Causal Discovery in Diabetes and its Complications

Abstract

Although previous studies have demonstrated that diabetes and its complications and other factors have high correlation, there is a lack of research using data-driven techniques to infer causality in this field. In this research, causal discovery is used to deal with a purified dataset containing 70,692 survey responses from the CDC’s BRFSS 2015, provided on Kaggle. The final causal graph result obtained through causal discovery strongly aligns with the theory of metabolic syndrome. The findings indicate that abdominal obesity is a leading factor in the development of diabetes and hypertension, and this causal relationship is strong. Hyperglycemia significantly leads to both hypertension and hyperlipidemia, while hypertension also markedly exacerbates dyslipidemia. There is a strong causal relationship between each pair of these conditions. hyperglycemia, hypertension, and hyperlipidemia contribute to the development of cardiovascular disease, with a moderate strength of causal relationship; hyperglycemia, hypertension, and hyperlipidemia all contribute to the development of stroke, with a weak strength of causal relationship. Cardiovascular disease leads to stroke with a weak strength of causal relationship. Aging is strongly causally linked to hyperlipidemia, while its causal relationship with cardiovascular disease is moderate. Using these understandings of the causal relationships between diabetes and its complications and other factors, the author can enhance the health status and quality of life of patients with diabetes on multiple levels by implementing precise prevention strategies, optimizing treatment options, reducing healthcare costs, and improving public health policies.