Development of a Multivariable Prediction Model for Risk of Hospitalization With Pressure Injury After Entering Residential Aged Care

Objectives: Although largely preventable, pressure injury is a major concern in individuals in permanent residential aged care (PRAC). Our study aimed to identify predictors and develop a prognostic model for risk of hospitalization with pressure injury (PI) using integrated Australian aged and health care data. 

Design: National retrospective cohort study. 

Setting and Participants: Individuals ≥65 years old (N = 206,540) who entered 1797 PRAC facilities between January 1, 2009, and December 31, 2016. 

Methods: PI, ascertained from hospitalization records, within 365 days of PRAC entry was the outcome of interest. Individual, medication, facility, system, and health care–related factors were examined as predictors. Prognostic models were developed using elastic nets penalized regression and Fine and Gray models. Area under the receiver operating characteristics curve (AUC) assessed model discrimination out-of-sample. 

Results: Within 365 days of PRAC entry, 4.3% (n = 8802) of individuals had a hospitalization with PI. The strongest predictors for PI risk include history of PIs [sub-distribution hazard ratio (sHR) 2.41; 95% CI 1.77–3.29]; numbers of prior hospitalizations (having ≥5 hospitalizations, sHR 1.95; 95% CI 1.74–2.19); history of traumatic amputation of toe, ankle, foot and leg (sHR 1.72; 95% CI 1.44–2.05); and history of skin disease (sHR 1.54; 95% CI 1.45–1.65). Lower care needs at PRAC entry with respect to mobility, complex health care, and medication assistance were associated with lower risk of PI. The risk prediction model had an AUC of 0.74 (95% CI 0.72–0.75). 

Conclusions and Implications: Our prognostic model for risk of hospitalization with PI performed moderately well and can be used by health and aged care providers to implement risk-based prevention plans at PRAC entry.