Sociodemographic and psychiatric predictors of attrition in a prospective psychiatric epidemiological study among the general population. Result of the Netherlands Mental Health Survey and Incidence Study-2

Compr Psychiatry. 2013 Nov;54(8):1131-9. doi: 10.1016/j.comppsych.2013.05.012. Epub 2013 Jun 27.


Background: In prospective psychiatric epidemiological studies, attrition at follow-up can be selective, and can bias the research findings. Therefore, knowledge of predictors of attrition and of its different types (noncontact, refusal, inability to participate) is of importance.

Methods: By means of (multinomial) logistic regression analyses, predictors of attrition were studied in the first 3-year follow-up of the second Netherlands Mental Health Survey and Incidence Study (NEMESIS-2), a prospective psychiatric epidemiological study among 6646 subjects of the general population aged 18-64 years. Baseline sociodemographic characteristics, physical health, mental disorders and their clinical characteristics, and experience with the previous interview were studied as predictors of attrition and of its different types.

Results: The attrition rate at follow-up was 20.2%. Refusal (14.2%) was more common than noncontact (4.6%) and inability to participate (1.4%). Compared to respondents, nonrespondents were more often younger, lower educated, unemployed and born outside the Netherlands. A less positive experience with the baseline interview and shorter interview duration also predicted attrition. Any 12-month mental disorder, the categories and separate mental disorders, and their clinical characteristics, were not significantly associated with attrition, after controlling for sociodemographics. Sociodemographic predictors and experience with the baseline interview differed between the three types of attrition, but these types were also hardly or not associated with previous mental disorders.

Conclusions: The authors conclude that bias due to selective attrition was limited to sociodemographics and experience with the baseline interview. Mental health status at baseline was not of influence, possibly due to the large time investment to persuade respondents to re-participate and to find them in case of noncontact or removal to an unknown address. During follow-up waves of future prospective studies it is important to implement an intensive recruitment period with special efforts among young adults and the lower educated.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adolescent
  • Adult
  • Age Factors
  • Female
  • Health Surveys / standards*
  • Humans
  • Interview, Psychological / standards
  • Male
  • Mental Disorders / epidemiology*
  • Middle Aged
  • Netherlands / epidemiology
  • Patient Dropouts / classification
  • Patient Dropouts / statistics & numerical data*
  • Patient Selection
  • Prospective Studies
  • Socioeconomic Factors
  • Young Adult