Effectiveness of pain assessment tools in non-verbal ICU patients: A meta-analysis-based review

Background: Assessing pain in ICU patients unable to self-report represents a significant clinical challenge. Observational tools such as the Critical‑Care Pain Observation Tool (CPOT), Behavioral Pain Scale (BPS), and Nonverbal Pain Assessment Tool (NPAT) have been developed to address this gap. Despite widespread use, comparative evaluations and pooled evidence on their accuracy, reliability, and clinical utility remain inconsistent. Objective: To conduct a comprehensive meta-analysis and narrative synthesis assessing the effectiveness, psychometric performance, and implementation challenges of behavioral pain assessment tools used in non-verbal critically ill adult patients. Methods: We systematically searched PubMed, Scopus, Cochrane, and Embase for validation studies, randomized controlled trials, observational cohorts, and implementation reports involving adult ICU patients incapable of self-reporting. We included studies that evaluated CPOT, BPS, NPAT, PAINAD, NCS‑R, and related scales. Primary outcomes comprised tool sensitivity, specificity, inter-rater reliability (ICCs/κ), internal consistency (Cronbach’s α), discriminant validity, and feasibility metrics. Quality assessments were conducted using QUADAS‑2 and GRADE; pooled estimates with random-effects meta-analysis; and heterogeneity quantified via I², with funnel plots and Egger’s test for bias. Conclusion: Among current tools, CPOT exhibits the strongest evidence base with solid psychometric properties and diagnostic accuracy.