Related Experiment Videos
The evolving concept of the healthy worker survivor effect
H M Arrighi1, I Hertz-Picciotto
1Department of Epidemiology, School of Public Health, University of North Carolina at Chapel Hill.
Epidemiology (Cambridge, Mass.)
|March 1, 1994
Summary
The healthy worker survivor effect, where employed individuals are healthier, can weaken study findings on occupational exposures. This review examines four methods to control for this bias in cohort studies.
Area of Science:
- Occupational Epidemiology
- Biostatistics
Background:
- The healthy worker survivor effect is a selection bias where employed workers are healthier than those who leave employment.
- This bias can attenuate (weaken) the observed adverse effects of occupational exposures in epidemiological studies, particularly for subtle associations.
Purpose of the Study:
- To review the concept of the healthy worker survivor effect.
- To discuss and evaluate four proposed methods for controlling this bias in occupational cohort studies.
Main Methods:
- Restriction of the cohort to long-term survivors.
- Lagging exposure assessment to exclude recent exposure.
- Adjusting for employment status as a confounder.
- Utilizing G-null test or G-estimation with structurally nested failure time models to address the effect as both intermediate and confounding.
Main Results:
- The healthy worker survivor effect generally attenuates adverse exposure-response associations.
- Internal referents do not eliminate this effect as it occurs within the occupational cohort.
- There is no consensus on the most appropriate control method.
Conclusions:
- The healthy worker survivor effect is a persistent challenge in occupational epidemiology.
- Understanding and controlling for this bias is crucial for accurate exposure assessment and risk evaluation.
- Further research and consensus are needed on the optimal methods for its control.