Participation, retention, and associated factors of women in a prospective multicenter study on Chlamydia trachomatis infections (FemCure)

Prospective studies are key study designs when attempting to unravel health mechanisms that are widely applicable. Understanding the internal validity of a prospective study is essential to judge a study’s quality. Moreover, insights in possible sampling bias and the external validity of a prospective study are useful to judge the applicability of a study’s findings. We evaluated participation, retention, and associated factors of women in a multicenter prospective cohort (FemCure) to understand the study’s validity.Chlamydia trachomatis (CT) infected adult women, negative for HIV, syphilis, and Neisseria gonorrhoeae were eligible to be preselected and included at three sexually transmitted infection (STI) clinics in the Netherlands (2016–2017). The planned follow-up for participants was 3 months, with two weekly rectal and vaginal CT self-sampling and online questionnaires administered at home and at the clinic. We calculated the proportions of preselected, included, and retained (completed follow-up) women. Associations with non-preselection, noninclusion, and non-retention (called attrition) were assessed (logistic and Cox regression).Among the 4,916 women, 1,763 (35.9%) were preselected, of whom 560 (31.8%) were included. The study population had diverse baseline characteristics: study site, migration background, high education, and no STI history were associated with non-preselection and noninclusion. Retention was 76.3% (n = 427). Attrition was 10.71/100 person/month (95% confidence interval 9.97, 12.69) and was associated with young age and low education. In an outpatient clinical setting, it proved feasible to include and retain women in an intensive prospective cohort. External validity was limited as the study population was not representative (sampling bias), but this did not affect the internal validity. Selective attrition, however (potential selection bias), should be accounted for when interpreting the study results.