Bad questions: addressing non‐response and respondent feedback in a social and structural determinants of health (SSDoH) questionnaire
Shana D. Stites, Kimberly Halberstadter, Jada Jeanette Houston, Rosalie Schumann, Dawn Mechanic‐Hamilton- Psychiatry and Mental health
- Cellular and Molecular Neuroscience
- Geriatrics and Gerontology
- Neurology (clinical)
- Developmental Neuroscience
- Health Policy
- Epidemiology
Abstract
Background
Scientists must be able to characterize and measure social and structural determinants of health (SSDoH) in order to study their influences on Alzheimer’s disease and related dementias outcomes. Constructing measures that capture this heterogeneity is challenging. We summarize SSDoH questions that participants skipped and critiqued and our response.
Method
An online survey of 225 items capturing SSDoH was administered to cognitively unimpaired participants (n = 135) in the UPenn Alzheimer’s Disease Research Center clinical cohort. Self‐report questionnaire categories were: (1) Education, (2) Occupation, (3) Economic Status and Strain, (4) Stressors and Subjective Stress, (6) Subjective Social Support and Status, and (7) Sociodemographic data, which included age, race, language, disability, neighborhood, sex, sexual orientation, and gender identity questions. We summarize pilot data on item and instrument completion rates, participant feedback, and survey revisions to address participant critiques.
Result
The response and completion rates were 66% and 95% respectively. Most often skipped items were: mother’s highest education level (17% skipped), number of years attending a segregated school for those who did attend (15% skipped), childhood zip‐code (14% skipped), transportation mode around one’s neighborhood (13% skipped), and current worldview compared to during childhood (12% skipped). Few participants (n = 5/135) reported dislike or discomfort with any questions, with 3 of the 5 citing that questions about childhood caused discomfort. Participants also critiqued questions about education for not capturing their individual schooling experiences. In response to feedback, we conducted follow‐up discussions with participants about their reactions to specific childhood questions and expanded questions to capture more educational pathways, including trade school experiences and apprenticeships. We also added answer choices to often‐skipped items with insufficient answer options.
Conclusion
Overall, most SSDoH items had high completion rates, and few questions were consistently skipped by participants. While investigators often worry about offending or upsetting research participants, those in our study wanted to help improve questions. Participants acknowledged that some questions elicited strong emotions, but in most cases were both willing to complete these questions and appreciative of the chance to share their experiences. Designing questions that capture a fuller range of educational pathways was a valuable contribution to data collection.