Related Experiment Videos
Hospital versus random digit dialing controls in the elderly. Observations from two case-control studies
S J Robertson1, S Grufferman, H J Cohen
1Department of Pediatrics, Duke University Medical Center, Durham, North Carolina.
Journal of the American Geriatrics Society
|February 1, 1988
Summary
Selecting hospital controls for case-control studies involving the elderly can introduce bias, particularly for chronic diseases. Community controls obtained via random digit dialing may offer a more reliable alternative for hypothesis generation in this demographic.
Area of Science:
- Epidemiology
- Geriatric Medicine
- Biostatistics
Background:
- Case-control studies are essential for etiological research but rely heavily on appropriate control group selection.
- Hospital-based controls, while convenient, may share underlying health conditions with cases, leading to selection bias (Berkson's bias).
- This bias may be amplified in elderly populations due to higher prevalence of multiple chronic diseases.
Purpose of the Study:
- To evaluate the suitability of hospital-based controls versus community-based controls in case-control studies of elderly populations.
- To identify potential biases introduced by using hospital controls in studies of diseases affecting older adults.
- To propose alternative control selection strategies for epidemiological research in geriatrics.
Main Methods:
- The study involved a case-control design with a mean case age of 63.4 years.
- Difficulties encountered with hospital-based controls were documented.
- Comparison was made with community controls recruited through random digit dialing.
Main Results:
- Hospital controls may present with multiple diseases, potentially confounding associations with exposures.
- Cancer cases referred to tertiary centers may be selectively chosen, while non-cancer controls are selected based on diverse morbidities.
- These factors exacerbate Berkson's bias in elderly populations with chronic conditions.
Conclusions:
- Hospital controls can introduce significant bias in case-control studies of the elderly, particularly concerning chronic diseases.
- Community controls selected via random digit dialing appear to be a preferable method for hypothesis-generating studies in this age group.
- Careful consideration of control selection is crucial for the validity of epidemiological research in geriatric populations.