Abstract
BACKGROUND: Supraclavicular brachial plexus block is a commonly used regional anesthesia technique for upper extremity surgeries. The addition of adjuvants to local anesthetics enhances the duration and quality of anesthesia. Recently, Machine Learning (ML) has emerged as a tool for predictive modeling in medicine, including pain management. This study investigates the predictive capability of a K-Nearest Neighbor (KNN) ML model for analgesic duration using different drug combinations. METHODS: A prospective randomized controlled trial was conducted on 60 patients scheduled for upper limb surgeries under ultrasound-guided supraclavicular brachial plexus block. Patients were divided into three groups: Control (bupivacaine + saline), Nalbuphine (bupivacaine + nalbuphine), and Morphine (bupivacaine + morphine). The analgesic duration, sensory and motor block onset and duration, and postoperative analgesic use were recorded. A ML model, K-Nearest Neighbor (KNN), was developed to predict analgesic time based on demographic and hemodynamic parameters. RESULTS: Both nalbuphine and morphine significantly prolonged the duration of analgesia compared to the control group. The KNN model demonstrated a strong correlation (R = 0.95) between the observed and predicted analgesic duration, indicating high predictive accuracy. CONCLUSIONS: Nalbuphine and morphine significantly extended the analgesic duration of bupivacaine in ultrasound-guided supraclavicular brachial plexus block. ML models, such as KNN, offer effective tools for predicting analgesic outcomes and can assist anesthesiologists in making informed decisions regarding drug combinations for enhanced patient care. TRIAL REGISTRATION: ClinicalTrials.gov (NCT07008443). Registered in June 2025. Retrospectively registered. ClinicalTrials.gov is a primary registry in the WHO International Clinical Trials Registry Platform (ICTRP) network. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12871-025-03537-6.