Although attrition is inevitable in longitudinal epidemiological studies, psychiatric studies are thought to be especially sensitive to attrition. This study aimed to evaluate the sociodemographic and psychiatric determinants of attrition at 2-year follow-up in the Netherlands Study of Depression and Anxiety.
Methods
Logistic regression was used to examine sociodemographic and psychiatric determinants of attrition and the influence of clinical psychiatric characteristics on attrition. In addition, differences in determinants between 3 types of attrition (refusal, noncontact, and not able to participate) were evaluated.
Results
The attrition rate at the 2-year follow-up assessment was 12.9% (385/2981), representing 6 deceased persons, 250 refusers, 51 noncontacts, and 78 persons unable to participate because of health reasons. Determinants of attrition were younger age, less years of education, not being of North European descent, being recruited in Amsterdam, no previous participation in research, and having major depressive disorder. Only the effects of age, sampling site, and previous participation in research differed between types of attrition. Furthermore, comorbid depressive and anxiety disorders and higher symptom severity were associated with attrition.
Conclusions
In contrast to the view that psychiatric epidemiological research is more prone to high attrition rates, this study revealed a relatively low attrition rate. Furthermore, both sociodemographic and psychiatric variables were independent determinants of attrition. Oversampling of subgroups that are at higher risk of dropout may be advisable for future psychiatric cohort studies.
Cited by (0)
Conflicts of interest: WA Nolen has received grants from the Netherlands Organisation for Health Research and Development, European Union, Stanley Medical Research Institute, Astra Zeneca, Eli Lilly, GlaxoSmithKline, and Wyeth, received honoraria/speaker's fees from Astra Zeneca, Eli Lilly, Pfizer, Servier, and Wyeth, and is advisory board member for Astra Zeneca, Cyberonics, Pfizer, and Servier. All other authors declare that they have no conflicts of interest.