Background: The aim of this study was to determine whether self-reported depression, coronavirus disease of 2019 (COVID-19) health risk profile, HIV status, and SARS-CoV-2 exposure were associated with the use of COVID-19 prevention measures.

Methods: This survey collected data electronically between June 29 and December 31, 2020 from a convenient sample of 5050 adults 18 years and above living in 12 West African countries. The dependent variables were: social distancing, working remotely, difficulty obtaining face masks and difficulty washing hands often. The independent variables were self-reported depression, having a health risk for COVID-19 (high, moderate and little/no risk), living with HIV and COVID-19 status (SARS-CoV-2 positive tests, having COVID-19 symptoms but not getting tested, having a close friend who tested positive for SARS-CoV-2 and knowing someone who died from COVID-19). Four binary logistic regression models were developed to model the associations between the dependent and independent variables, adjusting for socio-demographic variables (age, gender, educational status, employment status and living status).

Results: There were 2412 (47.8%) male participants and the mean (standard deviation) age was 36.94 (11.47) years. Respondents who reported depression had higher odds of working remotely (AOR: 1.341), and having difficulty obtaining face masks (AOR: 1.923;) and washing hands often (AOR: 1.263). People living with HIV had significantly lower odds of having difficulty washing hands often (AOR: 0.483). Respondents with moderate health risk for COVID-19 had significantly higher odds of social distancing (AOR: 1.144) and those with high health risk had difficulty obtaining face masks (AOR: 1.910). Respondents who had a close friend who tested positive for SARS-CoV-2 (AOR: 1.132) and knew someone who died of COVID-19 (AOR: 1.094) had significantly higher odds of social distancing. Those who tested positive for SARS-CoV-2 had significantly lower odds of social distancing (AOR: 0.629) and working remotely (AOR: 0.713). Those who had symptoms of COVID-19 but did not get tested had significantly lower odds of social distancing (AOR: 0.783) but significantly higher odds of working remotely (AOR: 1.277).

Conclusions: The study signifies a disparity in the access to and use of COVID-19 preventative measures that is allied to the health and COVID-19 status of residents in West Africa. Present findings point to risk compensation behaviours in explaining this outcome.