Validity, reliability and responsiveness of the Chelsea Critical Care Physical Assessment tool (CPAx) for patients in neurocritical care.

Abstract

Examine inter-tester reliability, construct validity, and responsiveness of the Chelsea Critical Care Physical Assessment tool (CPAx) in neurocritical care patients. Two physiotherapists conducted the baseline CPAx assessment, and one the follow-up assessment. Reliability was examined with Intraclass Correlation Coefficient (ICC) model 2.1 for the total score and Quadratic Weighted Kappa for each item. Measurement error was assessed with Standard Error of Measurement (SEM) and Minimal Detectable Change (MDC). Construct validity and responsiveness were examined using predefined hypotheses regarding the expected Spearman rank correlation coefficients between CPAx and similar instruments. In total, 74 patients were included. Excellent reliability was found with an ICC of 0.99 (0.985;0.994). SEM was 1.08 points, and MDC was 2.99 points. The quadratic weighted kappa of each item ranged between 0.81- 0.97, and the agreement between 97.68% - 99.51%. For construct validity, 67% of our predefined hypotheses were confirmed, and 78% of those for responsiveness. CPAx is a reliable and responsive tool for neurocritical care patients. Construct validity was below the recommended threshold for high validity. The absence of floor or ceiling effects indicates that CPAx is suitable for evaluating physical function, including aspects of coughing and respiratory function, in neurocritical care patients.