OBJECTIVES : To determine whether patients at higher risk for SARS-CoV-2-related complications had systematically different preferences pertaining to COVID-19 risk, social-distancing restrictions, and related economic impact.

METHODS : We designed a discrete-choice experiment to quantify Americans’ preferences across four attributes: the cumulative risk of contracting COVID-19 through 2020 (2%, 5%, 10%, 20%); when nonessential businesses could reopen (now, 2, 5 months); when the economy would recover to pre-COVID-19 levels (2, 3, 5 years); and the percentage of US households falling below the poverty threshold (16%, 20%, 25%). Respondents were recruited by SurveyHealthcareGlobus in May 2020. Latent-class analysis was used to model heterogeneity in preferences.

RESULTS : 5953 adults completed the survey. Thirty-five percent were <40 years, 36% were 40-59, and 29% were 60 or older; 67% were female; 23% were non-white; 34% percent had no comorbidities, 27% had one, and 40% had 2+ comorbidities. Although, on average, COVID-19 risk was the most important factor, four distinct classes emerged. Respondents with 1 or 2+ comorbidities (versus none) were more likely to be in COVID-19 “risk-minimizer” class (both p<0.05). Compared to ages 40-59, both older and younger respondents were more likely to be risk minimizers. A second class had stronger preference to delay reopening non-essential businesses. Younger respondents were more likely (p<0.01) while older respondents were less likely (p<0.01) to be in this class. Respondents with 2+ comorbidities (vs. none) were also more likely (p=0.05). Neither age nor comorbidities were predictive of being in a class prioritizing economic recovery. The fourth class strongly preferred to reopen nonessential businesses sooner. The likelihood of class membership was lower for respondents with 2+ comorbidities (p<0.01) and for those <40 versus those ages 40-59 (p<0.01).

CONCLUSIONS : Americans with comorbidities appeared to heed public-health messaging, generally making choices that prioritized risk reduction over economic outcomes. Respondents’ age was less consistently associated with COVID-19-related preferences.