Related Experiment Videos
Recall bias in case-control studies: an empirical analysis and theoretical framework
1Department of Medicine, McMaster University, Hamilton, Ontario, Canada.
Journal of Clinical Epidemiology
|February 1, 1995
Summary
This study found no link between lower study quality and recall bias in case-control studies. Data quality comparisons between cases and controls offer insights into potential bias and study rigor.
Area of Science:
- Epidemiology
- Biostatistics
Background:
- Recall bias is a significant concern in case-control studies.
- Previous research suggested recall bias worsens with lower study quality.
Purpose of the Study:
- To systematically review literature examining the relationship between study quality and recall bias.
- To assess if lower overall agreement correlates with increased recall bias.
Main Methods:
- Systematic literature search (1966-1990) for case-control studies with validated gold standards.
- Included studies reporting agreement rates, kappa, sensitivity, or specificity.
- Analyzed 16 identified studies.
Main Results:
- No relationship found between agreement differences and overall agreement levels.
- Linear relationships observed between data quality comparisons (cases vs. controls) and agreement metrics (correlations: 0.81-0.62).
- Kappas were lower than crude agreement; specificities higher than sensitivities.
Conclusions:
- Data quality comparisons are valuable for detecting differential recall and assessing study quality.
- A theoretical framework is proposed for evaluating recall bias and study rigor.