The effect of nudges on health insurance take-up in the United States

Nudges targeting a large array of barriers to uninsured populations increased health insurance take-up.
Interventions tested across eight randomized evaluations [1] [2] [3] [4] [5] [6] [7] [8] targeted a range of barriers to take-up, including lack of awareness or forgetfulness, complexity, low salience of costs and benefits, choice overload, procrastination, and preference for assistance in languages other than English. The interventions reviewed were generally able to reduce such barriers for populations eligible for Medicaid and those eligible for subsidized plans on the individual marketplace. Most interventions increased enrollment within the range of 0.3 to 14.8 percentage points, with an average impact of 5.1 percentage points and a median impact of 2.3 percentage points.

Nudges can be particularly effective at encouraging take-up among low-income populations. 
Three studies [3] [4] [7] found that low-income individuals were more likely to enroll in health insurance or remain insured after the intervention than higher-income populations. In California, letters that made subsidy and plan comparison information more accessible [4] and personalized phone calls [7] increased enrollment among low-income individuals compared to higher-income individuals. In a national study [1] researchers observed larger effects of the intervention on the take-up of insurance among individuals whose household income fell below the Medicaid threshold across all states and a larger effect in states that expanded Medicaid than in those that did not. 

Further, one study [4] found that for households with higher incomes (above 250 percent of the federal poverty level), providing subsidy information was counterproductive relative to basic reminders. Individuals with low subsidies may misunderstand or undervalue the benefits of insurance in a way that reduces take-up.  Including information about these benefits may be needed to overcome this misperception. 

Nudges are also generally effective in encouraging switching among consumers who already have insurance but are not enrolled in the lowest-premium plan. 
Barriers to switching include comparison friction, or the gap between the availability of information about different options and the consumer’s use of it to make comparisons, and choice error, particularly choosing a higher-cost plan when one is eligible for a lower-cost option that has similar or better insurance properties. To reduce potential choice error, consumers who were eligible for plans with higher subsidies but were enrolled in other plans were sent information regarding the higher-subsidy plans [9]. This information increased plan switching by 2 to 4 percentage points, for email-only and mail-plus-email interventions respectively (an 11 to 22 percent increase from a baseline of 18 percent). Letter recipients saved, on average, roughly $65 per year. Plan switchers saved, on average, $84 per month in premiums, or over $1000 annually, and $56 per month in reduced out-of-pocket expenses, or over $670 annually. 

In the context of comparison frictions in prescription-drug insurance, researchers sent personalized price information created by entering individuals’ prescription use data into Medicare's Plan Finder website. While this information was publicly available and widely advertised, the additional step of providing the information had an impact on plan switching, increasing switching by 11 percentage points (a 65 percent increase from a baseline of 17 percent) [10]. Letter recipients saved, on average, $100 per year. Plan switchers saved, on average, roughly $365 per year. 

However, a study showing no impacts raises questions on the effectiveness of nudges on inertia, in which consumers are reluctant to change plans after market offerings change and they might benefit from switching [11]. In this study, researchers provided consumers who would be re-enrolled in their current individual marketplace plans with personalized information about the savings on insurance premiums that they could realize from switching plans and encouraged them to shop for new plans. While this intervention increased shopping on the marketplace, it had no impact on consumers switching to lower-cost plans. Notably, this study did not include the lowest income, and therefore potentially most cost-sensitive, population. Because defaults are another common behavioral intervention, this finding indicates that more research on the effectiveness of defaulting individuals into their plans is needed, as well as a better understanding of whether individuals are intentionally choosing higher-cost plans. 

For both health insurance take-up and plan switching, nudges—particularly low-intensity, non-personalized efforts—are cost-effective.
Low-intensity efforts such as letters and emails cost an average of $0.78, with costs of the intervention ranging from $0 for emails to $0.55 to $1.75 for letters [1] [3] [4] [5] [8] [9]. The average cost per member enrolled for low-intensity interventions was approximately $60.48, ranging from $0 for emails to $12 to $191 for letters. More-intensive interventions (i.e. personalized phone call outreach), ranged in cost from about $4 [7] to $28 [3].  The cost per member enrolled for more intensive interventions averaged around $207. 

Because the marketplace receives a percentage of premiums from insurance companies, each member enrolled through the marketplace also leads to increased revenue for the marketplace. One study [7] calculated that the intervention had a two-to-one return on investment, making this intervention potentially profitable for state-wide marketplaces. 

Including a personalized component within the nudges may increase their effectiveness.  
Of the studies reviewed, eight out of eleven [1] [3] [4] [6] [7] [8] [10] [11] incorporated a personalized component ranging from a personalized estimate of a penalty, personalized subsidy or after-subsidy premium cost, personalized enrollment assistance, one-on-one phone conversation, guidance on plan selection, and case management.

Two studies [7] [10] evaluated the effects of personalized interventions compared to a control group, while six studies  [1] [3] [4] [6] [8] [11] included comparisons on the effects of non-personalized nudges to personalized ones. Additional personalized information included personalized telephone outreach [3] [6] or information on estimated subsidy, penalty, and monthly premium [1] [4] [8] [11].

Three of the six studies that incorporated a personalized component indicated that customized information, such as the inclusion of personalized penalty or subsidy estimates or the addition of personalized phone calls, increased enrollment compared to a generic letter [1] [6] [8]. The other three studies did not show a clear impact of personalized components on enrollment or switching rates [3] [4] [11] 

While including personalization may increase effectiveness, the positive impact of non-personalized interventions, such as handing out Medicaid/CHIP applications in the emergency department [2] or targeting a specific population but not personalizing on an individual level [9], demonstrates that low-intensity interventions can still lead to increased enrollment and plan switching. 

In addition, one study found that action language (such as “You’re almost done”) was more effective than generic messaging for individuals that started the process of enrolling [5].

While most nudges had a positive effect on enrollment and plan switching, the impact was often small. Given the potential of increased health care coverage to lead to beneficial health outcomes, more research is needed to understand what mechanisms can further increase take-up and switching.
Nudges can reduce some—but not all—barriers to enrollment for individuals who can afford coverage. It should also be noted that nudge-based interventions also do not address the barriers for millions of individuals in the United States who cannot afford insurance on the Marketplace and may not be eligible for Medicaid.

More research is needed on the impact of other, non-nudge interventions. For instance, comprehensive case management,³ where the counselors not only inform the consumers and provide information but help push the applications through the Medicaid office, could be more effective for populations facing the most barriers to take-up or selecting the lowest-cost plans. Simplified enrollment, such as “check the box” mechanisms [8], and automatic enrollment,⁴ where individuals must take an action to unenroll (rather than taking an action to enroll), may also be highly effective in certain settings at increasing enrollment by not requiring individuals to overcome common barriers to health insurance enrollment. Finally, more research on structural changes, such as increasing plan options, efforts to increase healthy behaviors, and impacts of insurance on health care utilization and health outcomes, continue to be needed.