Abstract
BACKGROUND: Biological sample collection and data linkage can expand the utility of population health surveys. The present study investigates factors associated with population health survey respondents' willingness to provide biological samples and personal health information. METHODS: Using data from the 2019 Centre for Addiction and Mental Health (CAMH) Monitor survey (n = 2,827), we examined participants' willingness to provide blood samples, saliva samples, probabilistic linkage, and direct linkage with personal health information. Associations of willingness to provide such information with socio-demographic, substance use, and mental health details were also examined. Question order effects were tested using a randomized trial. RESULTS: The proportion of respondents willing to provide blood samples, saliva samples, probabilistic linkage, and direct linkage with personal health information were 19.9%, 36.2%, 82.1%, and 17%, respectively. Willingness significantly varied by age, race, employment, non-medical prescription opioid use (past year), cocaine use (lifetime), and psychological distress. Significant question order effects were observed. Respondents were more likely to be willing to provide a saliva sample when this question was asked first compared to first being asked for direct data linkage. Similarly, respondents were more likely to be willing to allow for probabilistic data linkage when this question was asked first compared to first being asked for a saliva sample. CONCLUSION: A lack of willingness to provide biological samples or permit data linkage may lead to representivity issues in studies which rely on such information. The presence of question order effects suggests that the willingness of respondents can be increased through strategic ordering of survey structures.