An Automated Method of Causal Inference of the Underlying Cause of Death of Citizens

Area of Science:

• Public Health
• Medical Informatics
• Computational Biology

Background:

• Accurate cause-of-death (COD) inference is vital for resource allocation, particularly during global health crises like the coronavirus disease 2019 (COVID-19) pandemic.
• Traditional manual COD inference methods are resource-intensive and inefficient.
• The need for automated, efficient, and accurate COD inference methods is paramount.

Purpose of the Study:

• To develop and evaluate a novel mixed inference method, Sink-CF, for automated cause-of-death inference.
• To improve the efficiency and accuracy of COD determination compared to traditional methods.
• To reduce the reliance on large, manually labeled datasets for model optimization.

Main Methods:

• A hybrid approach combining mathematical statistics with a collaborative filtering and analysis algorithm from machine learning.
• Development of the Sink-CF algorithm based on confidence measurement for automated inference.
• Utilizing a mixed-method approach to leverage the strengths of both statistical and machine learning techniques.

Main Results:

• The Sink-CF method achieved high performance metrics, including a precision of 93.82% and a recall of 90.11%.
• Demonstrated superior performance compared to existing state-of-the-art machine learning algorithms for COD inference.
• The method showed efficiency by not requiring large amounts of manually labeled data for continuous model optimization.

Conclusions:

• The Sink-CF method offers a simple, easy-to-use, and efficient solution for automated cause-of-death inference.
• This approach conserves computational resources and time by minimizing the need for extensive manual data labeling.
• Sink-CF presents a promising advancement in public health informatics for accurate and efficient COD determination.