Pressure injuries in surgical patients: a comparison of Norton, Braden and Waterlow risk assessment scales

Objective: This study aims to determine the predictive power of the Norton, Braden and Waterlow scales in determining risk of pressure injury (PI) in surgical patients.

Method: This prospective study was carried out in the surgery clinic of a training and research hospital in Istanbul, Turkey between January and April 2017. The study sample consisted of adult patients aged ≥18 years and who did not have PI on admission to the clinic, had abdominal surgery under general anaesthesia and who stayed in the clinic for at least 48 hours. The data were collected using the Turkish versions of the Norton, Braden and Waterlow risk assessment scales. The predictive validity of PI risk assessment tools was assessed based on their sensitivity, specificity, positive and negative predictive values and the area under the receiver operating characteristic (ROC) curve. Predictive capacity was measured as relative risk (RR) with 95% confidence intervals (CI).

Results: The study sample included 250 patients, and the incidence of PI was 12%. The sensitivity, specificity, positive predictive value and negative predictive value were: 83.3%, 45.4%, 17.2% and 95.2%, respectively, for the Norton scale (a cut-off point of 14); 100%, 40.4%, 18.6% and 100%, respectively, for the Braden scale (a cut-off point of 16); and 100%, 48.1%, 20.8% and 100%, respectively, for the Waterlow scale (a cut-off point of 10). The areas under the ROC curve were 0.749 for the Norton, 0.771 for the Braden and 0.971 for the Waterlow scales. This study's findings produced the following predictive capacity indicators: Norton (RR=3.62; 95%CI=1.43-9.14), Braden (RR=33.88; 95%CI=2.09-547.66); and Waterlow (RR=45.01; 95%CI=2.78-727.97).

Conclusion: In this study, the Waterlow scale demonstrated the best values of predictive validity among the three scales in the assessment of PI risk. However, all three scales had low specificity despite high sensitivity in terms of a good risk prediction. No definitive decision could be reached on the predictive capacities of the scales because of wide CIs.