An Explainable Artificial Intelligence-Driven Decision Support Framework for Predicting Homoeopathic Treatment Utilization Among Patients with Diabetes and Hypertension: A Secondary Analysis of Multicentre Cross-Sectional Data

Utilization of homoeopathic treatment in patients with diabetes and hypertension can potentially affect the disclosure, reconciliation and clinical communication of medications; however, there is limited predictive evidence of this. This study designed and internally validated an explainable artificial intelligence (XAI) based decision support system with secondary multi-centre cross-sectional data of 400 Bangladeshi adults. Demographic knowledge and attitude variables, clinical knowledge and attitude variables and CAM related knowledge and attitude variables were used in the modelling of the Homoeopathic utilization. The models were penalized logistic regression, Random Forest and XGBoost models which were assessed with repeated nested cross validated analysis, uncertainty estimation using bootstrapping, probability calibration, explainability using SHAP values, sensitivity analysis and decision curve analysis. The number of participants who reported on Homoeopathy was 59 (14.75%). The calibrated extended random forest had the highest overall performance with an ROC-AUC of 0.7893, an average precision score of 0.3530 and a Brier score of 0.1082. Perceived CAM effectiveness, perceived CAM safety and age were the most influential and consistently ranked predictors. Observed utilization increased from 3.7% in the low-risk group to 22.1% and 33.3% in the intermediate- and high-risk groups, respectively. The framework showed a positive net benefit within clinically relevant thresholds and may facilitate targeted inquiry/disclosure, medication reconciliation, and communication of safety when using CAM. External validation is needed prior to clinical routine use.

Authors

Institutions

Publication Details

Journal
Computer and decision making.
Published
2026-11-13
DOI
https://doi.org/10.1922/comdem.v4i1.19080
Primary Topic
Complementary and Alternative Medicine Studies
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

An Explainable Artificial Intelligence-Driven Decision Support Framework for Predicting Homoeopathic Treatment Utilization Among Patients with Diabetes and Hypertension: A Secondary Analysis of Multicentre Cross-Sectional Data

Radha Mangukiya, Zameer Farooq Refai, Harshna S Goti, Akshay M. Patel et al.
Computer and decision making.
Complementary and Alternative Medicine Studies
article

An Explainable Artificial Intelligence-Driven Decision Support Framework for Predicting Homoeopathic Treatment Utilization Among Patients with Diabetes and Hypertension: A Secondary Analysis of Multicentre Cross-Sectional Data

Radha Mangukiya, Zameer Farooq Refai, Harshna S Goti, Akshay M. Patel, Dr. Pankaj N. Lathiya, Dr. Khushbu Navadia, Dr. Pathik kumar Jayeshbhai Patel
article en

Abstract

Utilization of homoeopathic treatment in patients with diabetes and hypertension can potentially affect the disclosure, reconciliation and clinical communication of medications; however, there is limited predictive evidence of this. This study designed and internally validated an explainable artificial intelligence (XAI) based decision support system with secondary multi-centre cross-sectional data of 400 Bangladeshi adults. Demographic knowledge and attitude variables, clinical knowledge and attitude variables and CAM related knowledge and attitude variables were used in the modelling of the Homoeopathic utilization. The models were penalized logistic regression, Random Forest and XGBoost models which were assessed with repeated nested cross validated analysis, uncertainty estimation using bootstrapping, probability calibration, explainability using SHAP values, sensitivity analysis and decision curve analysis. The number of participants who reported on Homoeopathy was 59 (14.75%). The calibrated extended random forest had the highest overall performance with an ROC-AUC of 0.7893, an average precision score of 0.3530 and a Brier score of 0.1082. Perceived CAM effectiveness, perceived CAM safety and age were the most influential and consistently ranked predictors. Observed utilization increased from 3.7% in the low-risk group to 22.1% and 33.3% in the intermediate- and high-risk groups, respectively. The framework showed a positive net benefit within clinically relevant thresholds and may facilitate targeted inquiry/disclosure, medication reconciliation, and communication of safety when using CAM. External validation is needed prior to clinical routine use.

Computer and decision making.Vol. 4(1)
P P Savani University (IN)
Complementary and Alternative Medicine Studies
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon