Search tips
Search criteria 


Logo of hsresearchLink to Publisher's site
Health Serv Res. 2004 August; 39(4 Pt 2): 1055–1070.
PMCID: PMC1361053

Awakening Consumer Stewardship of Health Benefits: Prevalence and Differentiation of New Health Plan Models



Despite widespread publicity of consumer-directed health plans, little is known about their prevalence and the extent to which their designs adequately reflect and support consumerism.


We examined three types of consumer-directed health plans: health reimbursement accounts (HRAs), premium-tiered, and point-of-care tiered benefit plans. We sought to measure the extent to which these plans had diffused, as well as to provide a critical look at the ways in which these plans support consumerism. Consumerism in this context refers to efforts to enable informed consumer choice and consumers' involvement in managing their health. We also wished to determine whether mainstream health plans—health maintenance organization (HMO), point of service (POS), and preferred provider organization (PPO) models—were being influenced by consumerism.

Data Sources/Study Setting

Our study uses national survey data collected by Mercer Human Resource Consulting from 680 national and regional commercial health benefit plans on HMO, PPO, POS, and consumer-directed products.

Study Design

We defined consumer-directed products as health benefit plans that provided (1) consumer incentives to select more economical health care options, including self-care and no care, and (2) information and support to inform such selections. We asked health plans that offered consumer-directed products about 2003 enrollment, basic design features, and the availability of decision support. We also asked mainstream health plans about their activities that supported consumerism (e.g., proactive outreach to inform or influence enrollee behavior, such as self-management or preventive care, reminders sent to patients with identified medical conditions.)

Data Collection/Extraction Methods

We analyzed survey responses for all four product lines in order to identify those plans that offer health reimbursement accounts (HRAs), premium-tiered, or point-of-care tiered models as well as efforts of mainstream health plans to engage informed consumer decision making.

Principal Findings

The majority of enrollees in consumer-directed health plans are in tiered models (primarily point-of-care tiered networks) rather than HRAs. Tiers are predominantly determined based on both cost and quality criteria. Enrollment in HRAs has grown substantially, in part because of the entry of mainstream managed care plans into the consumer-directed market. Health reimbursement accounts, tiered networks, and traditional managed care plans vary in their capacity to support consumers in managing their health risks and selection of provider and treatment options, with HRAs providing the most and mainstream plans the least.


While enrollment in consumer-directed health plans continues to grow steadily, it remains a tiny fraction of all employer-sponsored coverage. Decision support in these plans, a critical link to help consumers make more informed choices, is also still limited. This lack may be of concern in light of the fact that only a minority of such plans report that they monitor claims to protect against underuse. Tiered benefit models appear to be more readily accepted by the market than HRAs. If they are to succeed in optimizing consumers' utility from health benefit spending, careful attention needs to be paid to how well these models inform consumers about the consequences of their selections.

Keywords: Consumer-directed health plans, health reimbursement accounts, consumerism, tiered networks

Accelerating growth in health insurance premiums coupled with an economic downturn have generated a renewed focus on cost control in the U.S. health benefits sector. The prevailing vision for cost control in the current employer-sponsored health benefit market does not, however, call for increasingly restrictive managed care plans (Galvin and Milstein 2002). Desire for broad choice and rejection of explicit rationing is widespread, a phenomenon that was in part responsible for the managed care backlash. More than 40 percent of adults surveyed nationally do not support any restriction on choices of physicians, hospitals, or treatment options (Employee Benefit Research Institute 2003) even if such restrictions would result in lower health care costs.

A number of employers and health insurers have embraced new health benefit models with increased consumer incentives to select options that reduce health plan spending and possibly also to select higher-quality options, accompanied by more flexibility with regard to provider and treatment choices. Incentives may encourage more economical or higher-quality selections in all health care decisions or may target only a subset. The primary stimulus of this so-called consumer-directed health benefits movement has clearly been provided by the perceived need to reduce spending, but its stated goals also include enhancing quality or the ratio of health gain to health insurance spending (value). Sponsors of consumer-directed health benefits often suggest that enabling “consumerism” in health care is the primary objective of these new plans. Critics, however, worry that consumer-directed health plans merely shift more costs onto all consumers or to sicker consumers without conferring upon them the necessary tools to select higher value health care options.

Aside from financial incentives for consumers to select lower-cost and possibly higher-quality options, “consumerism” frequently incorporates two additional concepts: (1) informed choice and (2) active consumer participation in managing health and health care decision making (the consumer as “coproducer” of health as described in the literature) (Hibbard 2003). Informed choice of health plans on the basis of reported clinical quality and patient experience has been the primary emphasis of efforts to leverage consumer involvement to improve health care quality over the past several decades. Newer models more heavily emphasize informed selection of provider options. The typical assumption of consumer choice models is that consumers will not only select better (e.g., higher-quality) options resulting in better cost or quality outcomes in the short run but also that health plans, physicians, and hospitals will thereby be encouraged to compete on the basis of the performance measures that are reported. While health plan and provider report cards have met with relatively disappointing results to date (Scanlon et al. 1998; Schneider and Epstein 1998; Hibbard and Peters 2003), there have been improvements in both measurements and their communication to consumers.

Engagement of consumers in managing their own health risks and making informed decisions about treatment options (including not seeking treatment) builds on preexisting managed care methods; these include health risk assessments, information about self-care and management of chronic conditions, information and patient reminders about preventive health measures, nurse-staffed telephone help lines, and shared decision-making programs (Hibbard 2003). A growing literature documents the effectiveness of these methods, such as reminders and self-care education for improving health outcomes for individuals with diabetes, asthma, and depression (von Korff et al. 1997; Clark 2003).

At the present, the extent of these changes in health benefit plans are unknown, despite the abundance of articles on their policy and business implications (Fronstin 2002; Robinson 2002). The only published empirical analysis of this emerging trend found that, while growing, consumer-directed health plan enrollment remained low in 2002. The study, which relied on key informant interviews, reported a high degree of variation around plan models and features among the class of plans considered to be consumer-directed. It also suggested that large national and regional health plans were beginning to view consumer-directed models as strategically important products, which might consequently lead to wider diffusion in 2003 and beyond (Gabel, LoSasso, and Rice 2002). In addition to assessing the current prevalence of new models, a key puzzle to unravel is whether consumer-directed health plans provide the necessary tools to engage consumers in choosing and participating in managing their own health.

We sought to update and broaden previous research through a national health plan survey in the first quarter of 2003. Our research examines two broad categories of consumer-directed health plans: (1) health reimbursement account models, and (2) tiered benefit models. Our principal goal was to measure the uptake of these consumer-directed products and examine the extent to which they actively support consumerism. For comparison, we also wanted to gauge the extent to which mainstream health plans are incorporating incentives to select more economical health care options and providing information to support those selections (“decision support”). To this end, we examined the prevalence of such incentives and decision-support strategies among mainstream health plans—specifically, health maintenance organization (HMO), point of service (POS), and preferred provider organization (PPO) plans.

Consumer-Directed Health Plan Models

Most of the press and policy discussion about consumer-directed health plans has focused on so-called health reimbursement account (HRA) models. These plans represent the most distinct departure from mainstream managed care plans, presenting consumers with financial incentives to make cost-conscious choices over a wide range of health care spending decisions up to the plan's maximum out-of-pocket limit. Health reimbursement account models typically combine a high-deductible insurance plan (almost always in the form of a PPO) with an employer-funded account (called, variously, the health reimbursement account, personal care account, personal medical fund, and many other similar terms). The employer-funded account may be used to pay for covered health care services and is generally counted toward the deductible amount. Thus an HRA with a $500 employer-funded account and a $1,500 deductible implies that once the employer-funded account is depleted, the consumer must spend $1,000 out-of-pocket before insurance will begin sharing the costs of treatment. The psychological aim of the HRA is to induce consumer stewardship for the entire $1,500, rather than for $1,000, and perhaps to set in motion a more careful attitude toward all health care spending.1 Unexpended funds from the employer-funded account within an HRA, unlike a flexible spending account, may be rolled over (at the discretion of the employer). Also unlike flexible spending accounts, many HRA models use debit cards or require providers to file claims rather than ask enrollees to pay up front for services and file claims for reimbursement.

Tiered-benefit model plans include two distinct types: those that tier premium contributions and those that tier point-of-care cost sharing. Premium-tiered model plans require consumers to contribute more if they select a less-restrictive network, looser utilization management features, or more generous insurance coverage (e.g., lower copayments or coinsurance). Health Net's Vivius product and Humana's Smart Suite and Smart Select products are examples of premium-tiered models, which have also been referred to in the literature as “customized” plans. These plans do not necessarily introduce novel insurance models, although some include an HRA or point-of-care tiered model as an option. Point-of-care tiered models reduce point-of-care cost sharing if consumers select a provider deemed by the insurer to be preferred and therefore placed in a less costly tier. In our analysis, we explicitly exclude from this category mainstream PPO and POS products in which copayment differentials are primarily a function of whether the provider has agreed to accept a discounted reimbursement rate from the plan or to cooperate with care management requirements. Point-of-care tiered models typically start with a PPO or POS contracted network and then introduce differential cost sharing within the network based on broader measures of cost, quality, or both. For example, many of the models include responses to the Leapfrog Group safety survey as quality of care criteria for placing hospitals in tiers.

Data and Methods

We analyzed data from a national health plan inventory to describe the prevalence of consumer-directed health benefits in the United States. Mercer Human Resource Consulting collected information in 2003 from 680 health plans on a total of 986 HMO, POS, and PPO products to assist purchasers in contracting decisions. The plans report on the design and performance of these products for both self-insured and insured options. Mercer Human Resource Consulting attempts to gather information from the universe of commercial health plans in the United States by combining lists of potential respondents from a variety of sources including, but not limited to, Interstudy, state regulatory reports, and the Managed Care Information Center. As is common practice for health plan surveys, for plans that operate in multiple markets, we count each local or regional entity separately (for those that responded with separate information by market). In total, 70 percent of plans responded to Mercer Human Resource Consulting's request for information and completed an extensive web-based instrument.

The main product-line requests for information ask plans to report a wide variety of information on product characteristics and capabilities. Through a supplemental section, we added questions specifically related to consumer-directed health benefits. Questions in the supplemental section were developed with the aid of a panel of national subject matter experts in health economics, consumer decision making, and health policy.

To quantify the prevalence of consumer-directed health plans, we asked plans to report the number of enrollees that were covered by their HRA model and, separately, any premium-tiered or point-of-care tiered models. For the HRA model, we also asked about the dollar amount of the typical employer-funded account and deductible. For the tiered benefit products, we asked the plans to report whether provider-based tiering was a function of cost, quality, or both, and the typical annual cumulative out-of-pocket difference per enrollee between the most preferred and least preferred provider. The survey also included questions about decision support that were targeted to all consumer-directed health plans (HRA or tiered benefit models.) We first asked plans whether they provided enrollees with information: (1) regarding the average cost of procedures/services such as a routine office visit, (2) to help choose an individual physician or medical group based on comparative cost, (3) to help choose an individual physician or medical group based on comparative quality, (4) to help choose a hospital based on comparative cost, (5) to help choose a hospital based on comparative quality, (6) to help choose a drug based on comparative cost, (7) to help choose other types of options based on cost, (8) to help self-manage a chronic condition. We also ascertained the availability of a nurse-staffed telephone help line.

To put in context our findings about consumer-directed health plans, the survey asked respondents to report for their typical HMO, POS, or PPO plan whether they had raised, lowered, or left unchanged point-of-service consumer cost sharing and by how much. We asked respondents to include total annual estimated copayments, coinsurance, and deductibles in the calculation of increases in cost sharing. In addition, we included a series of questions to capture efforts of these mainstream plans to engage consumers in making informed health care decisions. First, we asked whether there was proactive outreach to help members with identified medical conditions manage their health. We also asked specifically whether reminders were sent to appropriately identified patients for preventive care services (which include both primary and secondary prevention): cervical cancer screening, cholesterol screening, colorectal cancer screening, diabetic retinal exam, influenza vaccine, childhood immunizations, mammograms, and prostate cancer screenings. Plans were also able to write in other services for which reminders were sent to patients. The survey also captured whether members could complete a health risk assessment on the plan's website, and also whether the website allowed members to develop a health profile. Because just offering health management tools may not be sufficient to motivate active participation, we also asked whether plans offered incentives to promote health improvement. Finally, as we did with the consumer-directed health plan models, we asked about the availability of nurse-staffed telephone help lines.

We report the numbers and percentages of respondents that offered consumer-directed products and totals for enrollment and contracting purchasers. For the dollar values of the employer-funded account and deductible portion of HRA models as well as the gap between the most- and least-preferred provider or drug in the point-of-care tiered products, enrollee-weighted mean and modal values are presented. To describe decision-support features, the direction of changes in cost sharing, and the bases for classifying providers into tiers, we report enrollee-weighted frequencies.


Health Reimbursement Account Models

Table 1 reports the number and percent of plans that offered an HRA model and describes enrollment in and selected features of plans. In total, there were 24 active HRA products in our sample as of January 1, 2003. These plans reported 466,039 enrollees. More than half of these enrollees were covered by four plans that offer only HRA models; the remaining half were scattered across 20 plans that also offer HMO, POS, or PPO products (data not shown). Only 13 plans had enrolled more than 1,000 beneficiaries as of the beginning of 2003. Nine health plans described an HRA product, but did not report any enrollment as of January 1, 2003. These plans may have launched later in the year or have been in the process of contracting for 2004.

Table 1
Health Reimbursement Account Models

The average enrollee received $824 in the employer-funded account and faced a deductible of $1,654 (the modal amounts for the account and deductible were $1,000 and $1,500, respectively).

Premium or Point-of-Care Tiered Models

Table 2 reports the number of employers and enrollees that participated in premium-tiered benefit plans as well as selected features of tiered point-of-care models. While a somewhat larger number of plans reported that they offered a premium-tiered benefit plan, enrollment was greatest in point-of-care tiered plans. Point-of-care tiered plans had enrolled more than 1.5 million beneficiaries as of January 1, 2003. For the point-of-care tiered plans, in which consumers are asked to pay more at the point of service for lower-tiered providers, we found that the enrollee-weighted average annual cumulative out-of-pocket cost difference between the most and least preferred providers was $609. Almost all enrollees in tiered benefit models were in plans that placed providers in tiers based on both cost and quality measures. No health plan reported that it placed providers (hospitals, physicians, or combinations thereof) in tiers solely on the basis of quality measures.

Table 2
Premium-Tiered and Point-of-Care Tiered Models

Decision Support in CDHB Plans

Table 3 presents the percent of HRA and tiered-benefit models (both premium of point-of-care) that offered decision support in the specific categories we identified. More than 90 percent of HRA enrollees had access to information on the typical cost of procedures and services, while only 13 percent of tiered-benefit model enrollees were offered this information. Comparative cost information on providers—information that would allow a consumer to “shop” across providers—was rarely provided even to HRA enrollees (16 percent and 17 percent of HRA enrollees were offered cost information on doctors and hospitals, respectively). In comparison 20 percent of tiered-benefit model enrollees were offered such information for physicians, while only 13 percent were given cost information to help choose a hospital. Comparative quality information was almost universally provided to HRA enrollees, for both hospitals and some physicians or medical groups. While more than half of tiered-benefit model enrollees received comparative quality information for hospitals, less than 10 percent received this information for physicians. Information to help consumers better manage a chronic condition was provided to only 34 percent of HRA enrollees and 13 percent of tiered-benefit model enrollees. The majority of enrollees in HRA models were offered cost information for selecting prescription drugs, while only 9 percent of enrollees in tiered benefit models had access to these tools. Finally, a nurse-staffed telephone help line was provided to virtually all HRA enrollees and 51 percent of tiered-benefit model enrollees.

Table 3
Decision Support in Consumer-Directed Plans

The enrollment-weighted frequencies of reported decision support mask some differences across subgroups of plans (data not shown). In particular, the few plans that offer only HRA models, in which a large share of HRA enrollment is now concentrated, are more likely to provide most elements of decision support than the mainstream managed care organizations with HRAs.

Consumer-Centered Health Management (and Cost Sharing) in Mainstream MCOs

In Table 4, we report the responses of mainstream HMO, POS, and PPO plans to a series of questions about consumer-centered health management and cost sharing. Nearly all HMO and POS plan enrollees are in plans with proactive outreach programs for members with identified health conditions (most of which are identified through claims data.) Moreover, about half of POS enrollees were in plans that reported using patient reminders for preventive care services. Enrollees in HMOs and PPOs were much less likely to be in plans that reported sending reminders for preventive care (differences significant with a p-value <.01.) By far the most common condition for which reminders were sent was asthma (medication reminder); reminders for hemoglobin A1c testing for members with diabetes, and immunizations were also frequently mentioned (data not shown).

Table 4
Consumer-Centered Health Management and Cost Sharing in Mainstream HMO/POS/PPO Plans

Among HMO and POS enrollees, respectively, 71 percent and 51 percent of enrollees had access to an online health risk assessment tool; fewer (69 percent and 28 percent, respectively) were offered the capability to create an online health profile. Among PPO enrollees, 55 percent were offered an online health risk assessment and could also develop a health profile. Almost half of HMO plans offered consumers incentives to undertake health improvement activities, while only 29 percent and 26 percent of POS and PPO plans respectively did so (pairwise differences between HMO and the other two products are significant with p-values <.01). Because such incentives may include discounted equipment or athletic club memberships, they may be designed for the purpose of attracting healthy enrollees or motivating enrollees with risky health behaviors to change. The majority of all plan types reported that they offered enrollees access to a nurse-staffed telephone advice line or were in the process of developing this capability.

Most mainstream managed care plans reported that cost sharing (copayments, coinsurance, or deductibles) increased in their typical plan in 2003 compared to 2002 and no plans reported decreased consumer cost sharing. Health maintenance organization enrollees were the least likely to face increased cost sharing (65 percent) compared to POS (91 percent) and PPO (78 percent) enrollees. The average increase in estimated annual consumer cost sharing, including deductibles, copayments, and coinsurance, among those plans that reported an increase was just under 5 percent (data not shown).


Consumer-directed health plans have been presented in the press as both a mechanism to shift the locus of decision making from managed care plans to consumers and as a palatable way for employers to reduce or share with enrollees double-digit premium increases. More mainstream managed care plans have also been reported to be developing updated models with increased choice, financial incentives for consumers to choose lower-cost options, and information to support their decisions.

In this study, we report findings from a national health plan survey that included questions designed to measure the uptake of health reimbursement accounts, premium-tiered and point-of-care tiered model plans, and consumer-centered elements of mainstream MCOs. Despite its high response rate (70 percent), the survey may not have captured all consumer-directed health plans. There may have been plans offering HRAs or tiered benefits that were not identified nor contacted by Mercer Human Resource Consulting and nonrespondents may also offer consumer-directed health benefit products. Naturally, this concern is particularly salient for our estimates of total enrollment. To address this concern we made every effort to compare the responses from our survey with other reports of HRA and tiered-benefit models and to ask that experts on consumer-directed health benefits within Mercer Human Resource Consulting identify any important omissions. In several cases, we contacted plans directly to confirm or amend enrollment data.

Another limitation of our approach is that responses to Mercer Human Resource Consulting requests for information are not primarily elicited for research purposes but rather for employer contracting. This accounts no doubt for the relatively high response rate. It might also be expected that health plans would attempt to cast their products in the most favorable light. This tendency, however, would be tempered by the fact that long-term relationships are at stake and exaggerated claims are likely to be detected.

Finally, because a health plan survey was relied upon by the authors and some models may be tailored in their design (including decision support) to meet the needs of particular purchaser segments, reported differences in features among plan types may reflect differences in the purchasers that selected them rather than characteristics of that plan type. For example, large self-insured employers may be more likely than small employers to offer HRAs. At the same time, these employers may typically contract directly for health management programs for all of their employees, so that the plans themselves do not provide such additional services. Health plan survey data cannot address this potential confounding.

The best available estimates of the diffusion of HRA models in 2002 suggested that perhaps 100,000 beneficiaries were then enrolled in these plans, most of whom were signed up with one of three plans specializing in consumer-directed health benefits (Definity, Destiny, and Lumenos) (Gabel, Lo Sasso, and Rice 2002). We estimate that in the first quarter of 2003 there were nearly half a million HRA enrollees. Plans that specialize in offering HRA models still dominate the HRA market, although to a lesser degree than previously reported. Large national managed care organizations have entered into the HRA market and some of the earliest entrants in this class enrolled tens of thousands of beneficiaries in HRA models by early 2003. Many more of these large organizations are launching HRA models in 2004, consistent with reports from the field that most health plans view their ability to offer a consumer-directed plan as a strategic necessity.

While the rate of enrollment growth is substantial, HRA enrollees remain an exceedingly small percentage of the roughly 160 million people with employer-sponsored insurance. If HRA models are to play a major role in changing the dynamics of the U.S. health system—either by encouraging consumerism or in controlling the expenditure trend—more dramatic diffusion will need to occur in the future. Perhaps this will ensue in coming years. Early results from the field suggest roughly a doubling of enrollment in 2004 and recently legislated health savings accounts will further stimulate growth of account-based plans. Nonetheless, projections attributed to industry insiders such as “20 percent of the market by 2005,” are difficult to reconcile with our survey responses (Gabel, Lo Sasso, and Rice 2002).

Our findings support the notion that there is greater marketability of tiered managed care offerings with increased choice (of either benefit design or point-of-care options) accompanied by incentives to choose lower-cost or higher-quality options. Respondents reported a 2003 enrollment of nearly two million covered lives in premium-tiered or point-of-care tiered models. Point-of-care tiered models comprise the majority of this category, accounting for more than three-quarters of the enrollment.

Rather than simply increase cost sharing, consumer-directed health plans are purported to empower individuals to make informed choices with regard to their health and health care. To meet this goal, point-of-care tiered models offering consumers incentives to select a subset of providers or treatment options must also offer information to help consumers decide whether and when selection of higher-cost options is worth the outlay. We found, however, that information to support value-based choices of provider or treatment is not universally provided by HRA models and tiered-benefit products. In particular, comparative cost information for both physicians and hospitals is typically lacking. Consumer-directed heath plans frequently make available hospital quality information, possibly because there are some off-the-shelf products that derive quality information from Medicare and state all-payer administrative data. Average costs for services or procedures and drugs are also common elements of decision support for HRA model plans, perhaps because these are relatively easy for companies to provide, although comparisons of the likely cost implications of alternative types of treatment options beyond drugs for a given condition are typically not available.

The RAND Health Insurance Experiment suggested that consumers (without decision support) rationed necessary care to the same degree as unnecessary care in the face of greater cost sharing. Given this result, it may be a concern that more HRA models are not offering information on optimal care for a chronic condition. This is particularly troubling in light of the fact that just over half of HRA plans reported that they screen all claims against evidence-based practice algorithms to detect underuse and only about one-third of HRA plans notify providers and members of deviations from evidence-based practice (data not shown). On the other hand, perhaps it should not be very surprising that decision support for these products is so incomplete. Such systems entail extensive fixed investments and thus require some scale to support.

Alongside the evolving phenomenon of consumer-directed plans, mainstream MCOs also are sharing more costs with consumers, in order to shift costs, create consumer incentives to spend more prudently, or both. Most plans report percentage increases in cost sharing in the single digits. To a limited degree, MCOs, particularly HMOs, also support the consumer “coproducer” role as well, through nurse help-lines, health risk assessments, and health profiles as well as member outreach. The apparent scramble by large health plans to gain a foothold in the consumer-directed health plan market may support the adoption of additional consumer-centered health management tools because of the economies of scale mentioned previously. That is, rolling out a consumer-directed plan offering with complementary programs and decision support to help consumers manage their health and health spending may spill over onto mainstream health plans because of low or zero incremental costs for extending these programs to enrollees in all types of products.

Health reimbursement accounts and tiered-benefit models viewed together represent the latest vehicles for cost sharing and, potentially, for engaging consumers in stewardship of their health and health benefit costs. What differentiates them from one another is the point in time at which consumers are engaged, the scope of decisions that are targeted, and the degree to which support is provided to inform consumer selections. Health reimbursement account models essentially put consumers fully in charge and at risk for a range of health care decisions until spending reaches the deductible amount, usually about $1,500 per year. Premium-tiered models emphasize consumerism at open enrollment by drawing direct connections between the premium contribution and a variety of plan features including cost sharing and scope of network. Point-of-care tiered models typically engage consumers in making better provider selections, and could be extended to include better treatment option selections particularly for services deemed discretionary. Decision support for all of these models, most importantly for HRAs because of the broad range of choices consumers are expected to manage, does not seem quite up to the task of mobilizing consumers to be successful in making more cost-efficient and health-improving selections. As consumer-directed health benefits grow, it will be of central interest to track the evolution of these decision-support systems and of complementary efforts by plans to monitor underuse and proactively engage both consumers and providers when care falls short of established clinical guidelines.


Financial support for this research was provided by the Robert Wood Johnson Foundation's Changes in Health Care Financing and Organization Initiative. We are grateful for input from our Advisory Panel: Arnold Epstein, Paul Ginsburg, Judith Hibbard, Joseph Newhouse, and Joel Weissman.


1From an incentive perspective, the important question is whether consumers view the account dollars as having the same opportunity cost as out-of-pocket spending. While the rollover provision of most account-based plans would make this more likely, it is unclear whether consumers perceive the account dollars to be fully fungible.


  • Clark NM. “Management of Chronic Disease by Patients” Annual Review of Public Health. 2003;24:289–313. [PubMed]
  • Employee Benefit Research Institute (EBRI), Consumer Health Education Council 2003. Survey April 24–May 24, reported by Roper Center for Public Opinion Research, University of Connecticut. Available at
  • Fronstin P. Can ‘Consumerism’ Slow the Rate of Health Benefit Cost Increases? Washington, DC: Employee Benefit Research Institute; 2002. [PubMed]
  • Gabel JR, LoSasso AT, Rice T. “Consumer-Driven Health Plans: Are They More Than Talk Now?” 2002. [accessed November 2003]. Health Affairs Web exclusives. Available at [PubMed]
  • Galvin R, Milstein A. “Large Employers' New Strategies in Health Care” New England Journal of Medicine. 2002;347(12):939–42. [PubMed]
  • Hibbard JH. “Engaging Health Care Consumers to Improve the Quality of Care” Medical Care. 2003;41(1, supplement):I61–70. [PubMed]
  • Hibbard JH, Peters E. “Supporting Informed Consumer Health Care Decisions: Data Presentation Approaches That Facilitate the Use of Information in Choice” Annual Review of Public Health. 2003;24:413–33. [PubMed]
  • Robinson JC. “Renewed Emphasis on Consumer Cost Sharing in Health Insurance Benefit Design” 2002. [accessed November 2003]. Health Affairs Web exclusives. Available at [PubMed]
  • Scanlon DP, Chernew M, Sheffler S, Fendrick AM. “Health Plan Report Cards: Exploring Differences in Plan Ratings” Joint Commission Journal on Quality Improvement. 1998;24(1):5–20. [PubMed]
  • Schneider EC, Epstein AM. “Use of Public Performance Reports: A Survey of Cardiac Surgery Patients” Journal of the American Medical Association. 1998;279(20):1638–42. [PubMed]
  • von Korff M, Gruman J, Schaefer J, Curry SJ, Wagner EH. “Collaborative Management of Chronic Illness” Annals of Internal Medicine. 1997;127(12):1097–102. [PubMed]

Articles from Health Services Research are provided here courtesy of Health Research & Educational Trust