|
FROM THE FIELD(De)constructing ‘Basic’ Benefits: Citizens Define The Limits Of Coverage
Many initiatives for covering the uninsured call for "basic" health care coverage, yet few define that term. The Just Coverage project used a computer-based simulation exercise to learn how nearly 800 community members in northern California identified the inclusions and exclusions that would constitute basic coverage. Working with a limited budget, participants distinguished essential from nonessential health care needs, resisted high patient cost sharing, and tolerated tight restrictions on provider choice. They also supported practice guidelines and standards of effectiveness, and they excluded high-cost, low-value interventions. These results reinforce the importance of community input to policymakers.
Everyone needs some kind of coverage. You know, something to just get by until they hit the lottery. —Uninsured project participant, Sacramento DESPITE HEIGHTENED, broad-based interest, achieving universal health care coverage faces the formidable obstacle of rising health care costs. Communities’ attempts to reduce the number of uninsured residents—use of safety-net clinics, lean benefits, donated services—can be useful as stop-gap measures, but few are likely to survive as long-term strategies.1 Covering the uninsured might not be a realistic pursuit without commitment to affordability. The new universal coverage law in Massachusetts appears to be tackling this indirectly: The coverage that individuals are required to purchase must be "affordable," which legislators indicate should be about two-thirds the amount of the average employer-based premium.2 It is not yet known, however, how health plans are going to design such a low-cost product. Research shows that for many with lower incomes or chronic illnesses, or both, higher deductibles, premiums, and copayments might be counterproductive, making this approach inadvisable for basic coverage.3 Another alternative is limiting the comprehensive coverage model, the foundation of most employer-based insurance. Yet identifying excess benefits is not obvious. If there is a Holy Grail of "basic coverage," few have attempted to define it in any detail.4 Sacramento Healthcare Decisions (SHD) designed the Just Coverage project to solicit public input on establishing the elements of basic coverage. The project asked participants, If everyone were to have health care coverage, what is the minimum they would need?
To explore the dimensions of basic coverage, SHD used a computer-based simulation program called CHAT (Choosing Healthplans All Together) to identify individual and group priorities for coverage within a fixed budget.5 CHAT had been used previously to explore coverage trade-offs with a variety of audiences.6 However, the Just Coverage version was unique in that it structured trade-offs around health care needs (such as chronic illness and episodic care) rather than health care services (such as hospital care or pharmacy). A needs-based approach encouraged participants to assess the basic rationale for insurance coverage beyond simply paying the bills, inducing them to think in terms of "how important is this need relative to other needs?" CHAT offered fourteen categories for consideration, each offering one, two, or three tiers. Higher tiers were more expensive but brought augmented benefits or fewer restrictions. Eleven categories dealt with health care needs, and three related to system features: restrictiveness of provider networks, the individual’s monthly premium share, and copayments. All categories and tiers were displayed on a computer-based pie chart on which participants specified their choices.7 Key questions. Participants had many options for creating this basic plan: increase cost sharing, restrict provider network, eliminate entire categories of health care needs, include stricter coverage criteria within various categories, or exclude coverage of subsets of needs within categories. Key questions included the following: (1) Do participants distinguish between essential and nonessential health care needs, and, if so, what characterizes those distinctions? (2) How do they balance system features (provider network and cost sharing) with other restrictions on coverage? (3) Are there values about coverage that dominate the views of most participants? CHAT process. SHD assigned each of fourteen categories and their respective tiers an actuarially defined number of markers to indicate their cost. Selecting all categories at their highest tiers required seventy-six markers; however, participants had only fifty markers to allocate, an amount representing roughly two-thirds of the average premium paid by employers in California in 2005.8 Using the Health Cost Guidelines, actuaries from Milliman Inc., a national health care consulting firm, helped determine the approximate cost requirements (that is, the number of markers needed) for each category/tier.9 Participants were instructed to develop a coverage plan that represented the "floor below which no one should fall." They were also told that those most likely to use basic coverage would be the 18–20 percent uninsured in California, typically lower-income, working-class individuals and families. This would not apply to those with Medicare, Medicaid, or other public coverage.10 Although basic coverage would be available to everyone, richer plans were possible if individuals or their employers were willing to pay more. Each CHAT group of ten to twelve people met for a two-and-a-half-hour session. This session involved four rounds: (1) On individual laptops, participants created their own version of a basic plan; (2) groups of three worked to gain skill in creating a consensus plan; (3) the whole group worked with the facilitator to create one uniform plan; and (4) individuals created their own plan again. After the first two rounds, a health event "lottery" exposed participants to scenarios that illustrated how the plans they had created would affect those with basic coverage. Throughout the process, participants were asked to make decisions as citizens influencing health policy, not as patients deciding coverage for themselves.
Participants.
The project involved a nonrandom sampling of volunteers from community organizations, workplaces, college classes, and religious congregations in eight northern California counties. To encourage broader representation, some people (for example, lower-income, less-educated, or uninsured) received stipends. Following institutional review board (IRB) approval, 71 sessions were held with 798 participants from June 2005 to April 2006; 52 percent of participants were in Sacramento County. Compared with the general population, participants were overrepresented by women, Caucasians, and people with health insurance, higher education, and higher incomes (Exhibit 1
SHD collected all computer data anonymously, including pre- and post-CHAT survey questions. Group discussions in round 3 were tape-recorded, transcribed, and analyzed to identify the themes and values that underlie individual and group decisions, a process that was critical to understanding the rationale for trade-offs. Although round 3 results show the decisions of entire groups (voting was often required when consensus was elusive), the round 4 results reflect individuals’ final decisions. Both sets of responses are instructive. Data were analyzed to identify group and individual benefit selections. Bivariate analyses examined the association between coverage choices and participants’ characteristics, particularly sex, race/ethnicity, income, education, and insurance status.11
We present here selected findings that are relevant to current health policy; comprehensive results are available elsewhere.12
Finding 1: Not all needs are considered equal.
Participants put the highest priority on coverage for responding to life-threatening situations; preventing or delaying illness, disease, or disability; and enabling or restoring vital capabilities (Exhibit 2
Defined as "problems in function, appearance, or comfort that are not seriously disabling but impact personal quality of life," QOL is a catch-all for medical situations that are meaningful for individuals but have little impact on the most essential activities of daily living. Typical participant comments were as follows: "It’s nice to have, but it doesn’t affect your ability to go to work," "It’s not going to shorten your lifespan," and "Why is it that all of us should pay for your ability to run a marathon if in fact the rest of your mobility functions in life are normal?" The QOL category was included by only 15 percent of the groups during the round 3 discussion. When participants made individual decisions in round 4, however, the QOL inclusion increased to 40 percent. Said one QOL advocate, "How many people can afford an infertility work-up? We owe them a chance." Support of the QOL category was particularly prevalent among lower-income participants, of whom 60 percent selected the QOL category in the final round (odds ratio = 2.31, 95 percent confidence interval = 1.21, 4.40); African American and Hispanic participants’ choices were similar—58.7 percent and 59.2 percent, respectively.13 Participants with a high school education or less were three times more likely than those with higher education to select QOL (OR = 3.14, 95% CI = 1.30, 7.58). There was no statistically significant difference in selection among any of the other demographic subgroups, including the uninsured. If exclusion of the QOL category is validated with participation of more diverse groups, this might suggest a tighter standard for distinguishing between medically essential and medically desirable treatment. But if additional inquiry reinforces a striking difference in how certain demographic groups perceive the relative importance of this category, it raises an essential question about the use of communal resources: Whose opinion counts?
Finding 2: Tighter coverage criteria are acceptable.
Participants could also reduce rigid coverage criteria by choosing higher (and more costly) tiers in the categories of prevention, maintenance, catastrophic, maternity, and end-of-life care. Most groups and individuals did not do so, as reflected in the "not covered" descriptions in Exhibit 2 One example was catastrophic care, where only 4 percent of the groups in round 3 included tier 2, which covered "treatments that are not likely to help but are the ‘only hope left’." When individuals decided for themselves in round 4, the inclusion of tier 2 increased to 16 percent, and among those with a high school education or less, it increased to 35 percent (OR = 3.12; 95% CI = 1.25, 7.75). Variance among other subgroups showed that those with lower incomes and who were uninsured also had a stronger preference to include tier 2 (26 percent and 27 percent, respectively). However, these findings were not statistically significant because of the small sample size. Some health plans and medical groups aspire to the high standards of effectiveness as described in several of these categories. However, physicians and patients have considerable leeway in deciding for themselves when an intervention is medically necessary and what treatment is preferable. Implementing these results in a basic coverage plan would require a reduction in physicians’ and patients’ autonomy in determining the interventions that are paid for by insurance.
Finding 3: A restrictive provider network is tolerable.
Participants had to weigh three features of the delivery system (Exhibit 3
SHD based its description of tier 1 providers on a restricted model of health care delivery in which there is no excess physician capacity, referrals to specialists are tightly controlled, and there is extensive use of cost-efficient alternative providers (such as nurse practitioners). Research by Milliman indicates that such an approach greatly reduces costs compared with delivery systems that lack these measures.14 This assessment is reinforced by studies indicating that high use of specialists increases health care spending without improving the quality of care.15 The tier 1 providers description was intentionally stark to expose participants to the high cost of open-ended, loosely managed provider networks. Although few participants had experienced the specter of such limits, many viewed it with alarm. With reluctance, 63 percent of the groups choose this tier, allowing them to keep cost sharing reasonable or to increase coverage of other categories. During discussion, participants said that tier 1 was tolerable if the quality of providers’ care was acceptable, if one could change doctors when needed (albeit with limited selection), and if specialty care wasn’t completely inaccessible. Among the comments were the following: "To some degree if you’re on this type of program you have to accept that you’re going to make those kinds of concessions." "I think it’s kind of odd that people are willing to spend the extra markers to have a choice of doctor but they won’t let you have your knee replaced in Complex Chronic. I can see a problem with that balance." When individuals created their final plan in round 4, 54 percent still chose tier 1 for provider network. Three demographic subgroups (lower income, lower education, and African American) were in the 45–47 percent range for choosing tier 1, but the differences did not show statistical significance. All other subgroups were in the same range, around 54 percent, as the total results for round 4. There was relatively little debate about the "copayments" and "premiums" categories. Participants thought that those needing basic coverage are more likely than not to have lower-than-average incomes. To keep coverage affordable and services accessible, most groups choose the mid-range for copayments and premiums. Although debate was minimal, efforts during round 3 to increase patient cost sharing (to use the markers for providers or other categories) were met with resistance: Most participants regarded these two categories as nonnegotiable. Round 4 results showed little variation among individuals.
Dealing with different values. Despite the limits, participants reacted favorably to the basic plans they had designed. In response to post-CHAT survey questions exploring participants’ attitudes, 82 percent thought that the plan they had developed adequately represented basic coverage, and 85 percent thought that they would definitely or probably find this coverage acceptable if they needed it themselves. When these two questions were analyzed by demographic groups, there were no meaningful differences in response rates by income, education, ethnicity, sex, or age. Yet some of the differences previously noted between round 3 and round 4 results raise the issue of how coverage decisions are made. One could argue that round 3, when participants must work through their differences in how to allocate markers to various categories and tiers, is the only legitimate way to make decisions involving communal funds. Others might insist that it is only the individual perspective at the conclusion that really matters, when people are free to express their own preferences unhindered by group pressure. Differences in the views of demographic groups also raise the question of whose opinion has more validity: the group made up of higher-income citizens whose taxes will likely support basic coverage, or the people most likely to use the coverage? Although the QOL category was the only one with markedly varied results, it illustrates differences in values and priorities that cannot be ignored. Creating a specific plan. A greater challenge is creating a specific benefit plan based on the Just Coverage priorities. Participants rejected a model that would have been relatively simple to implement: increasing patient cost sharing and eliminating coverage of several easily segmented categories such as maternity or mental health care.16 Instead, they clearly preferred the most comprehensive coverage possible with reasonable cost sharing. These values drove much of the group discussions. The resulting plan imposes high standards on physicians’ accountability for efficient and effective care and on patients’ acceptance of reduced treatment options and provider choice. Although implementing these conditions would face barriers, the rising cost of health care might require a radical reconsideration of the principles of coverage. Limitations. Because the project focused on health care needs, not services, actuarial estimates of the costs were less precise than in a service-oriented range of options. Since participants were heavily influenced by the number of markers required for each category or tier, even small errors in cost assignment might have influenced their choices. Also, although most responded that the plan they had developed was acceptable as "basic," we do not know, without further testing, how acceptable more or less restrictive plans would be. Particularly limiting was the small number of participants who would be likely recipients of basic coverage. Had there been more participation by the uninsured, those with less education, and Hispanics (who constitute a disproportionate percentage of California’s uninsured population), SHD could have reported more definitively the preferences and values of those most likely to be using basic coverage. Just Coverage dealt with certain trade-offs but not with all possible ways to reduce health care costs. Other cost drivers (such as the market-driven prices of drugs and devices, administrative costs, profit margins, and unnecessary duplication of services) represent formidable obstacles to an efficient system. If strategies to address these cost drivers were implemented and effective, there is no doubt that consumers would prefer them to any actions restricting coverage. DECIDING WHAT NOT TO COVER is vexing for policymakers trying to craft a less costly but socially responsible benefit package. Yet they rarely ask the public to assist with this task in a substantive way. Just Coverage was one process for eliciting public input with a focus on explicit, tangible trade-offs. The recent passage of Massachusetts’ universal health care plan—as well as programs being implemented in other states and communities—gives particular relevance to this process of citizen engagement. Although the views of these 798 community members are by no means conclusive, they provide a starting point for discussions about compromises and trade-offs. These conversations are long overdue.
Marjorie Ginsburg (ginsburg{at}sacdecisions.org) is executive director of Sacramento Healthcare Decisions in Rancho Cordova, California. Susan Goold is director of the Bioethics Program at the University of Michigan School of Medicine in Ann Arbor. Marion Danis is head, Section on Ethics and Health Policy, at the National Institutes of Health in Bethesda, Maryland. The Just Coverage project was supported by a grant from the California HealthCare Foundation. The conclusions presented here are those of the authors and do not necessarily reflect those of the funders, the National Institutes of Health, or the U.S. Department of Health and Human Services. The authors thank Howard Slyter, Jill Yegian, Michael Johnson, Carol Parise, Joanne Garrett, and, in particular, Kathy Glasmire for their insightful comments and suggestions.
| |||||||||||||||||||||||||||||||||||||||
Top
Study Methods
Study Results
