Developing an artificial intelligence-based clinical decision support system for nursing diagnoses.

BACKGROUND

Nursing diagnoses are crucial for care planning and communication, yet achieving diagnostic accuracy is often challenging due to workload and complex patient data. Integrating artificial intelligence (AI) into clinical practice may enhance clinical reasoning and promote a standardized nursing language.

AIM: This methodological study aimed to develop an AI-based clinical decision support system that ensures diagnostic accuracy with fewer defining characteristics.

METHODS

Based on Gordon's Functional Health Patterns Model, data were collected from 122 patients, including 117 defining characteristics, 59 risk factors, and 19 nursing diagnoses. Machine learning algorithms were tested for accuracy.

RESULTS

The Naive Bayes algorithm achieved 96.33% accuracy using all indicators, while Gradient Boosting reached 82.94% with the 50 most important variables. Both effectively identified key diagnoses such as "Obesity" and "Acute Pain."

CONCLUSION

The developed AI-based decision support system demonstrated high accuracy and potential to assist nurses in making faster and more consistent diagnostic decisions.