescription drugs, put beneficiaries at risk for facing continuously high out-of-pocket spending. Findings
from this study warrant the urgent need for a cap on out-of-pocket spending in Medicare.
70

Lastly, we examine the recent enrollment growth in the Part D plans with enhanced benefits. We
find considerable growth in the enhanced plans despite the more rapidly increased premiums as opposed
to plans with basic benefits. Findings reveal the fact that big plan sponsors are strategically shifting plans
with basic benefits to enhanced plans, by holding on premiums at the change and increasing premiums
over time. Meanwhile, there is substantial inertia among beneficiaries even with the large premium
increase. Future work can examine whether plans with large premium increase still remain attractive in
the competitive market of Part D. In addition, understanding how enhanced benefits overall have evolved
over time is key to understanding the dynamics at play, particularly since the basic Part D benefit design
has become considerably more generous as the doughnut hole has been filled in since 2011.

Abstract (if available)

Abstract The rapidly increasing healthcare spending, particularly on prescription drug, is a pressing issue in the United States and deserves comprehensive and in-depth analysis. In this dissertation, I investigate three topics that are highly relevant yet less researched under the current policy discussion of reforming Medicare by using empirical data from Part D beneficiaries and plan sponsors. The first paper examines the variation in medication use and finds considerable variation in use of generics among plans despite the increasing overall generic dispensing rate in Part D, as well as suggestive evidence of rebates influencing plan’s formulary design. Focusing on changing the underlying benefit design to reduce incentives to use high-cost drugs or encouraging the use of more tightly managed formularies could increase generic use and thus generate cost-savings for both the program and beneficiaries. Second paper describes the long-term spending persistence of prescription drugs and I observe substantial degree of spending persistence among Part D enrollees. Findings from this study warrant the urgent need for a cap on out-of-pocket spending in Medicare. The last paper focuses on the recent trend of increasing enrollment in plans offering enhanced benefits. Findings reveal the fact that big plan sponsors strategically increase premiums over time by shifting plans with basic benefits to enhanced plans, while choice behaviors of beneficiaries show substantial inertia.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Essays on access to prescription drugs

PDF

Economic, clinical, and behavioral outcomes from medical and pharmaceutical treatments

PDF

Essays in opioid use and abuse

PDF

Health care utilization and spending of the U.S. aging population

PDF

Value in health in the era of vertical integration

PDF

Value in oncology care and opportunities for improvement

PDF

Essays on health insurance programs and policies

PDF

Understanding primary nonadherence to medications and its associated healthcare outcomes: a retrospective analysis of electronic medical records in an integrated healthcare setting

PDF

Three essays on comparative economic and clinical outcomes of antipsychotics using retrospective claims data

PDF

Impact of pharmacy-based transitional care on healthcare utilization and costs

PDF

The impact of Patient-Centered Medical Home on a managed Medicaid plan

PDF

Delivering better care for children with special health care needs: analyses of patient-centered medical home and types of insurance

PDF

US biosimilar market entry and uptake: US commercial payer preferences in access decision making

PDF

Three essays on estimating the effects of government programs and policies on health care among disadvantaged population

PDF

The value of novel antihyperlipidemic treatments in the U.S. healthcare system: Reducing the burden of cardiovascular diseases and filling the gap of low adherence in statins

PDF

Pharmacotherapy among patients with type 2 diabetes

PDF

Essays on the economics of infectious diseases

PDF

The impact of treatment decisions and adherence on outcomes in small hereditary disease populations

PDF

Behavioral approaches to industrial organization

PDF

Essays in health economics and provider behavior

Asset Metadata

Creator Xu, Yifan (author)

Core Title Thesis on Medicare Part D

School School of Pharmacy

Degree Doctor of Philosophy

Degree Program Pharmaceutical Economics and Policy

Publication Date 12/11/2019

Defense Date 10/08/2019

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag enhanced benefits,generic utilization,Medicare,OAI-PMH Harvest,Part D,prescription drug spending,spending persistence

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Joyce, Geoffrey (committee chair), Romley, John (committee member), Trish, Erin (committee member)

Creator Email yifanxu@usc.edu,yvonnexu198777@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-251423

Unique identifier UC11673877

Identifier etd-XuYifan-8044.pdf (filename),usctheses-c89-251423 (legacy record id)

Legacy Identifier etd-XuYifan-8044.pdf

Dmrecord 251423

Document Type Dissertation

Rights Xu, Yifan

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA