Overcoming Missing Data: Accurately Predicting Cardiovascular Risk in Type 2 Diabetes, A Systematic Review

Area of Science:

• Medical Informatics
• Epidemiology
• Biostatistics

Background:

• Cardiovascular disease (CVD) is a major complication in type 2 diabetes mellitus (T2DM).
• Accurate prediction models are crucial for managing CVD risk in T2DM patients.
• Missing data presents a significant challenge in the development and validation of these models.

Purpose of the Study:

• To investigate the prevalence and handling strategies of missing data in CVD prediction models for T2DM.
• To assess the reporting quality of missing data methodologies in relevant studies.

Main Methods:

• Systematic literature search of MEDLINE for English-language studies up to June 30, 2024.
• Extraction and summarization of missing data percentages, mechanisms, and handling strategies.
• Analysis of 51 articles focusing on prediction model development and validation.

Main Results:

• Missing data was reported in most development (40/51) and external validation (12/16) studies.
• Imputation was the predominant method for handling missing data in both development (74.5%) and validation (68.8%) stages.
• Reporting of missing data increased significantly after 2016, but clarity on methodologies remains insufficient.

Conclusions:

• While missing data is frequently encountered and addressed in T2DM CVD prediction models, reporting practices are often inadequate.
• Improved transparency and application of robust missing data handling techniques are essential for enhancing prediction model quality and reliability.
• Further research should focus on standardized reporting guidelines for missing data in clinical prediction models.