Introduction
Medicare is the federal entitlement program which was signed into law on July
30l , 1965 by President Lyndon B. Johnson. The original Medicare benefit provided
health care services to beneficiaries in the form of an inpatient/hospital (Part A) and
outpatient medical (Part B) benefit. The Medicare Prescription Drug, Improvement, and
Modernization Act (MMA) was signed into law by President George W. Bush on
December 8, 2003 and went into effect on January 1, 2006. This act was the most
substantive legislation affecting the Medicare benefit since its inception. The MMA
created the provision which offers Medicare beneficiaries optional prescription drug
coverage. Beneficiaries who choose to receive their prescription drug benefit (Part D)
through a Medicare-approved plan have one of two choices: a) enrolling in a stand-alone
prescription drug plan (PDP) or b) enrolling in a Medicare Advantage prescription drug
plan (MA-PD) in which the inpatient/hospital, outpatient medical, and prescription drug
benefit are offered through a private insurance company (Part C).
As of January 2008, an estimated 25.4 million beneficiaries were enrolled in
either a PDP or a MA-PD. The ratio of beneficiaries who signed up for a PDP versus a
MA-PD is approximately 2:1.' An impetus behind the Medicare Part D (MPD) benefit
was to provide patients a broad spectrum of drug plans from which to choose, thereby
increasing competition among plan sponsors, while allowing beneficiaries flexibility in
coverage thereby enabling them to select a plan best suited to meet their specific
individual needs.2 As seen in Table 2.1, the MMA fulfilled this objective with the
provision of between 45 and 66 PDPs offered in each of the 34 MPD regions in 2007.3 In
2008, between 47 and 63 PDPs were offered.4

The large number of PDPs notwithstanding, the Part D benefit is complicated by
the fact that each plan can differ with respect to plan benefit design parameters. At
minimum, PDP sponsors must offer a plan that is "actuarially equivalent" to the standard
benefit plan as defined by the Centers for Medicare and Medicaid Services (CMS). The
standard MPD prescription drug benefit (Table 2.2) includes four coverage levels
consisting of the deductible, initial coverage, gap coverage (also referred to as the "donut
hole"), and catastrophic coverage.

The patient pays 5% of total drug cost with the annual minimum co-pay for
generic/brand name medications set at $2.15/$5.35 and $2.25/$5.60 in 2007 and 2008,
respectively Catastrophic coverage is subsidized by the Centers for Medicare and Medicaid Services
at 80% of drug costs increased maximum deductible, increased ranges of initial and gap coverage, and a higher
spending limit required to reach catastrophic coverage. According to the Kaiser Family
Foundation, only 12% of PDPs nationwide offered the standard benefit in 2008 \ Most
plans (~60%>) did not have a deductible and 86% charged tiered co-payments as opposed
to a flat co-insurance for covered drugs in 2008.'

With the additional media focus on changes to certain plan parameters as key
indicators of beneficiary costs, it is not surprising that some beneficiaries may use only
certain plan parameters (e.g., premium and/or deductible) when choosing their PDP.5'6'7
While each individual plan parameter comprises a portion of a beneficiary's total yearly
out-of-pocket costs, no single parameter can accurately predict a beneficiary's total
annual out-of-pocket costs for a given PDP. An estimate of the annual cost for each PDP
is based on a beneficiary's medication profile (e.g., medication names, doses, and
administration frequencies) and other individual needs (e.g., preferred pharmacy) and can
be obtained using the Plan Finder Tool on the Medicare website (www.medicare.gov).
This tool is provided by CMS and is only available online via the official Medicare
website. By using the Plan Finder Tool, beneficiaries are able to compare each plan's
estimated annual cost (EAC) for a given year based on the medication profile and other
beneficiary-specific parameters (e.g. zip code). The EAC includes beneficiary co-
payments/co-insurance, deductible, plan premium, and if applicable, the amount paid in
the coverage gap and catastrophic coverage. These plan parameters are regularly updated
by CMS throughout the year as they receive updates from PDP sponsors. In summary,
the EAC is the sum of all out-of-pocket costs that are expected to be incurred by a
beneficiary during a given year.

Considering that EAC represents all inclusive out-of-pocket patient costs, few
studies have relied on it to determine beneficiary drug costs. Previous research that has
attempted to approximate EAC has either relied on a small (four) group of hypothetical
patients, rather than actual Medicare recipients, or on chain pharmacy claims data,
which cannot account for a beneficiary's monthly prescription drug plan premium and
therefore underestimates total out-of-pocket costs. This is the first study which examines
the national between-year variation in EAC of PDPs from 2007 and 2008 using a cohort
of Medicare-eligible patients.

Methods
A database of approximately 18,000 Medicare-eligible patients was provided by
the data analytics company, Medlnitiatives M. This database consisted of pharmacy
claims data provided by the pharmacy benefits management company Catalyst Rx®. All
protected health information and unique patient identifiers were stripped from the
database prior to provision. A sample of 50 patients was randomly selected from the
provided database. Pharmacy claims records for each study patient were obtained from
the provided database during the six-month period from January 1 to June 30, 2007 in
order to determine the patient's medication profile. Prescription medications taken on a
regular basis, defined as having at least an 84-day supply filled during the six month
interval, were included in the patient's medication profile. Patients without any
medications meeting this criterion over the six month time frame of interest were
excluded from further analyses and replaced by a new randomly selected patient.
A patient medication profile (e.g., medication name, strength, dosage form,
frequency of use) was created and saved on the Medicare website (www.medicare.gov)
for each of the 50 sample patients. The Medicare Plan Finder Tool was used to
determine the EAC of each PDP offered in all 34 MPD regions for all 50 patient profiles.
The EAC for all offered plans in both 2007 and 2008 were obtained during December
2007 to reflect the plan costs reported to patients during the annual open enrollment
period (November 15th through December 31st). All Plan Finder results were web-captured as Adobe Acrobat files. For each patient in each MPD region, the lowest (least
expensive), 25th percentile, median, and highest (most expensive) EAC PDP were
identified for both 2007 and 2008. In region 18, only 48 patients were included in the
25th percentile and median EAC analysis, while in regions one, 6, 8, 22, and 23 only 49
patients were included due to file corruption during web-capture.
Estimated annual cost values from 2007 were adjusted using the medical
commodities Consumer Price Index (CPI) to account for the average change in
prescription drug prices prior to comparison with 2008 values. Consumer Price Index
was calculated at the regional level (e.g., West, Midwest, South, and Northeast) from
November 2006 to November 2007 in order to reflect drug price changes occurring
between the 2007 and 2008 open enrollment periods.10 The CPIs for MPD regions with
states in two different CPI regions (MPD regions 6, 15, and 25) were calculated using
population (over 65 years old) weighted CPI values from each state in the MPD region.11
The normality of cost data was tested via the Kolmogorov-Smirnov test. Cost
data were not consistently normally distributed. As such, pair-wise comparisons between
the CPI-adjusted 2007 EAC and 2008 EAC were made for the sample cohort in each of
the 34 MPD regions using the Wilcoxon Signed-Ranks test at an a priori alpha level of
0.05. Additionally, in each MPD region, the mean (standard error), median, and range of
the least expensive, 25th percentile, median and most expensive EAC plan were
calculated for each member of the cohort for both 2007 (CPI-adjusted) and 2008. All
statistics were performed using SPSS, version 15.0 (Chicago, 111.).

Results
Table 2.3 displays the demographic characteristics of the 50-patient sample.
While most patients in the sample were between the ages of 65 and 84 (80%), 12% were
under the age of 65 and thus able to qualify for Medicare by some reason other than age
(e.g., end-stage renal disease or permanent disability). The study cohort was comprised of
roughly equal numbers of males and females. The mean (SD) number of prescriptions per
sample patient was 5.4 (3.0). Patients in the cohort had between one and 14 prescriptions
filled each month. The proportion of generic:brand name prescriptions filled each month
in the study sample was 60%:40%.
Descriptive statistics for the least expensive (Table 2.4), 25l percentile (Table
2.5), median (Table 2.6) and highest (Table 2.7) EAC plan for 2007 (CPI-adjusted) and
2008 for the study cohort in all 34 Medicare Part D regions are reported.
The lowest EAC PDPs in 2008 were significantly higher than the CPI-adjusted
lowest EAC PDPs of 2007 in 21 of the 34 regions (Table 2.4). The medians of the lowest
EAC plans were smaller than the means of the lowest EAC plans in both years and for all
regions. The range of lowest EAC plan costs widened from 2007 (CPI adjusted) to 2008.
The 25th percentile and median EAC revealed statistically significant increases
between 2007 (CPI-adjusted) and 2008 in all but six and two Medicare Part D regions,
respectively (Table 2.5 and Table 2.6, respectively). In comparison, the highest EAC
comparisons between years revealed a statistically significant decrease in 20 regions, an
increase in 12 regions, and no significant change in the remaining two regions (Table
2.7). In both years, the mean of the highest cost plan was 200-350% greater than the
EAC of the lowest cost plan, and approximately 150-350% greater than the EAC of the
25th percentile plan, for the cohort in each Medicare Part D region.

Discussion
This is the first study to report estimates of the total annual out-of-pocket MPD
beneficiary costs on a national level using a sample cohort of Medicare-eligible patients
and to determine the annual variability in plan costs between 2007 and 2008.
The between-year variation in the lowest EAC plans (Table 2.4) indicates that
patients in 13 of the 34 MPD regions may not have paid significantly more for their PDP
Table 2.4 Descriptive statistics of the least expensive estimated annual cost stand-alone
prescription drug plan (PDP) in each region for 2007 (CPI-adjusted) and 2008 for the
sample cohort PDP 2007 2008 2007

Based on Wilcoxon Signed-Ranks test between 2007 (CPI-adjusted) and 2008 EAC
in 2008 as compared to the CPI-adjusted 2007 values. However, this assumes that each
patient was enrolled in the least expensive PDP during each year, requiring the patient to
reassess the PDP offerings during the annual open enrollment period. From surveys of
beneficiaries enrolled in MPD PDPs conducted by the Kaiser Family Foundation, it is
clear that many MPD enrollees did not intend to re-evaluate the PDP offerings for 2008.12
Furthermore, considering that the number of regions with no statistically significant
increase in EAC between 2007 and 2008 costs are reduced to six regions for the 25l
percentile (Table 2.5) and two regions for the median (Table 2.6) EAC plans,
beneficiaries may be paying significantly more if they are not enrolling in the least
expensive plan. Additionally, median EACs increased by approximately $300-$400 from
the lowest to the 25th percentile plan and from the 25th percentile to the median plan.
The significant decrease in the highest EAC plans seen in 20 regions between
2007 (CPI-adjusted) and 2008 (Table 2.7) likely indicates improvements in plan
formularies. This can be inferred from the $7,000 to $15,000 decrease in the upper end
of the range in 17 regions, as well as the approximate $2,000 decrease in mean and
median EAC in 15 regions (Table 2.7). The logical implication of this magnitude of cost
decrease would be through the inclusion of non-formulary medications in 2007 onto plan
formularies in 2008.

Together these findings indicate the necessity for beneficiaries to re-evaluate the
PDP offerings annually in order to select the lowest cost plan each year. Beneficiaries
without computer or online access to the Medicare website, those who do not want to call
Medicare to discuss their MPD drug plan options over the phone, or beneficiaries
requiring assistance with their medication lists should seek the help of trained
pharmacists or other healthcare providers proficient with MPD plan selection. Note that
while pharmacists are unable to direct a patient to a particular MPD plan, they, along
with other healthcare professionals, can assist a beneficiary with the Plan Finder Tool and
answer their questions regarding plan enrollment.

The sample size used in this study may limit the generalization of these results to
the Medicare population as a whole; however additional sampling was precluded by the
amount of data that could be collected during the MPD open enrollment period and
before removal of 2007 plan cost data from the Medicare website. Notwithstanding, the
characteristics of the study cohort (Table 3) are similar to those previously reported from
a 2006 national survey of the MPD senior population.13 Patients in our sample cohort
filled an average of 5.4 prescriptions per month with 30% obtaining seven or more
prescriptions monthly, compared to 5.0 prescriptions per month and 27.6% of seniors
enrolled in MPD plans obtaining seven or more prescriptions per month in the
aforementioned national survey.13 Our sample cohort had fewer patients over 85 years of
age (8%), compared to 11.6%, and fewer female patients (52%), compared to 63.5% in
the 2006 national survey.13 These differences are reasonable considering that Neuman et
al. excluded patients under 65 years of age while our cohort consisted of Medicare-
eligible patients of any age. As such, the comparison and similarity to the characteristics
presented in Neuman et al. appear to validate this cohort as a representative sample of the
general MPD population.

This study is intended to be a preliminary investigation of the trends in out-of-
pocket beneficiary drug costs through the MPD benefit as, to date, there have been no
other published studies reporting such data. Future studies observing the MPD costs of a
larger sample cohort over several years are justified to confirm the significant changes
observed in beneficiary drug costs through the MPD drug benefit. Considering many
seniors in the US are on a fixed income with medical expenses representing a large
fraction of their monthly incomes, even a small increase in drug costs could greatly
impact these individuals.

This study represents a novel approach to estimating MPD costs from the
beneficiary's perspective using medication profiles from Medicare-eligible patients and
calculating each patient's EAC using the Plan Finder Tool. In general, it appears that
MPD PDP costs are predominantly increasing across the U.S. from 2007 to 2008 based
on the increases of the lowest, 25l percentile, and median EAC plans. This is contrasted
with large MPD PDP cost decreases in highest EAC plans, potentially indicating wider
formulary coverage in many regions. Although using the Plan Finder Tool is a time-
intensive method to obtaining beneficiary-specific data, it is currently the only available
approach for beneficiaries or their advocates to compare the EAC of the various available
PDP options. The wealth of PDP data provided through the Plan Finder Tool could be
extracted to further educate MPD beneficiaries, healthcare providers, and the media about
the relevance of various plan parameters (e.g. monthly premiums), as compared to total
out-of-pocket beneficiary costs. Considering the annual flux of PDP offerings and plan
parameters, future studies utilizing Plan Finder Tool data could also be used to determine
the opportunity costs involved in failing to reassess PDP offerings during the annual open
enrollment period. In all cases, these studies would further educate Medicare
beneficiaries so that they can maximize their MPD benefit, while decreasing their out-of-
pocket drug expenditures.

References
1. Kaiser Family Foundation. The Medicare Prescription Drug Benefit. Publication
7044-08. Washington, DC: Kaiser Family Foundation; 2008.
2. President George W. Bush. Remarks following a meeting on Medicare
Prescription Drug Benefit. Weekly Comp Pres Docs. 2007;43:511.
3. Kaiser Family Foundation. Medicare Part D plan characteristics, 2007.
Publication 7426-02. Washington, DC: Kaiser Family Foundation; 2006.
4. Kaiser Family Foundation. Medicare Part D plan characteristics, by state, 2008.
Publication 7426-04. Washington, DC: Kaiser Family Foundation; 2007.
5. Kline RR, Gupta K. Drug benefit decisions among older adults: a policy-
capturing analysis. Med Decis Making. 2006;26:273-81.
6. Medicare Part D premiums slated to rise for 2008. Pharmacy Times.
2007;73(12):1.
7. Saul S. There are alternatives: insuring to bridge the gap or opting out. The New
York Times. 2007 Nov 24.
8. Davis MM, Patel MS, Halasyamani LK. Variation in estimated Medicare
prescription drug plan costs and affordability for beneficiaries living in different
states. J Gen Intern Med. 2007;22(2):257-63.
9. Yin W, Basu A, Zhang JX, Rabbani A et al. The effect of the Medicare Part D
prescription benefit on drug utilization and expenditures. Ann Intern Med.
2008;148:169-77.
10. Bureau of Labor Statistics, US Department of Labor. CPI Detailed Report Tables
for November 2006 thru November 2007. http://www.bls.gov/cpi/, accessed on
July 19, 2008.
11. US Census Bureau, Population Division. Intercensal Estimates of State
Population, http://www.census.gov/popest/archives/index.html, accessed January
20, 2008.
12. Kaiser Family Foundation. Voices of beneficiaries: attitudes toward Medicare
Part D open enrollment for 2008. Publication 7722. Washington, DC: Kaiser
Family Foundation; 2007.
13. Neuman P, Kitchman Strollo M, Guterman S et al. Medicare prescription drug
benefit progress report: findings from a 2006 national survey of seniors. Health
Affairs. 2007;26:w630-w643.