An efficient data integration scheme for synthesizing information from multiple secondary datasets for the parameter...

Area of Science:

• Biostatistics
• Epidemiology
• Clinical Trials

Background:

• Secondary outcomes in studies are often analyzed separately from primary endpoints.
• These secondary outcomes can be highly correlated with primary endpoints, offering potential for improved analysis.
• Existing methods do not fully leverage secondary outcome data for primary analysis efficiency.

Purpose of the Study:

• To introduce Multiple Information Borrowing (MinBo), a new statistical method.
• To improve the estimation precision of primary analyses by borrowing information from secondary data.
• To demonstrate the robustness and efficiency gains of MinBo against model misspecification.

Main Methods:

• Developed the Multiple Information Borrowing (MinBo) statistical framework.
• Employed theoretical analyses to establish the method's properties.
• Conducted case studies, including application to the Atherosclerosis Risk in Communities (ARIC) study data.

Main Results:

• MinBo demonstrated superior efficiency gains compared to existing methods in both theoretical and case studies.
• The method proved robust against potential model misspecification in secondary data.
• Application to ARIC study data provided insights into hypertension risk factors.

Conclusions:

• MinBo offers a statistically sound and efficient approach to integrate secondary outcome data into primary analyses.
• The method enhances precision without compromising robustness, advancing secondary data utilization in research.
• MinBo has significant implications for improving the analysis of observational studies and clinical trials.