Development of an Intelligent Predictive Indicator System for Weaning and Extubation Timing in Mechanically Ventilated Patients Based on the Delphi Method

Objective To develop an intelligent predictive indicator system for determining weaning and extubation timing in mechanically ventilated patients, providing a theoretical foundation for clinical decision support systems and intelligent assessment in critical care nursing. Methods A literature review was conducted to identify and extract factors influencing weaning and extubation timing in mechanically ventilated patients, forming the initial item pool for the indicator system. A Delphi expert consultation questionnaire was designed and administered to 21 experts from relevant fields over two rounds. Item screening and revision were performed using the boundary value method, based on the arithmetic mean of importance scores, coefficient of variation (CV), and qualitative expert feedback. Results The effective response rate was 100% for both rounds. The expert authority coefficient (Cr) was 0.89 in the first round and 0.93 in the second round. Kendall's W was 0.244 ({chi}{superscript 2} = 1038.944, P < 0.001) in the first round and 0.079 ({chi}{superscript 2} = 350.343, P < 0.001) in the second round. Following the first round of evaluation of 203 items--which resulted in 13 additions, 23 modifications, and the separation of weaning/extubation dimensions--the second round comprised 211 evaluation items. The final system encompassed 6 primary indicators and 58 secondary indicators (with tertiary indicators under each), covering the complete clinical trajectory: medical history, ICU admission assessment, pre-weaning/extubation assessment, weaning assessment, extubation assessment, and post-extubation outcomes. Conclusion This indicator system was developed through a rigorous process with high expert authority and good consensus. The indicators are clinically comprehensive, nursing-operable, and suitable for intelligent modeling. This system can serve as a core framework for risk warning, timing judgment, and intelligent prediction system development for weaning and extubation in mechanically ventilated patients.