Population inference with mortality and attrition in longitudinal studies on aging: a two-stage multiple imputation...

Area of Science:

• Gerontology
• Biostatistics
• Longitudinal Studies

Background:

• Longitudinal studies of aging face challenges with participant attrition and mortality.
• Accurate inference is difficult when complete follow-up, including age at death, is not available.
• Incomplete mortality data is common in aging research.

Purpose of the Study:

• To propose a statistical method for incorporating time-to-death into longitudinal models with incomplete follow-up.
• To address the challenge of making inferences conditional on mortality when death data is incomplete.
• To demonstrate the utility of the proposed method using real-world data.

Main Methods:

• A two-stage multiple-imputation procedure is introduced.
• The method is designed to handle missing time-to-death information.
• Statistical modeling is employed to integrate mortality predictors.

Main Results:

• The proposed statistical approach effectively incorporates time-to-death.
• The method facilitates inferences conditional on mortality even with incomplete data.
• The OCTO-Twin study data demonstrated the procedure's practical application.

Conclusions:

• The developed statistical procedure offers a viable solution for incomplete mortality data in aging studies.
• This method enhances the accuracy of longitudinal aging research by accounting for mortality.
• Researchers can improve inferences in studies with missing death information using this technique.