Bias in hair cortisol measures for psychological stress: Self vs. professional collection

Population-representative samples of hair-based cortisol have the potential to revolutionize studies of psychological stress. However, to integrate hair-based cortisol measurement into large-scale studies, it is crucial to understand the limitations and potential biases associated with self-collection. This study examines the associations between individual, household, and community characteristics, sample collection method, and hair cortisol concentration (HCC). We conducted a random-assignment experiment within a cohort design, comparing self-collection to professional-collection of hair samples among young adults. We attempted contact and hair collection among n = 1181 individuals and obtained hair samples from n = 617 individuals ages 18-22 participating in the Chitwan Valley Family Study in Nepal. The primary outcomes were the ability to provide an analyzable hair sample (defined by sufficient weight, length, and identification of the root-end) and HCC in the top quintile. There was no difference in the probability of providing a hair sample between the self-collection and professional-collection groups. However, self-collection decreased the odds of providing an analyzable sample (OR, 0.47; 95 % CI, 0.23-0.94). Women had higher odds of providing analyzable samples (OR, 9.86; 95 % CI, 3.94-24.72), while being married reduced sample suitability (OR, 0.36; 95 % CI, 0.14-0.92). Women also had decreased odds of high HCC (OR, 0.50; 95 % CI, 0.30-0.83). Our findings show that self-collection of hair samples is feasible for large-scale studies. However, gender influences successful self-collection and is also associated with HCC. Adjusting for gender related bias is both essential and feasible to fully realize the potential of large-scale hair cortisol measurement.