Medicare's drug discount card program: beneficiaries' experience with choice.

This article describes Medicare beneficiaries' experience with the choice among Medicare drug discount cards and is based primarily on surveys and focus groups with beneficiaries as well as interviews with other stakeholders. Although competition and choice have the potential to reduce cost and enhance quality in the Medicare Program, our findings highlight some of the challenges involved in making choice work in practice. Despite the unique and temporary nature of the drug discount card program, these findings have considerable relevance to the Part D drug benefit and to other Medicare initiatives that rely on choice.

Introduction

Over the past decade, the Medicare Program increasingly relied on private market strategies to both reduce cost and enhance quality of care. The underlying rationale is that inviting private companies to offer Medicare products and offering beneficiaries choices among these products will drive the Medicare Program toward greater efficiency and value. In order for these market strategies to deliver the expected benefits, the products must differ in ways that are meaningful to beneficiaries, and they must be willing and able to understand these differences and make appropriate choices.

Several programs initiated by the 2003 Medicare Prescription Drug, Improvement, and Modernization Act exemplify the interest in market strategies. These included the Medicare Drug Discount Card and Transitional Assistance programs, active between spring 2004 and the end of 2006, and the Medicare Part D drug benefit, launched in January 2006. In the former program, Medicare beneficiaries were able to enroll in privately provided Medicare-approved drug discount cards that offered discounts on drug prices. Card sponsors were able to vary the benefit design. Beneficiaries could choose among 39 national and 33 regional cards at the start of the program (Kaiser Family Foundation, 2004), but could enroll in only one card at a time. In addition, those Medicare beneficiaries whose income did not exceed 135 percent of the Federal poverty level and who did not have access to other drug coverage were also eligible for transitional assistance of up to $600 annually, to be applied directly to the cost of prescription drugs. The benefit was administered via the Medicare drug discount cards, greatly enhancing their value for eligible beneficiaries. More than 6.3 million Medicare beneficiaries enrolled in drug discount cards, with more than 1.8 million of these also qualifying for transitional assistance.

This article describes beneficiaries' experiences with the choice among Medicare drug discount cards, including beneficiaries' (1) awareness and understanding of the drug discount card program, (2) sources of information and information seeking behavior, and (3) actual decision-making. The study draws on surveys and focus groups with beneficiaries as well as interviews with other stakeholders, all conducted as part of two CMS-sponsored evaluations of the drug discount card program (Abt Associates Inc., 2005; 2006a; 2006b).

Background

Research by Gold, Achman, and Brown (2003) has shown that choosing health plans is challenging for Medicare beneficiaries, who rarely embark on the decisionmaking process unless they have to (e.g., when they turn age 65 and become eligible for Medicare). Elderly people make many more errors when using comparative information in charts and tables than younger people, and those in poor health or with less education are the most likely to err in interpreting comparative information presented in such formats (Hibbard, 2001). Nearly one-fifth of Medicare beneficiaries have medical or psychiatric conditions that make it difficult for them to make decisions (Moon and Storeygard, 2001). While most beneficiaries agree that decisions about choosing health plans are both important and difficult to make, few seek help in choosing a health plan (McCormack and Garfinkel,2001).

CMS created a helpline and Web site to assist Medicare beneficiaries find information about Medicare insurance options. A study by the Kaiser Family Foundation (2004) found that 60 percent of Medicare beneficiaries had heard of the Medicare helpline, but only 10 percent had called it; 13 percent had heard of the Medicare Web site, but only 4 percent had used it.

One study suggests that Medicare beneficiaries' information-seeking behavior may differ for Medicare drug discount cards. An AARP survey of Medicare beneficiaries found that 46 percent of respondents sought information to help them choose a drug discount card; one-third chose a Medicare drug discount card because it was offered by a trusted source, and 11 percent chose a card that was recommended by their pharmacist; and only 20 percent said they chose their card because it offered the best discounts (Love, 2004). A Kaiser Family Foundation (2006) study found that 27 percent of beneficiaries talked to a pharmacist about Medicare drug benefits.

Study Methods

Our findings stem from three sources of data: (1) two national surveys of Medicare beneficiaries enrolled in drug discount cards, (2) 54 focus groups with Medicare beneficiaries, and (3) 243 interviews and other qualitative data collection with stakeholders.

Beneficiary Surveys

Two national surveys of Medicare drug discount card enrollees were conducted, in fall 2004 and spring 2005. The sample frame, drawn from CMS databases, consisted of all Medicare beneficiaries enrolled in drug discount cards except those who (1) were enrolled in special endorsement cards, (2) had end-stage renal disease, (3) were over age 85, or (4) had effective card dates after July 1, 2004, for the first survey and January 1, 2005, for the second. The latter exclusion ensured that survey participants had had time to use their cards.

Since the findings from the two surveys were quite similar, unless otherwise noted the results presented here are from the later of the two surveys. For a full description of the methods and results for both surveys and for the focus groups refer to (Abt Associates Inc., 2006a).

The sample for the 2005 survey was selected in two stages. First, a purposive sample of 27 drug discount cards was selected from among those cards with an enrollment of at least 600 beneficiaries with and 600 beneficiaries without transitional assistance. Following CMS instructions, most of the largest national cards were selected, with an effort to achieve a rough geographic balance. Of the 27 cards selected, 4 were large exclusive cards offered by Medicare Advantage plans to their members. Most of the 27 selected cards were national, but a few were regional cards with the requisite number of enrollees. Overall enrollment in the 27 selected cards was 3,282,793 which represented 58.7 percent of all Medicare drug discount card enrollees who met eligibility criteria for the survey. Beneficiaries who enrolled in the largest drug discount cards may have differed from those who enrolled in smaller cards; we are unable to estimate the size or direction of any bias such a difference may have introduced.

In the second stage of the sampling process an independent sample of 600 enrollees with and 600 enrollees without transitional assistance was selected from each of the 27 drug discount cards, for a total sample of 32,400 enrollees.

The 2005 survey was mailed in mid-April 2005. Of the 32,400 drug discount card enrollees selected, 22,021 returned a survey with at least the first question answered; another 258 indicated that they were not aware of being enrolled in a Medicare drug discount card, and 40 indicated that they had a card, but had not yet used it. With these responses included, the total number of respondents was 22,319, resulting in a response rate of 69 percent; 6 percent were completed by proxy. Responses were weighted to reflect the size and composition of each of the individual cards' enrolled populations, and adjusted for non-response.

Awareness of enrollment was quite low among enrollees in exclusive drug discount cards (23 percent) compared to those in non-exclusive cards (64 percent); therefore exclusive drug discount card enrollees were removed from further analysis.

Beneficiary Focus Groups

We conducted 54 focus groups with 436 total participants in 15 cities: 30 focus groups in 8 cities during September-October 2004, and 24 focus groups in 7 cities during February-March 2005. Cities were selected for geographic diversity and to concentrate on places where early card enrollment was high. The sample frame for the focus groups consisted of all Medicare beneficiaries, whether or not they were enrolled in drug discount cards. The same four exclusionary criteria were used as for the surveys. There were six types of beneficiary focus groups (Table 1). Separate focus groups were held with drug discount card enrollees and with non-enrollees.

Table 1

Number of Beneficiary Focus Groups, by Group Type

Focus Group Type	Number of Groups	Number of Participants
Card Enrollees Without Transitional Assistance, Age 65 or Over	16	151
Card Enrollees With Transitional Assistance, Age 65 or Over	12	88
Non-Enrollees (Not Eligible for Transitional Assistance), Age 65 or Over	12	89
Non-Enrollees with Limited Incomes (Eligible for Transitional Assistance), Age 65 or Over	6	32
Card Enrollees Medicare Eligible Due to Disability	4	37
Card Enrollees with Transitional Assistance, Medicare Eligible Due to Disability	4	38
Total	54	436

SOURCE: Abt Associates Inc., 2004 and 2005 Focus Groups on Medicare-Approved Drug Discount Cards.

Screening questions were used to verify the enrollment status and eligibility criteria of each beneficiary during recruitment, including income (i.e., meeting income eligibility criteria for transitional assistance among non-enrollees). Many focus group candidates, who were enrolled in drug discount cards according to CMS data, were not aware of their enrollment. We excluded these beneficiaries from enrollee focus groups because we wanted participants who could comment on their reasons for enrollment. Thirty-nine percent of card enrollees without and 62 percent of enrollees with transitional assistance contacted during recruitment in 2005 said they were not aware that they had a Medicare drug discount card.

Stakeholder Interviews, Site Visits, and Focus Groups

We conducted indepth interviews, site visits, and focus groups with several types of stakeholders. Data were collected in two phases in order to capture stakeholders' early perspectives on the drug discount card program during Phase I (November 2004-February 2005) and to document changes in stakeholders' perspectives on the program during Phase II (August-November 2005). Many respondents' perspectives did not change appreciably, therefore unless otherwise noted, the themes presented recurred in both phases.

In Phase I, we conducted 117 indepth interviews with representatives from drug discount card sponsors, community pharmacies, prescription drug manufacturers, and organizations helping to educate beneficiaries about the new program. We also interviewed 20 experts with other vantage points including industry professionals and consultants, and academic researchers. We invited all 137 Phase I respondents to participate in a second interview in Phase II, and 92 repeat interviews were completed. In Phase II, we also talked with 14 representatives from SPAPs (Table 2).

Table 2

Number of Participants in Stakeholder Interviews, by Participant Type

Participant Type	Number of Phase I Interviews	Number of Phase II Interviews
Pharmacy Sector	39	28
Pharmacy Executives	17	10
Pharmacists in Chain Pharmacies	10	9
Pharmacists in Independent Pharmacies	12	9
Card Sponsors (Private Companies Offering Drug Discount Cards) General, Exclusive, Special Endorsement	32	21
Organizations Helping Beneficiaries	30	18
SHIP Program Directors	22	11
Information Intermediaries and Beneficiary Advocates	8	7
Experts	20	15
Professional Associations	10	7
Thought Leaders	10	9
State Pharmacy Assistance Programs	N/A	14
Drug Manufacturers	16	9
Total	137	106

NOTES: Phase I: November 2004-February 2005; Phase II: August-November 2005. SHIP is State Health Insurance Assistance Program. N/A is not applicable.

SOURCE: Abt Associates Inc., 2004 and 2005 Stakeholder Interviews on Medicare-Approved Drug Discount Cards.

In Phase I we visited four drug discount card sponsors, and in Phase II we conducted four community case studies, which included multiple in-person interviews and focus groups with community pharmacists. As part of the case studies, we also analyzed coverage of the drug card program in six major newspapers and two senior monthlies. (Refer to Abt Associates Inc., 2005; 2006b for a full description of the methods and results.)

Findings

Beneficiaries' Experiences with Choice

Awareness and Understanding of Card Options

Fifty-nine percent of the 2005 survey respondents reported that they were aware they had a choice among Medicare drug discount cards (Table 3). Low income survey respondents with transitional assistance were less likely than those without transitional assistance (52 versus 64 percent, p<0.01) to be aware that they had a choice among Medicare drug discount cards (data not shown).

Table 3

Enrollees' Awareness of Card Choices and Consideration of Multiple Cards: 2005 Survey

Respondent Awareness	Before you got this survey, were you aware that there is more than one Medicare-approved drug discount card that you could apply for?	Did you consider more than one Medicare-approved drug discount card before you settled on the one you have now?
	Percent
Yes	59	31
No	34	63
Do Not Know	6	3
Not Answered	1	3
Total	100	100

NOTES: n=10,935.

SOURCE: Abt Associates Inc., 2005 Survey on Medicare-Approved Drug Discount Cards.

The survey asked about beneficiaries' understanding of other relevant program features. Depending on the question, approximately one-quarter to one-half of beneficiaries demonstrated a correct understanding of program features (Table 4). Forty-one percent knew that they could have only one Medicare drug discount card at a time; 32 percent knew that a Medicare drug discount card was not the same as drug insurance; 31 percent understood that a drug discount card did not guarantee discounts on all drugs at any pharmacy; but only 23 percent understood that Medicare drug discount card holders could also have a discount card from another source like a drug manufacturer or drug store. And 52 percent understood that the price paid when using a Medicare drug discount card depended on generic versus brand name drug purchases. Understanding among enrollees about these programmatic features did not improve between the two survey periods.

Table 4

Enrollees' Knowledge of Drug Discount Card Program Features: 2005 Survey

Survey Question	Agree	Disagree	Don't Know	No Answer
	Percent
You can only have one Medicare-Approved drug discount card at a time.	411	11	38	10
A Medicare-Approved Drug Discount Card is the same as having insurance for prescription drugs.	25	321	34	9
With a Medicare-Approved Drug Discount Card you get discounts on all prescription drugs, at any pharmacy.	21	311	37	10
If you have a Medicare-Approved Drug Discount Card, you can also have other discount cards sponsored by drug manufacturing companies or drug store chains.	231	15	53	9
When you use your Medicare-Approved Drug Discount Card, the price you pay will depend on whether you are buying a generic drug or a brand name drug.	521	7	33	8

Indicates the correct answer.

NOTES: n=10,935.

SOURCE: Abt Associates Inc., 2005 Survey on Medicare-Approved Drug Discount Cards.

Depending on the focus group, one-quarter to one-half of participants were unaware that they could choose among many drug discount cards. Although awareness of the program was reasonably high in all focus groups, detailed understanding was not. Most participants could describe a few features of the program such as the $600 subsidy, price discounts, and the temporary nature of the program, but few were able to explain it in more detail. The aspect of the program that enrolled focus group participants understood most clearly was that they needed to present their Medicare drug discount card to the pharmacist when filling a prescription (at least the first time) in order to receive a discount. The aspect of the program that seemed to cause the most confusion in fall 2004, especially among those with transitional assistance, was how the $600 subsidy worked in conjunction with discounts/benefits from other programs such as SPAPs. By the winter 2005 fewer focus group participants were confused about benefit coordination.

Sources of Information and Information-Seeking Behavior

Survey respondents, all of whom were enrolled in Medicare drug discount cards, were asked to indicate the sources of information they used when deciding about a Medicare drug discount card, including CMS' information channels. Over one-third of survey respondents cited pharmacists as an important information source (Table 5). Other commonly mentioned information sources were mass media (especially television), the Medicare helpline, and friends and family members (especially among respondents with transitional assistance). Those with transitional assistance were also more likely to get information from a government agency, and less likely to get information from an insurance company or counselor. Fewer than 10 percent of survey respondents had used the Medicare Web site and only 3 percent mentioned using any type of information counselor or service (including State Health Insurance Assistance Programs [SHIP]). While many respondents had not used these resources, 59 percent of 2005 survey respondents reported that they had adequate information to make the enrollment decision (data not shown).

Table 5

Enrollees' Sources of Information About the Program: 2005 Survey

Please check all the places where you got information when you were deciding about your Medicare-approved drug discount card	All Respondents1	According to CMS Records
		With Transitional Assistance2	Without Transitional Assistance3
	Percent
Pharmacist or Pharmacy	38	39***	37
Media	24	25***	23
1-800-Medicare	20	24***	16
Family or Friends	15	19***	12
Web site (e.g., www.Medicare.gov)	9	6***	12
Doctor or Other Medical Person	6	7***	5
Drug Manufacturer	6	6***	7
Other Source of Information	6	6***	7
State/County/City Agency	5	7***	3
Health Insurance Company or Agent	5	3***	6
Got No Information When Choosing Card	4	3***	5
Health Insurance Counselor or Information Service	3	3***	4
Did Not Answer	3	3***	3
Employer or Former Employer	0	0***	0

p<0.01.

n=10,935.

n=7,129.

n=3,806.

NOTES: Respondents could check more than one category. Therefore, tests of significant differences (using chi-square) between enrollees are at the category level (rows) and totals do not sum to 100 percent. Due to rounding, some percentages appear to be identical, but are in fact statistically different.

SOURCE: Abt Associates Inc., 2005 Survey on Medicare-Approved Drug Discount Cards.

Almost no one in any of the focus groups had used (or recognized the name of) their local SHIP, although many expressed a desire to meet with an impartial expert to get advice. Among those who recalled getting a CMS mailing, but who did not enroll, most commented that the material they received was either difficult to understand or not sufficiently detailed. Some also reported that they did not read CMS mailings. Most of those who used the Medicare helpline did eventually enroll in a Medicare drug discount card, while few of those who did not enroll had called the Medicare helpline for information.

Decisionmaking Process

Choice Among Cards

Fifty-nine percent of survey respondents knew there were multiple Medicare drug discount cards, but only 31 percent reported that they considered more than one drug discount card when enrolling (Table 3). Respondents with transitional assistance were less likely than those without transitional assistance to consider more than one drug discount card and to know that multiple cards existed (data not shown). The top three reasons survey respondents reported for choosing their particular card were that (1) it was accepted by their pharmacy (77 percent), (2) it lowered costs and helped pay for drugs (64 percent), and (3) it provided discounts on their particular drugs (56 percent). Twenty-eight percent of respondents chose their card because of their pharmacist's recommendation (Table 6).

Table 6

Enrollees' Reasons for Choosing Their Medicare Drug Discount Card 2005 Survey

Check all of the reasons that you chose your Medicare-Approved Drug Discount Card	All Respondents1	According to CMS Records
		With Transitional Assistance2	Without Transitional Assistance3
	Percent
Pharmacies I use will accept my card	77	80***	74
Lower Costs and Helps Pay for Drugs	64	76***	52
Discounts on Drug Bought	56	61***	51
Only Card I Looked Into or Considered	35	37***	33
Annual Enrollment Fee for Card Was Acceptable to Me	35	28***	41
My Pharmacist Recommended Card	28	30***	25
A Friend or Family Member Recommended Card	14	18***	9
A Medicare Counselor or Information Service Recommended Card	11	14***	8
Other Reason	9	8***	10
A Doctor or Other Medical Person Recommended Card	7	9***	5
A Health Insurance Agent or Company Recommended Card	5	3***	7
Did Not Answer	2	1	2

p<0.01.

n=10,935.

n=7,129.

n=3,806.

NOTES: Respondents could check more than one category. Therefore, tests of significant differences (using chi-square) between enrollees are at the category level (rows) and totals do not sum to 100 percent.

SOURCE: Abt Associates Inc., 2005 Survey on Medicare-Approved Drug Discount Cards.

Many focus group participants reported that they enrolled in a card that was connected in some way with insurance companies or other firms with which they already had relationships. For example, in Jacksonville, many participants who had supplemental insurance policies with Blue Cross® of Florida also enrolled in that firm's Medicare drug discount card. Focus group participants who were aware of choices reported on the factors that were most important in their enrollment decision: the prices for their medications, which pharmacies accepted the card, and the price of the card itself. Many beneficiaries were not aware of the variations in discounts and formularies among cards. Others noted that they were influenced by the recommendations of physicians and pharmacists or by the plan's customer service staff.

Reasons for Not Enrolling

Among focus group participants who had not enrolled in Medicare drug discount cards, some had heard about the program, but had decided not to enroll because of misperceptions about eligibility or doubts about the value of the cards. The most common misperception was that only persons with limited incomes could enroll in a Medicare drug discount card. Some beneficiaries with limited incomes who were not enrolled were under the mistaken impression that they would have to pay a monthly premium to obtain a drug discount card and the transitional assistance benefit. Several focus group participants reported that the standard prices at discount retailers (Costco, Sam's Club) were lower without the card than with it. Others had few prescriptions to fill or felt that the senior discount offered by their local pharmacy was better than the discount available from Medicare drug discount cards. Some focus group participants received information about several cards, but were overwhelmed by the number of options and unable to choose. A few focus group participants knew that the program would be temporary and did not want to engage in a complicated choice process for a short-term benefit.

Stakeholders' Comments on Beneficiaries' Choice Process

Beneficiary Awareness and Understanding

Stakeholders were asked about their own experiences with beneficiary choice; some stakeholders emphasized that beneficiaries (especially low income beneficiaries) were hard to reach, and that beneficiaries were confused about the program. Drug discount card sponsors remarked that reaching this population was much harder than they expected and community-level stakeholders pointed out that it was particularly hard to identify and locate beneficiaries who were eligible for the $600 transitional assistance subsidy. Factors stakeholders identified that made these beneficiaries challenging to reach and educate included: low literacy levels, physical and cognitive impairments, skepticism of government programs, reluctance to divulge income information or identify themselves as low-income, program complexity, and beneficiaries' doubts as to whether the Medicare drug discount cards offered greater savings than other, more familiar alternatives.

In one area, stakeholders' perceptions clearly diverged from beneficiaries': few stakeholders mentioned that beneficiaries were unaware of the need to choose among cards; rather they felt beneficiaries were confused by the large number of cards and the need to choose among them.

Beneficiary Sources of Information and Information-Seeking Behavior

In 2004, stakeholders stated that CMS communications directed at beneficiaries and the public and were very important to the program's success. Many commented on the well-publicized delays and queues for people calling 1-800-Medicare and suggested that beneficiaries' initial difficulties in gaining information through this channel contributed to low levels of enrollment and to negative public attitudes toward the entire program.

Stakeholders took notice of the Medicare drug discount card price comparison tool on the CMS Web site. Some recognized it as an innovative feature of the drug discount card program, both because it offered accessible information on drug prices and because it facilitated individualized decisionmaking. However, virtually all respondents strongly agreed that whatever the Web site's strengths, CMS could not rely on it as a primary mode for communicating directly with beneficiaries because at that time most beneficiaries did not yet have the computer access and skills required to use it.

In 2004, stakeholders observed that negative and politicized media coverage had hurt the program. Our own analysis of newspaper coverage of the drug discount card program in four case study communities indicated that articles provided basic information about the program, but were generally negative about the Medicare Web site and call center.

In discussing their efforts to market to beneficiaries in 2004, drug discount card sponsors stated that traditional marketing methods, especially the use of print materials, did not reach this population. Because initial direct marketing was not successful, most sponsors worked through new or existing partnerships with organizations that already had relationships with beneficiaries, such as insurers, employers, or unions. This is consistent with beneficiaries' stated preference for cards from trusted sources.

In 2005, we asked stakeholders about the informational materials they themselves needed when beneficiaries came to them for information and help with decisionmaking. Pharmacists and other organizations serving beneficiaries expressed a desire for clear, concise, accessible information about the Medicare drug discount card program and the drug discount cards being offered, ideally in a format that facilitated inter-card comparisons. In addition, stakeholders wanted instructional information that would help beneficiaries and staff through the process of making decisions and using the available Internet tools.

Beneficiary Decisionmaking Process

In 2004 and 2005, stakeholders were asked about their experiences with helping beneficiaries decide whether to enroll in a drug discount card and which card to choose. They said beneficiaries needed personalized, one-on-one support to compare the large number of choices and to understand which Medicare drug discount card would best meet their needs. Organizations serving beneficiaries (SHIPs, information intermediaries, and community service organizations) remarked that it was challenging, time-consuming, and expensive for them to offer these services to beneficiaries. Many stakeholders used and appreciated the automated tools on the CMS Web site (or similar private-sector products).

In 2005 stakeholders defined some best practices for helping beneficiaries to make decisions about whether to enroll in a drug discount card and which card to choose: ensuring computer access at outreach sites, training volunteers in use of existing tools, and providing general program education through print media or in a group setting followed by one-on-one assistance in choosing a plan.

In 2005, using both interviews and focus groups, we asked pharmacists about their role in assisting beneficiaries with issues related to drug discount card benefits and costs and found pharmacists to be conflicted on the subject. Many said that they would prefer to focus on clinical and drug therapy issues and not drug insurance benefits, but many accepted these additional responsibilities as part of their job and even found it rewarding to help their customers in this way.

In both rounds of interviews we asked about State/sponsor partnerships in which members of SPAPs were automatically enrolled into a preferred drug discount card (always with an option to select another). Auto-enrollment effectively softened the requirement that beneficiaries choose among cards, by giving them a passive/default option. In 2005, several of the sponsors and approximately one-half of the SPAPs interviewed had been involved in these partnerships, and in general, were pleased with the increased enrollment and reduced administrative burden. In addition, several information intermediaries observed that auto-enrollment had worked well; none expressed concerns about curtailing active choice.

Discussion

The original purpose of this study was to yield lessons and insights relevant to the launch of the Part D drug benefit. While that original purpose is now moot, the drug card experience remains relevant to the Part D program over the long term and to the Medicare Program more generally.

Our primary finding was that beneficiaries did not embrace the concept of choice among drug cards. Our data suggest that many beneficiaries approached information gathering and card comparison passively, not seeking the best card, but rather being satisfied with one that was good enough in terms of discounts and other features. This minimal effort may stem from low levels of program knowledge, poor access to information and assistance, and/or an informed judgment that the differences among cards and the value of the best choice were not sufficiently significant to merit more effort.

Stakeholders' comments emphasized the pain (not the gain) of choice, noting the difficulty of reaching beneficiaries, the effort required to help them to choose among cards, and stakeholders' own unmet needs for informational resources. The contrast between stakeholders' emphasis on beneficiaries' confusion/need for help and beneficiaries' general satisfaction may stem from the fact that stakeholders only saw those beneficiaries who requested help, not those who remained unenrolled or made decisions independently. Like many beneficiaries, stakeholders were not fully aware of the resources available from CMS.

The drug card experience highlighted some of the challenges in making choice work in practice. Addressing these challenges requires effective educational campaigns, widely available one-on-one assistance, and program options that beneficiaries perceive as truly different and valuable. The health care sector must recognize that some beneficiaries need encouragement and help to compare multiple options, and that beneficiaries lean toward products offered by familiar and trusted organizations. Finally, proponents of choice must communicate its potential to contribute to widely shared goals, such as access, quality, and efficiency.

The study raises three specific issues that deserve further comment. First, while auto-enrollment offered increased enrollment and reduced administrative burden, it carried potential unintended consequences: reduced awareness of enrollment and program knowledge. We found that only 23 percent of those enrolled in exclusive cards (according to CMS records) were aware of this enrollment, an outcome that may be due to managed care plans enrolling their members automatically into associated drug cards without beneficiaries realizing that the change had occurred. Similarly, we found that beneficiaries with transitional assistance were less likely than those without to be aware of card enrollment or to know they had a choice among cards. Again, auto-enrollment could have contributed to this discrepancy if beneficiaries with transitional assistance were disproportionately likely to be receiving assistance from SPAPs, be auto-enrolled, and never fully understand the new Federal Program. If benefits coordination was smooth, some of these beneficiaries could have used cards and spent Federal dollars without ever understanding that the Medicare drug discount card program was involved.

Second, this study confirmed the importance of CMS communications with beneficiaries. After the pharmacist and family/friends, the CMS helpline was the most commonly used information source. Paradoxically, the helpline delays actually demonstrate the importance of this source of information. We observed a correlation between calling the 1-800-Medicare helpline and program enrollment, although it is not clear whether the helpline actually promoted enrollment or was just a logical step for those already on that path. While beneficiaries did not commonly use the Medicare Web site, stakeholders who helped beneficiaries appreciated this resource, and the proportion of seniors using the Internet is likely to increase dramatically in the next decade (Kaiser Family Foundation, 2005).

Lastly, during the drug card program, pharmacies offered an important channel for reaching beneficiaries: beneficiaries named pharmacists more often than any other source of program information. While pharmacists acknowledged this role, they expressed ambivalence about it. Given that both this potential opportunity and this ambivalence are likely to persist in the Part D drug benefit, it will be important for the health care sector to identify and take any possible steps to enlist the pharmacy sector as a partner in outreach about Part D.

While the drug card experience sheds light on choice in Medicare, it is important not to over-generalize. The Medicare drug discount card program was new and required information and skills unfamiliar to both beneficiaries and the organizations that offered the benefit. Had the program persisted, some of the challenges associated with the choice might have dissipated as beneficiaries and the wider community gained experience with the Medicare drug card.

The program was also unusual in that the financial and health consequences of the wrong choice were relatively low and the program was known to be temporary. By contrast, decisions among Part D drug plans and Medicare Advantage health plans have greater consequences, and these programs are intended to be permanent features of the Medicare Part D Program. Given these differences, it is reasonable to expect that stakeholders will invest greater effort in publicizing the Part D program and the advantages of various plans. In addition, beneficiaries and their families may be willing to work harder to understand the program and make informed choices.