Related Experiment Videos
Designs for synthetic case-control studies in open cohorts
J M Robins1, R L Prentice, D Blevins
1Occupational Health Program, Harvard School of Public Health, Boston, Massachusetts 02115.
Biometrics
|December 1, 1989
Summary
This study introduces efficient case-control designs for open cohorts, optimizing relative risk estimation. Design C is most efficient when control costs are per-person, but alternatives exist if costs depend on distinct controls.
Area of Science:
- Epidemiology
- Biostatistics
Background:
- Case-control studies within cohorts are valuable for epidemiological research, especially with open cohorts allowing late entry.
- Efficient parameter estimation is crucial for reliable relative risk assessment in these designs.
Purpose of the Study:
- To propose and evaluate novel case-control designs for open cohorts.
- To compare the efficiency and consistency of proposed designs against existing methods for relative risk estimation.
Main Methods:
- Development of several case-control designs for open cohorts.
- Examination of designs regarding consistency and efficiency of relative risk parameter estimation.
- A small simulation study to assess design performance.
Main Results:
- Design C is proposed as the most efficient when study costs are proportional to the total number of "at-risk" controls.
- Design C involves random selection of controls from eligible cohort members at the time of a case's failure.
- The efficiency gain of Design C over Thomas's standard design may be small; alternative designs are more efficient if costs are based on distinct "at-risk" controls.
Conclusions:
- The choice of the most efficient case-control design for open cohorts depends on the cost structure related to control selection.
- Design C offers an efficient approach when costs are associated with the total number of controls utilized.
- Alternative designs are necessary and proposed for scenarios where costs are tied to the number of unique "at-risk" controls.