Development and preliminary clinical validation of a mobile health application for pressure injury staging in ICU patients.

BACKGROUND: Pressure injury is a common adverse event among critically ill patients and is associated with poor clinical outcomes, increased nursing workload, and substantial healthcare costs. Accurate staging of pressure injuries is essential for guiding nursing interventions; however, conventional visual assessment is highly dependent on individual experience and often shows limited consistency.

METHODS: A prospective cross-sectional study was conducted in the intensive care units (ICUs) of tertiary hospitals. A smartphone-based application integrating wound image acquisition and intelligent assisted staging was developed. Pressure injury images were collected by trained nurses following a standardized protocol. Each injury was independently staged by a wound, ostomy, and continence nurse (WOCN), and a consensus reference standard was established. The staging results generated by the application were compared with expert consensus using an agreement analysis.

RESULTS: A total of 52 patients with 269 pressure injuries ranging from Stage 1 to Stage 4 were included in the study. To assess the consistency of expert-based staging, the weighted kappa value was calculated. The results showed that the weighted kappa value between the two WOCNs was 0.872 (95% CI: 0.698-0.944, p < 0.001), indicating a high level of agreement between the expert assessments. Further comparison of the mobile health application and the expert consensus revealed that the inter-rater agreement coefficient between the iOS version of the application and expert consensus was 0.769 (95% CI: 0.632-0.906, p < 0.001), while the inter-rater agreement coefficient between the Android version and expert consensus was 0.821 (95% CI: 0.698-0.944, p < 0.001), both showing substantial agreement. The classification consistency for each stage of pressure injuries ranged from 76.9 to 84.6%. Additionally, the inter-rater agreement between the registered nurse and the mobile health application for wound length, width, and area measurements was good (range: 0.843-0.914, p < 0.001). The intra-rater reliability measurements for wound length, width, and area in both the iOS and Android systems were excellent (iOS range: 0.952-0.986, Android range: 0.959-0.975, p < 0.001).

CONCLUSION: The application was easy to operate and was well accepted by nursing staff. This study provides preliminary evidence that a mobile health application can serve as a feasible and objective adjunct tool for pressure injury staging, supporting more standardized and consistent nursing assessments.