Is the matched extreme case-control design more powerful than the nested case-control design?

Area of Science:

• Epidemiology
• Biostatistics
• Survival Analysis

Background:

• Time-to-event data analysis commonly employs the nested case-control design, where controls are matched to cases at the time of the event.
• An alternative, the extreme case-control design, samples controls at a single, pre-specified time point, potentially yielding long-term survivors as controls.

Purpose of the Study:

• To investigate the potential information gain and compare the statistical power of 'extreme' case-control designs against the standard nested case-control design.
• To evaluate the efficiency of extreme case-control designs based on the timing of control sampling.

Main Methods:

• Simulation studies were conducted to compare the power of various extreme case-control designs with the nested case-control design.
• A theoretical expression for average information in nested and extreme case-control pairs was derived for a single binary exposure.
• An application to a dementia study was performed using a 1:1 extreme case-control design.

Main Results:

• The efficiency of the extreme case-control design was found to increase as controls were sampled further out in time.
• The extreme case-control design demonstrated greater power and provided a hazard ratio estimate with a smaller standard error compared to a nested case-control design.
• Apolipoprotein E was identified as a risk factor for dementia using the 1:1 extreme case-control design.

Conclusions:

• Extreme case-control designs, particularly when controls are sampled further from the event time, offer a more informative and powerful alternative to nested case-control designs for time-to-event data.
• The findings suggest that extreme case-control designs can enhance the identification of risk factors and improve the precision of effect estimates in epidemiological studies.