The healthy worker effect: Do health problems predict participation rates in, and the results of, a follow-up survey?

Int Arch Occup Environ Health. 2016 Feb;89(2):231-8. doi: 10.1007/s00420-015-1066-6. Epub 2015 Jun 24.

Abstract

Purpose: To determine the impact of the healthy worker effect (HWE) as a bias for the external and internal validity of the follow-up assessment in prospective survey research. Specifically, the study examined (1) whether the health status of respondents at the baseline measurement influenced response at the follow-up survey (external validity) and (2) whether HWE is a threat to internal validity by differential attrition, i.e., whether associations between work and health at baseline differ between stayers and dropouts.

Methods: In a two-wave questionnaire survey with a 2-year time lag comprising 6283 persons, 4392 responded at both time points (response rate 70%). Mental distress and somatic symptoms served as indicators of health. Role conflict and role clarity were indicators of work factors.

Results: There were few differences in response rate at follow-up between persons with and without health complaints at the baseline measurement. As response rate increased incrementally with educational level, there seems to be a socio-educational bias, rather than a HWE bias on survey participation. Baseline relationships between work factors and health indicators were equal in magnitude among stayers and dropouts.

Conclusion: The health status of participants at baseline seems to have little impact on the external and internal validity of the follow-up assessment in prospective survey research. Hence, the findings provide little support to the HWE as a potential bias in prospective studies within occupational health research. A limitation of the study is that the findings do not inform about the impact of the HWE on participation in the baseline assessment.

Keywords: Attrition; Generalizability; Non-response bias; Participation; Research design.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Abdominal Pain / epidemiology
  • Adolescent
  • Adult
  • Aged
  • Aged, 80 and over
  • Back Pain / epidemiology
  • Bias
  • Chest Pain / epidemiology
  • Educational Status
  • Female
  • Follow-Up Studies
  • Headache / epidemiology
  • Health Surveys / statistics & numerical data*
  • Healthy Worker Effect
  • Humans
  • Male
  • Middle Aged
  • Neck Pain / epidemiology
  • Norway
  • Occupational Health*
  • Pain / epidemiology*
  • Prevalence
  • Reproducibility of Results
  • Stress, Psychological / epidemiology*
  • Young Adult