Issue: Policymakers have sought to improve the shopping experience on the Affordable Care Act’s marketplaces by offering decision support tools that help consumers better understand and compare their health plan options. Cost estimators are one such tool. They are designed to provide consumers a personalized estimate of the total cost—premium, minus subsidy, plus cost-sharing—of their coverage options. Cost estimators were available in most states by the start of the fourth open enrollment period. Goal: To understand the experiences of marketplaces that offer a total cost estimator and the interests and concerns of policymakers from states that are not using them. Methods: Structured interviews with marketplace officials, consumer enrollment assisters, technology vendors, and subject matter experts; analysis of the total cost estimators available on the marketplaces as of October 2016. Key findings and conclusions: Informants strongly supported marketplace adoption of a total cost estimator. Marketplaces that offer an estimator faced a range of design choices and varied significantly in their approaches to resolving them. Interviews suggested a clear need for additional consumer testing and data analysis of tool usage and for sustained outreach to enrollment assisters to encourage greater use of the estimators.
The Affordable Care Act’s (ACA) marketplaces were designed to make it easier for people to view their health insurance options in one place, compare key features of plans, and choose coverage that is best for them. Three years after their launch, the marketplaces’ record is mixed. The uninsured rate is at an historic low; financial assistance, available exclusively through the marketplaces, has encouraged millions to sign up for insurance; and most marketplace enrollees say they are satisfied with their coverage and are using it to obtain care they would not have been able to access otherwise.1 Still, for many consumers, shopping for a suitable health plan remains daunting; individuals who are not eligible for a subsidized premium have shown relatively little interest in using the marketplaces to find a plan; and the costs of coverage persist as a major barrier to care.2 For these reasons and others, marketplace enrollment has lagged initial expectations.3
One area marketplaces have sought to improve is the shopping experience. Health insurance is complicated; many people have difficulty understanding and applying key insurance concepts.4 The complexities are compounded when individuals must weigh the features of competing plans—a task most people dread and do not do well.5 Consumers who are confused or overwhelmed by their options may be discouraged from enrolling, while those who choose poorly are likely to face higher costs or be unhappy with their plans.
Consumers are especially interested in—although often confused by—a health plan’s cost. People may understand differences in plan premiums, but evidence shows many have a hard time evaluating out-of-pocket spending, a calculation that requires consumers to understand plans’ cost-sharing structures and predict the amount and kind of care their household is likely to use.6 Consequently, consumers often overweight premium relative to other spending, leading to suboptimal plan choices.7
To support better enrollment decisions, most marketplace websites offer a tool to help consumers estimate the total cost—premium, minus subsidy, plus expected cost-sharing—of their coverage options. Total cost estimators are available on HealthCare.gov and in eight state-run marketplaces, reaching consumers in the vast majority of states (Exhibit 1).8 We interviewed more than 40 marketplace officials, consumer assisters, technology vendors, and other subject matter experts to see how having an estimator can change consumers’ experiences in the marketplace and to understand the concerns of policymakers as they have considered whether and how to implement these tools.
Informants Agree: Marketplaces Should Have Total Cost Estimators
— Vermont Marketplace
The policymakers we interviewed strongly suggested that marketplaces should have total cost estimators—this was true in states with and without such tools. Several marketplace officials said estimators were a “critical” component of the enrollment process, providing consumers a more holistic view of health care costs that prompted them to look beyond premiums and consider actual utilization. Policymakers noted that after the first year of enrollment, it was apparent that consumers needed a quick, simple way to determine costs and decipher plan options. For many informants, the tool helped fill this void by giving consumers a snapshot of costs and familiarizing them with their choices before they started the application process. In addition to improving health insurance literacy and driving informed decision-making, policymakers said the tool offered other benefits, like strengthening customer service and equipping broker and assister communities with potential time-saving processes. Marketplace informants say that feedback from consumers and enrollment stakeholders has been overwhelmingly positive.
— Colorado Marketplace
Even in marketplaces without an estimator, policymakers have long noted the value of having such a tool. However, officials have prioritized other implementation issues and, in some instances, were set back in their efforts by IT systems failures during the first open enrollment period. Nevertheless, informants suggested ongoing interest in developing an estimator. Several states suggested starting simple, arguing even a basic model could do much to help consumers understand the value of different plans and could be enhanced over time. Informants cautioned, however, that poorly designed decision aids could add to consumer confusion or undermine decision-making and emphasized that even a simple tool required significant planning.
Exploring Design Choices: The Devil in the Details
Who Builds the Tool and with What Data?
Marketplaces that chose to go forward with an estimator first had to determine who would build it. Most relied on a vendor, in part because of limited staff capacity. Three marketplaces—Minnesota, the District of Columbia, and Vermont—selected an outside vendor that specialized in consumer decision support and that offered an off-the-shelf option. Three others leveraged existing vendor relationships. Kentucky, for example, tapped its longstanding marketplace IT developer so that it could launch the tool as quickly as possible.9 All marketplaces sought solutions that would integrate well with existing IT infrastructure and each said they were satisfied with their choices. Several marketplaces, including Connecticut, expressed a preference for a tool that would use state-specific data—so cost estimates would more closely align with the experiences of the local population—and chose to build in-house. Though most states currently use estimators that rely on national databases, many marketplace informants expressed interest in incorporating state-specific data in the future.
How Much Detail to Provide?
Marketplace informants reported trying to balance how much detail to include in their estimators—in terms of the number and complexity of inputs and outputs and the information provided to explain them—with a goal of making the tool helpful but not overwhelming. Approaches have varied. The federal platform has the simplest and perhaps most-streamlined process. Its output is driven by a single question that asks consumers to estimate their expected medical care (low, medium, or high). Other marketplaces have incorporated additional inputs in hopes of offering consumers a more personalized result (Exhibit 2).
— Colorado Marketplace
Marketplaces weighed several factors when determining how many and which inputs to use. Informants in Connecticut, Minnesota, and Vermont said they carefully considered how long it would take for a consumer to use the estimator and therefore strove for brevity. Many marketplaces were reluctant to use numerous inputs out of concerns, according to one informant, that “granular questions might give consumers a false sense of accuracy” and said they emphasized to users that estimators produced just that—estimates. At the same time, many informants suggested it was worthwhile to aim for at least a moderately personalized estimate and suggested inputs could be increased if marketplaces were thoughtful about how they explained them—for example, by providing information via hover boxes or links where consumers could elect to learn more.
Tool inputs also depended on designers’ methodological approach to calculating cost estimates. Some marketplaces give significant weight to anticipated utilization, asking consumers to estimate how often they will visit the doctor or the approximate number of prescription drugs they will take. Others, including three marketplaces using similar tools developed by the same vendor, provide consumers with estimates based primarily on an actuarial analysis of costs incurred by people similar to them. This tool relies on inputs for age and health status. However, it also allows consumers to indicate if they expect to undergo one of several enumerated medical procedures.
Informants agreed that regardless of the underlying design, estimators should receive prominent placement on marketplace websites to encourage usage. Stakeholder feedback prompted several marketplaces to look for opportunities to better integrate and display the tool for the fourth open enrollment period.
Engaging Stakeholders Before and After Implementation Is Critical to Take-Up
Marketplaces varied in the degree to which they engaged consumers and assisters prior to implementing the tool. While a few reported developing the tool through a series of interactions with focus groups and after “painstaking” attention to the wording of questions, others said they pursued an expedited process that allowed for demonstrations only after the estimator was complete. National advocates and vendors emphasized that consumer testing is an essential component when developing a tool. In practice, however, this did not always occur.
— DC Marketplace
Some of the assisters we interviewed used cost estimators regularly; others almost never. Take-up frequently appears to have been strongly influenced by first impressions formed shortly after a tool was released. If the initial experience was rocky, assisters sometimes did not give the estimator another try. Assisters described a range of barriers to utilization, including time pressure, insufficient training, a reluctance to change established ways of providing enrollment assistance, and concerns about accuracy. To overcome these challenges, many assisters recommended that marketplace officials do more to highlight the estimators within the enrollment community, for example, by clearly articulating the value of the tools, describing how to use them, and soliciting feedback from assisters. Today, the majority of marketplaces with total cost estimators are evaluating their tools using utilization data, surveys, and observational testing. Some marketplaces, like Minnesota and the District of Columbia, reported that tool promotion and training have been key factors in increasing utilization. Other marketplaces recognize that more can be done to engage stakeholders and dispel early misconceptions.
The marketplaces have made finding coverage easier, while the ACA’s consumer protections have ensured that people cannot be denied coverage or charged a higher premium because of their health status. Still, the goal of a consumer-friendly shopping experience, where individuals are empowered to sift through competing plans and easily identify the one best suited to their needs, is as yet largely unrealized.
Delivering on this vision is difficult because health insurance is complicated. But there is reason to believe that improving the design of and support provided by marketplace websites could yield real benefits. A simpler shopping experience, enhanced by tools that allow consumers to compare critical information about plans, may generate higher enrollment by reducing confusion and attrition during the sign-up process. Increasing the value proposition of the marketplaces may encourage new enrollment from consumers—particularly from those not eligible for subsidies—who have so far seen little reason to use them. More fundamentally, decision support tools are worthwhile if they help consumers find less costly and potentially more satisfying plans. For these reasons, the marketplace officials we spoke with saw particular value in providing consumers a cost estimator tool.
Interviews with informants reveal several important issues facing policymakers as they consider how best to implement total cost estimators. First, marketplaces must strive to provide an estimator that is neither so detailed as to be overwhelming nor so simplistic as to be largely without value. Informants involved in the design process said their goal was a tool capable of generating personalized estimates of plan costs. But they were cognizant that an estimator with too many inputs or outputs that are poorly explained might create confusion rather than mitigate it, or be perceived by consumers as too complicated or time-intensive to use. Informants said, however, that the difficulty of striking a design balance was not a reason to avoid using an estimator. Rather, many emphasized the importance of consumer testing and data analysis—in particular, collection and study of marketplace website usage patterns—to inform policymakers about how estimators are being used. In practice, however, the degree to which marketplaces have undertaken these activities has varied widely, and all marketplaces would benefit from more rigorous solicitation of feedback and data analysis going forward.
Second, additional testing is needed to evaluate these tools’ accuracy. Several informants stressed that cost estimators are not intended to produce an exact measure of costs, but instead offer value by encouraging consumers to think more broadly about expenses and by facilitating comparisons between plans. Indeed, some expressed real concern about giving consumers a false impression of precision, given that actual and estimated costs are likely to diverge. No one suggested the dollar value of the estimate is immaterial; it can anchor consumer expectations and be incorporated into household budgeting decisions. Going forward, it will be important for marketplaces to assess the accuracy and efficacy of these tools.
Finally, while nearly all informants saw at least theoretical benefits in a total cost estimator, a significant number of assisters interviewed were not regularly using the tools.10 Some were unsure how the estimators work; others were skeptical of their value in practice and viewed them as extraneous to established routines for providing assistance. Opinions regarding the tools often appear to have been formed quickly, following their launch, and in some cases may have been colored by marketplaces’ initial technology troubles. To overcome these barriers, marketplaces may need to do more to engage assisters and to make a clear case for using these decision aids.
1 According to the Gallup-Healthways Well-Being Index, the uninsured rate in the United States reached a historic low of 11 percent during the second quarter of 2016, down 6.1 percentage points from the level measured immediately before ACA’s largest reforms to the insurance markets took effect. S. Marken, U.S. Uninsured Rate Remains at Historical Low of 11.0% (Gallup, July 11, 2016). By March 31, 2016, there were about 11.1 million consumers with active coverage through the marketplaces, about 85 percent of whom were receiving premium tax credits. U.S. Department of Health and Human Services, March 31, 2016 Effectuated Enrollment Snapshot (HHS, June 30, 2016). When asked to rate their coverage, 66 percent of marketplace enrollees surveyed by The Commonwealth Fund reported that coverage was good, very good, or excellent. S. R. Collins, M. Z. Gunja, M. M. Doty, and S. Beutel, Americans’ Experiences with ACA Marketplace and Medicaid Coverage: Access to Care and Satisfaction (The Commonwealth Fund, May 2016). Moreover, according to three national surveys, most Qualified Health Plan (QHP) enrollees obtaining coverage through the marketplaces reported overall satisfaction with their plans in 2014–2016. Government Accountability Office, Most Enrollees Reported Satisfaction with Their Health Plans, Although Some Concerns Exist (GAO, Sept. 2016).
2 See, e.g., M. Z. Gunja, S. R. Collins, M. M. Doty, and S. Beutel, Americans’ Experiences with ACA Marketplace Coverage: Affordability and Provider Network Satisfaction (The Commonwealth Fund, July 2016); see also R. Garfield, A. Damico, C. Cox et al., New Estimates of Eligibility for ACA Coverage Among the Uninsured (Henry J. Kaiser Family Foundation, Jan. 2016).
3 For an assessment of factors likely responsible for lower-than-anticipated marketplace enrollment, see, e.g., L. Levitt, G. Claxton, A. Damico et al., Assessing ACA Marketplace Enrollment (Henry J. Kaiser Family Foundation, March 2016).
4 For a helpful overview, which includes a scan of recent literature on the topic, see Z. Parragh and D. Okrent, Health Literacy and Health Insurance Literacy: Do Consumers Know What They Are Buying? (Alliance for Health Reform, Jan. 2015); see also L. Blumberg, S. Long, G. Kenney et al., Public Understanding of Basic Health Insurance Concepts on the Eve of Health Reform (Urban Institute, Dec. 2013).
5 For a recent overview of the research, see A. Frakt, “Why Consumers Often Err in Choosing Health Plans,” New York Times, Nov. 1, 2015; and L. Quincy, What’s Behind the Door: Consumers’ Difficulties Selecting Health Plans (Consumers Union, Jan. 2012).
6 Consumer sensitivity to premiums may be heightened in the ACA context by the common marketplace practice of presenting plans in premium order, as well as by the fact that the amount of financial assistance a consumer receives is tied to plan premiums. See, e.g., J. Holahan, L. Blumberg, and E. Wengle, Marketplace Plan Choice: How Important Is Price? An Analysis of Experiences in Five States (Robert Wood Johnson Foundation, March 2016).
7 See, e.g., G. Lowenstein, J. Y. Friedman, B. McGill et al., “Consumers’ Misunderstanding of Health Insurance,” Journal of Health Economics, Sept. 2013 32(5):850–62.
8 For analyses of the decision support tools available on the marketplaces during the third open enrollment period, see C. A. Wong, D. E. Polsky, A. T. Jones et al., “For Third Enrollment Period, Marketplaces Expand Decision Support Tools to Assist Consumers,” Health Affairs, April 2016 35(4):680–87; and National Partnership for Women & Families, Supporting Informed Decision-Making in the Health Insurance Marketplace: A Progress Report for 2016 (NPWF, May 2016).
9 At the time of analysis, Kentucky was using its own marketplace platform and offered consumers a state-designed total cost estimator. The state transitioned to the federal marketplace platform, HealthCare.gov, and the federally designed total cost estimator, at the start of the fourth open enrollment period in November 2016.
10 A survey of enrollment assisters conducted in April 2016 noted, similarly, that while a majority of assisters found total cost estimators to be useful, the tools had not yet been widely used. C. Fish-Parcham, Improve the Display of Plan Information on Marketplace Websites to Help Enrollment (Families USA, July 2016).
The authors thank the many marketplace officials, consumer assisters, agents and brokers, and policy experts who shared their time and valuable insights with us. Thanks also to Kevin Lucia, JoAnn Volk, David Machledt, Cheryl Parcham, and Sophie Stern for their thoughtful review and comments.