Effect of Availability of COVID-19 Testing on Choice to Isolate and Socially Distance

Overview

The purpose of this research is to conduct a cross-sectional survey to investigate how people's lifestyle decisions and social distancing choices are affected by the medical information they receive. The hypothesis is that a positive COVID-19 test result will lead to study participants having the greatest self-isolation intentions compared to those who are only clinically diagnosed for COVID-19 without a confirmatory diagnostic test result or those who receive a negative COVID-19 test result.

When people make choices about how to socially distance, they may make different choices when they have information from a doctor or other medical provider versus when they have information from a test result. Are people's lifestyle decisions and choices to socially distance or isolate affected by how information about their health status is communicated? Does the presence of testing change the lifestyle decisions people make? We study and attempt to answer these questions through a cross-sectional, online survey. Two pilot studies will be run prior to the launching the main survey. The preliminary results from these two surveys will be analyzed through EFA (exploratory factor analysis) and power analysis in order to determine subscales, effect size, and appropriate sample size for the main study. Additionally, a focus group consisting of approximately 15 college-educated individuals will be asked to take the main survey in order to evaluate a reasonable time of completion for the survey. The lower bound of this study completion duration will then be utilized to establish reasonable survey completion time and thus create an exclusion criterion based on time to survey completion. Study participants will be first invited to complete a 9 question, pre-test survey on Amazon's Mechanical Turk. Respondents who pass the 4 attention check questions within the pre-test will then be invited to complete the main survey. An estimated 1400 participants will be recruited using Amazon's Mechanical Turk to complete the main survey. Of those, we anticipate 1194 participants to meet all inclusion criteria who will then be included in the main study analysis. Participants will be included in the main survey analysis if they read and agree to the English language consent form, are U.S. residents (based on zip code data), correctly answer the attention check questions in both the pre-test and main survey, and complete the main survey within a reasonable amount of time (deemed to be 120 seconds or more). Participants will be first invited to complete a 9 question, pre-test survey on Amazon's Mechanical Turk. Respondents who pass the 4 attention check questions within the pre-test will then be invited to complete the main survey. After consenting to participate, participants will be randomized to take one of three surveys each describing a different scenario: one where they likely have COVID-19 but testing is not available, one where they likely have COVID-19 and testing results show a positive result, and one where they likely have COVID-19 and testing results show a negative result. Then, participants will be asked questions about their activity and behavior intentions (e.g., stay in a specific room in my home and stay away from all other people and pets, visit a friend or family member in person). Participants will also be asked construct questions based off of Theory of Planned Behavior/Reason Action Approach (located in the pre-test survey), along with a set of demographic questions. Survey responses will be summarized for the full sample, as well as stratified by testing scenario. Quantitative responses will be summarized using means, standard deviations and quartiles, and categorical and ordinal responses will be summarized using frequency distributions. Comparisons between scenarios will be performed using one-way analysis of variance (ANOVA) for quantitative variables, Kruskal-Wallis tests for ordinal variables, and chi-squared or Fisher's exact tests as appropriate for categorical variables. The primary outcome is a difference in the behavioral sum score constructed using 11 items, composed of two subscales. Secondary outcomes include the 'personal decisions' and 'social expectations' subscales respectively. The 'personal decisions' subscale will consist of the items pertaining to masking, self-isolation, visiting friends, purchasing supplies, undertaking physical activity, eating at a restaurant and having dinner at home with friends. The 'social expectations' subscale will consist of the items pertaining to getting a haircut, attending weddings, funerals, and birthday parties. Other secondary outcomes including likelihood of voting, protesting/political gathering, and public transportation 1-item questions will also be analyzed in a similar fashion across the three different arms. The primary hypothesis is that there will be a statistically significant difference in willingness to engage in risky behavior based on COVID test results. This will be evaluated using a linear regression model of the total 11-item score. The primary model term will be scenario, and covariates will include age, sex, race/ethnicity, political affiliation, education level, location, and type of residence. We will perform pairwise comparisons of the 3 scenarios, and use an 0.017 significance level (3-fold Bonferroni correction for an overall alpha of 0.05). Secondary analyses will evaluate the subscales separately using a similar approach. We will also perform exploratory analyses evaluating individual item responses using ordinal logistic regression models with similar specifications. A 5% significance level will be used for all secondary and exploratory hypothesis tests. All analyses will be performed using R v. 3.6.2 (http://www.r-project.org).

Conditions

Interventions

Eligibility

Locations (1)

Outcomes