Predictors of Adolescents' Willingness to Participate in Long-term Studies: Results of the Ohrkan Cohort Study on Leisure Time Noise Exposure

Abstract

The loss of participants during the follow-up of cohort studies can introduce bias if persons who leave the study differ from those who continue to participate over time. Many studies have investigated socio-demographic and health factors associated with participation/non-participation in epidemiological studies. However, relatively few studies are available for school-based cohorts. The aim of this study was to determine predictors of willingness to participate in the Ohrkan study, an ongoing longitudinal survey of adolescents that collects detailed information on leisure time noise exposure and potential effects on hearing. Using the Ohrkan cohort of 2,148 persons recruited in 2009-2011 (O-I) among the ninth graders of any secondary school type in Regensburg and followed-up in 2012-14 (O-II) and 2015-16 (O-III). The probability of participating in O-II and O-III was higher for persons without a migration background, who attended university-preparatory secondary school at baseline and whose parents had a higher educational level or lived together. Persons who had not changed their place of residence in the last 2.5 years also participated 3 times as often as respondents who moved inside or outside of Regensburg. The knowledge of determinants of attrition and retention makes it easier to specify which groups of a cohort are most likely to be lost to follow-up. These groups can be targeted in order to minimize their loss in the future follow-ups. Moreover, the predictors can be used in statistical missing data techniques such as multiple imputation.