Resident Physician, Internal Medicine McGill University, Canada
Background: Cardiovascular cohort studies rely on comprehensive data acquisition. Current methodologies may impose logistical demands on participants, leading to poor compliance and incomplete data. This missingness may not be random. Women, ethnic minorities, and those with lower socioeconomic backgrounds face structural barriers that hinder full participation. Yet, there is a paucity in Canadian cardiovascular research formally identifying these disparities. Our study aimed to determine which sociodemographic barriers were associated with incomplete participation across key study domains within a cardiovascular patient cohort.
METHODS AND RESULTS: We performed a nested case-control study of N=1803 participants enrolled in the Courtois Cardiovascular Biorepository (CCB), a longitudinal cohort of healthy participants aged 35-79 years at a tertiary healthcare center in Montreal, Quebec, Canada. A categorical response variable was defined based on completion of five core study domains: survey, laboratory, dietary, biospecimen, and physical measures. Participants were classified as full responders (completion of 5/5 domains), partial (3-4/5 domains) or low (≤2/5 domains). Bivariate analyses were conducted using chi-square and t-tests. Multivariable logistic regression was used to assess independent predictors of incomplete response (partial or low), adjusting for age, sex, ethnicity, education, income, Patient-Health Question-9 (PHQ-9) score (a depression severity measure), and employment hours.
There were 398 participants (22.1%) classified as full responders, 1335 (74.1%) as partial responders, and 70 (3.9%) low responders. In the bivariate analysis, younger age, lower income and non-white ethnicity were associated with incomplete participation (p < 0.05). In the multivariable analysis, lower income remained independently associated with incomplete response (aOR 2.82 for $30,000 - $49,999 vs ≥$100,000, p < 0.001). Conversely, Afro-Caribbean ethnicity (aOR 0.45, p=0.037), lower education (aOR 0.42 for trade vs post-secondary, p=0.020), and higher depression (higher PHQ-9, aOR 0.96/point, p=0.034) were independently associated with lower odds of incomplete response. Sex and employment hours were non-significant predictors. Detailed results are listed in Table 1.
Conclusion: Lower household income was independently associated with incomplete response across five key study domains within this large population-based cohort study. These findings highlight the need to proactively address structural barriers, particularly financial, that may hinder full participation in cardiovascular research. Without recognizing and addressing these disparities, cardiovascular research risks perpetuating the systemic exclusion of participants, reducing generalizability to underserved populations. Ensuring culturally sensitive engagement, flexible scheduling and financial accessibility may promote equitable research and help reduce missingness in cardiovascular cohort studies.
