Photo Credit: Ekaterina

The following is a summary of “Utilizing machine learning algorithms for predicting Anxiety-Depression Comorbidity Syndrome in Gastroenterology Inpatients (ADCS-GI),” published in the March 2025 issue of BMC Psychiatry by Tan et al.

Diagnosing Anxiety-Depression Comorbidity Syndrome in Gastroenterology Inpatients (ADCS-GI) is challenging due to patient hesitance and non-psychiatric professionals’ limitations. 

Researchers conducted a retrospective study to evaluate machine learning for detecting ADCS-GI.  

They recruited 1,186 patients with ADCS and conducted data quantification, equilibrium, and correlation analysis. About 8 machine learning models—Gaussian Naive Bayes (NB), Support Vector Classifier (SVC), K-Neighbors, RandomForest, XGB, CatBoost, Cascade Forest, and Decision Tree—were used to compare prediction efficacy while minimizing reliance on subjective questionnaires. 

The results showed that Decision Tree and K-Neighbors achieved over 81% accuracy and sensitivity in detecting ADCS. For moderate and severe cases, accuracy exceeded 88% with 90% sensitivity. Models without subjective questionnaires performed well, supporting questionnaire-free early detection. 

Investigators used machine learning to identify ADCS in patients with gastroenterology, aiding early detection and intervention in their care. 

Source: bmcpsychiatry.biomedcentral.com/articles/10.1186/s12888-025-06666-x