Developing a decision support system for classifying care needs and determining nursing diagnoses in older patients with cardiometabolic multimorbidity

INTRODUCTION AND AIM: The aim of the first phase of this three-stage study was to determine nurses' opinions on the development of a clinical decision support system (CDSS), the second phase aimed to develop the "Symptom Severity Scale in Cardiometabolic Multimorbid Patients (SSS-CM)," and the final phase aimed to develop a machine learning-based CDSS to classify patients' care needs and determine nursing diagnoses. MATERIALS AND METHODS: In the first phase, semi-structured interviews were conducted with 20 nurses who cared for older adults with cardiometabolic multimorbidity. The data were analyzed through thematic analysis using an inductive approach via MaxQDA software. In the second phase, 388 patients aged 65 years and older with at least two cardiometabolic diseases were included. The final phase was conducted with 700 patients. Based on the dataset containing various features, Random Forest, XGBoost, and SVM machine learning algorithms were used to predict care needs based on frailty level. For each of the ten nursing diagnoses, separate decision tree models were trained and evaluated. RESULTS: Three main themes were identified: experiences during the care process, classification of care needs and prioritization of nursing diagnoses, and views on the development of a decision support system. The Cronbach's alpha coefficient of the developed scale was found to be 0.978, and it showed good model fit overall. For binary classification with the top 19 features, the XGBoost model performed best (86%), while for three-class classification, the SVM model showed the highest accuracy (76%). The ten nursing diagnoses evaluated using the decision tree method were classified with high accuracy, ranging from 97% to 100%. CONCLUSION: Utilizing machine learning-based CDSSs based on multidimensional features may facilitate nurses' clinical decision-making when planning care for older individuals with complex care needs. Keywords: Nursing care, cardiometabolic multimorbidity, older adults, clinical decision support system, machine learning, artificial intelligence