Dynamic Prediction of Sleep Protection Demand on the First Postoperative Night and Its Influence on Ward Process After Thoracoscopic Lobectomy or Segmentectomy

This study combines retrospective data analysis with a single-center prospective non-interventional cohort study to understand postoperative sleep disturbance on the first night after video-assisted thoracoscopic (VATS) lobectomy or segmentectomy. It mainly wants to answer the question: Can perioperative data collected along the clinical timeline dynamically predict which patients will experience clinically significant sleep disturbance on their first postoperative night? Participants undergoing elective VATS lung resection will have multidimensional data systematically collected from preoperative screening (including PSQI and GAD-7), intraoperative anesthetic parameters, PACU-to-ward handover, pre-bedtime symptom burden, and follow-up through 1 month. Three dynamic prediction models (M0 preoperative screening, M1 early postoperative update, and M2 pre-bedtime main model) will be developed using logistic regression with RCSQ-assessed sleep disturbance as the primary outcome, while an explanatory model will quantify the impact of nighttime disruptions such as intravenous infusions, vital sign monitoring, nursing entries, and awakenings.