DRPM: An advanced predictive model for early diabetes detection and risk stratification

Area of Science:

• Computational biology
• Medical informatics
• Public health

Background:

• Diabetes mellitus is a widespread chronic condition with substantial health and economic impacts.
• Early identification and management of diabetes are crucial for preventing complications and improving patient outcomes.
• Existing risk prediction methods may benefit from advanced computational approaches for enhanced accuracy.

Purpose of the Study:

• To develop and validate a deep learning model for predicting the risk of developing diabetes.
• To identify key predictive features for diabetes risk from a comprehensive dataset.
• To create an accessible tool for public use in diabetes risk screening.

Main Methods:

• Utilized data from the National Health and Nutrition Examination Survey (NHANES) from 2011-2018.
• Employed feature selection techniques to identify five critical variables from an initial set of 36.
• Developed a deep learning model for risk prediction and validated its performance on test and external datasets.

Main Results:

• The deep learning model demonstrated high precision in predicting diabetes risk.
• External validation confirmed the model's effectiveness in identifying at-risk individuals.
• A user-friendly online tool and risk stratification system were developed for practical application.

Conclusions:

• The developed deep learning model offers a robust and accurate method for diabetes risk prediction.
• The online tool facilitates personalized early detection and risk screening for diabetes.
• This approach supports public health initiatives for diabetes prevention and management.