In this, third post on how bad people are at choosing health plans, I continue to summarize studies relating to commercial market plans, which I started in my second post. (See post 1 for research pertaining to Medicare plans.)
A study by Saurabh Bhargava, George Loewenstein, and Justin Sydnor examined the choices made by over 50,000 workers at a large U.S. firm among 48 health plans in 2010 and 2011. They found that the majority of them made objectively worse choices than they could have, costing themselves an average of $373 per year, or 42% of the average employee premium contribution. Lower income employees and those with chronic conditions were even more likely to make cost-increasing choices.
Apart from cost-sharing and premiums, the 48 plans were identical (e.g., services covered, networks, carrier, manner of presentation in marketing material). Across plans, there were four deductible levels ($1000, $750, $500, $350), three maximum out-of-pocket spending levels ($3000, $2500, $1500), two coinsurance rates (10%/40%, 20%/50% for in/out of network), and two copayment levels ($15/$40, $25/$35 for primary/specialist care). Do the math and that makes 48 possible options (4 × 3 × 2 × 2 = 48).
Because of premium levels, some plans were "dominated" by others, meaning they were objectively worse choices for everyone. For instance, a plan with a $750 deductible cost $500 more in premium than an otherwise identical plan with a $1000 deductible. Trading a guaranteed $500 cost for at most a $250 benefit is an objectively bad deal. (If one's marginal tax rate is at least 50%, it might not be bad, since premiums are excluded from taxation. The study's findings hold up even accounting for taxes.) All but one of the 36 lower deductible plans (<$1000 deductible) offered was dominated in this way by a higher deductible plan.
Though no single plan would have been optimal for every employee, one of the $1000 deductible level plans would have been a better choice (or a no worse choice) for 97% of employees who chose a lower deductible plan and 84% of all employees. Though for some it would not have been the best option, no more than 5% of employees would have lost more than $20 in switching to that high deductible plan. To a large extent, the employer did its employees no favors by offering this particular menu of choices.
On different populations, the authors also conducted controlled experiments based on simulated plan choices. In one experiment, they found that even with a more simplified presentation of plan options than those provided to the employees, 66% still selected dominated plans. In another experiment, they found that 71% couldn't identify basic cost-sharing features and that this type of illiteracy was highly predictive of selecting dominated plans.
Decision aids and smart presentation of options can help consumers make fewer errors. Charlene Wong and colleagues examined their availability on healthcare.gov and the online marketplaces for 12 other states (Maryland and Hawaii excluded). For the second open enrollment period (last year), most marketplaces presented plans in premium order, cheapest to most expensive. Other work suggests that this may cause consumers to overweight the significance of lower premiums.
Similar phenomena have been found in other contexts. After a study showed that papers listed at the top of the National Bureau of Economic Research's weekly email of new working papers garnered more downloads and citations, NBER began randomizing the order across its distribution list. Peter Ubel and colleagues wrote,
Political scientists have shown, for example, that all else being equal, politicians listed at the top of ballots receive more votes than those whose names appear lower on the list. Similarly, people choose different wines if the wine list presents them in order of quality (as judged by experts) than if they're ordered according to price.
Charlene Wong and colleagues also found that only three states offered cost estimators, six offered the ability to search on provider network, and pop-up definitions were offered in nine states. These represented increases from the first open enrollment period, but still, many marketplaces don't offer much assistance to shoppers.
Plan shoppers probably don't even know they need assistance, a necessary condition to protecting themselves from mistakes. George Lowenstein and colleagues conducted a survey to assess privately insured, non-elderly, adult consumers' understanding of private health insurance. Over 90% of respondents expressed a high degree of confidence in their understanding of the concepts of copays, deductibles, and maximum out-of-pocket costs; 57% expressed confidence in understanding coinsurance. But, according to responses to survey questions, people were broadly overconfident in their assessments. Only 78% answered questions about deductibles correctly; only 34% did so for coinsurance, consistent with other work showing people's lack of understanding of their own insurance plans or choices.
Respondents also had difficulty in accurately assessing out-of-pocket costs under a hypothetical insurance plan. For example, about 40% could identify that cost for an MRI scan; only 11% could report the correct cost of a four day hospital stay and only 14% were as close as $1,000. Interestingly, respondents with greater experience with the health care system did not exhibit better comprehension of cost concepts or more accurately estimate out-of-pocket costs.
Some consumers are easily fooled by arbitrary labels. Peter Ubel and colleagues tested hypothetical marketplace-like plan choices in a convenience sample of riders of public buses in Durham, North Carolina. Most participants were drawn to "gold" labeled plans. However, for half the participants, the investigators reversed the meaning of "gold" and "bronze" labels such that the former was of lower premium and higher cost sharing than the latter. It didn't matter.
[A]mong participants who were below the median in mathematical ability, the majority said they preferred gold plans over bronze plans, regardless of which plan was labeled as gold.
Even reporting premiums per month or per week matters.
[P]articipants were significantly less likely to choose the higher-premium, lower-deductible plan when we presented them with monthly premiums than when we cited weekly ones.
A mountain of research documents that consumers can be overwhelmed by a large number of choices, causing them to make decision errors or stick with the status quo; that market opacity can benefit firms, which exploit consumers' decision errors to increase prices and profits; and that provision of more salient and relevant comparison information can improve decisions.
Nevertheless, it is not the case that offering consumers only one or a very small number of choices is necessarily better. The fewer the choices, the less competition and the less ability for consumers to potentially match preferences to products. It's a balancing act. What's certain to help is greater education that errors are common, greater recognition that most of us need help, and more widespread provision of easily accessible, personalized health plan comparison tools and calculators.
Despite how good you think you may be at it, choosing a health plan is very hard. If you've learned nothing else from the literature, learn at least that.
Austin B. Frakt, PhD, is a health economist with the Department of Veterans Affairs and an associate professor at Boston University’s School of Medicine and School of Public Health. He blogs about health economics and policy at The Incidental Economist and tweets at @afrakt. The views expressed in this post are that of the author and do not necessarily reflect the position of the Department of Veterans Affairs or Boston University.