Comparison of the Validity and Reliability of Five Pressure Injury Risk Scales in Intensive Care

BACKGROUND: The lack of consensus on the best pressure injury (PI) risk assessment tools demonstrates the need for further comparative research to identify the most effective options for intensive care unit (ICU) populations. AIM: This study compared the predictive validity and reliability of five PI risk assessment tools in adult ICU patients. STUDY DESIGN: A prospective cohort study was conducted with patients aged ≥ 18 years, admitted to the ICUs for at least 24 h and without pre-existing PI. Five tools (Braden Scale, CALCULATE, Cubbin & Jackson, EVARUCI and Sunderland) were used daily to assess PI risk until either a PI developed or the participant was discharged, died or completed 21 days in the ICUs. The main outcome measures included area under the receiver operating characteristic (ROC) curve, sensitivity, specificity, relative risk and reliability. RESULTS: Of the 150 participants, 40 (26.7%) developed PIs. None of the five tools demonstrated good predictive accuracy. AUC values ranged from 0.605 to 0.692 across the five tools, with none exceeding the predefined threshold of 0.70. Participants classified as high risk by Sunderland, Cubbin & Jackson and EVARUCI had a 3.0, 2.5 and 2.1 times higher risk, respectively, of developing a PI. Cubbin & Jackson was the only tool to achieve acceptable reliability values in Cronbach's alpha, intraclass correlation coefficient (ICC) and Cohen's weighted kappa. CONCLUSIONS: High-risk classification using the Sunderland, Cubbin & Jackson and EVARUCI scales was more strongly associated with subsequent PI development than classification using the other tools. Reliability findings varied across parameters, with Cubbin & Jackson being the only tool to meet acceptable thresholds across all assessed reliability measures. RELEVANCE TO CLINICAL PRACTICE: PI assessment in ICU patients should integrate validated tools with clinical judgement to address patient-specific and device-related risk factors.