Identifying risk patterns for suicide attempts in individuals with diabetes : a data-driven approach using LASSO regression
Metadata[+] Show full item record
Diabetes is a major health concern in the United States, with 34.2 million Americans affected in 2020. Unfortunately, the risk of suicide is also elevated in individuals with diabetes, with around 90,000 people with diabetes committing suicide each year. People with type 1 diabetes are three to four times more likely to attempt suicide, and those with newly diagnosed type 2 diabetes are twice as likely to attempt suicide compared to the general population. However, poor mental health comorbidity is still neglected, and more recommendations are needed to support for people with diabetes. It is widely acknowledged that the comorbidity of depression with diabetes is considered a higher risk factor for suicide attempts Previous studies have used logistic regression to identify risk factors for suicide attempts in individuals with diabetes. However, this technique can be prone to overfitting when the number of variables is high. To address this issue, we used the LASSO (Least Absolute Shrinkage and Selection Operator), a regularization technique, to reduce overfitting in a logistic regression model. It works by adding a penalty term ([lambda]) to the log-likelihood function, which shrinks the estimates of the coefficients. This process allows LASSO to act as a feature selection method, effectively setting coefficients that contribute most to the error to zero. Because few studies have focused on un derstanding the relationship between suicide attempts and diabetes, we used association rule mining ARM an explainable rule based machine learning technique, for knowledge discovery to reveal previously unknown relationships between suicide attempts and diabetes. This approach has already proved useful in the medical field, where it has been applied to electronic health record (EHR) data to discover associations such as disease co-occurrences, drug-disease associations, and symptomatic patterns of disease. However, no previous studies have used ARM to determine risk factors and predict suicide attempts in people with diabetes. The aim of this dissertation is to identify patterns of risk factors for suicide attempts in individuals with diabetes, with the long term goal of developing a clinical decision support system that can be integrated into EHRs. This system would allow healthcare providers to identify patients with diabetes at high risk of suicide attempts and provide appropriate preventive measures during outpatient clinic visits. To achieve this goal, we have three specific aims: (1) to identify potential risk factors for suicide attempts in individuals with diabetes through a literature review; (2) to investigate risk factors for suicide attempts in individuals with diabetes using LASSO regression; (3) to identify risk patterns for suicide attempts in individuals with diabetes using association rule mining. In this dissertation, we have reviewed the literature and compiled a list of data elements for suicide attempts in people with diabetes. We then retrieved data on patients with diabetes from Cerner Real-World Data [trade mark]. LASSO regression was used for feature selection, and ARM was used for investigating the risk patterns. We discovered risk patterns that are understandable and practical for healthcare providers. The findings of this research can inform suicide prevention efforts for people with diabetes and contribute to improved mental health outcomes.