Design considerations for estimation of exposure effects on disease risk, using aggregate data studies

Area of Science:

• Epidemiology
• Biostatistics
• Cancer Etiology

Background:

• Aggregate data studies offer a novel approach to estimating exposure effects on disease rates.
• Understanding the role of diet in cancer etiology is a key public health challenge.
• Existing methods may not fully leverage population-level data for disease rate analysis.

Purpose of the Study:

• To propose and evaluate an aggregate data study design for estimating exposure effects.
• To differentiate aggregate data studies from traditional ecologic studies.
• To provide guidance on optimizing the design of aggregate data studies for cancer research.

Main Methods:

• Development of a random effects relative rate model aggregated from individual-level models.
• Utilizing population-based disease rates and risk factor survey data.
• Conducting simulation studies to assess the impact of study design parameters.

Main Results:

• Aggregate data studies yield relative rate parameter estimates comparable to individual-level studies.
• Increasing the number of populations (from 20 to 30-40) enhances statistical power more effectively than increasing individual sample size.
• The proposed design effectively uses between-group variations in the data.

Conclusions:

• Aggregate data study designs provide a valuable framework for epidemiological research, particularly in cancer etiology.
• Optimizing the number of populations is crucial for maximizing statistical power in these studies.
• This approach offers a distinct and powerful alternative to ecologic studies for exposure-effect estimation.