Factors Affecting Non-response in a Cohort Study of Female Participants: The Korea Nurses' Health Study (Preprint)

The major drawback of a cohort study design is the loss of follow-up participants, which increases selection bias and threatening external validity, particularly in online website surveys. It is important to identify factors beyond population or demographics that influence non-response rates relevant to cohort studies. This study aimed to examine the non-response rate and factors over a 10-year follow-up period among female participants, using data from initial and subsequent surveys. The Korea Nurses' Health Study recruited 20,613 female nurses in 2013, using simple random sampling. The participants were followed up ten times through 2022. We identified the demographic, work-related, survey-related, and psychological characteristics of non-responding nurses with 10-year follow-up and compared them with those who continued to participate. Descriptive, chi-square, and multivariate logistic regression models were used. The non-response rate of the follow-up surveys from the 2nd to the 11th varied between 25.5% (second survey) and 61.2% (sixth survey). The influence of age, education, and the usability of survey websites on non-response lasted up to the 11th survey. Nurses in their 20s were more likely not to respond to the follow-up surveys than those in their 30s. Those who had an associate degree and neutral feelings about the usability of the survey websites were less likely to respond to follow-up surveys than those who were satisfied with the initial survey. The influence of geographic region, hospital size, and psychological factors-including stress, fatigue, and sleep disturbance-was evident from the second to the sixth survey. When designing and recruiting female nurse participants for community-based cohort studies, researchers should consider the factors that influence non-response and adopt tailored strategies based on demographic characteristics. In addition, it is recommended that the usability of survey websites be improved to reduce non-response at follow-up in cohort studies involving female participants. RR2-DOI: https://doi.org/10.4178/epih.e2024048.