Background: Multiple factors predict unplanned presentations to the Emergency Department (ED) or hospital admissions for patients with diabetes. We aimed to determine the contribution of clinical and psychosocial/self-management factors towards the risk of hospitalisation by using the Hospital Admission Risk Profile (HARP) tool (developed in Victoria, Australia).
Methods: The HARP tool was completed for patients attending the Sydney Diabetes Health Assessment Unit at Royal North Shore Hospital, Sydney in 2013. The tool includes a clinical assessment score (part A) and psychosocial/self-management impact score (part B) and categorises patients into low, medium, high or urgent risk for hospitalisation.
Results: The cohort comprised 285 patients (mean ± SD age 65 ± 13 years; 62.8% male; 95.8% T2D; diabetes duration 10.6 ± 6.5 years). Most patients had a low risk classification (75.8%; n = 216), fewer had a medium risk (24.2%; n = 69) and none had high or urgent risk. 20.3% of patients with medium risk scores, compared with 9.3% of patients with low risk scores, presented to ED or had an unplanned hospitalisation in the 12 months following HARP tool completion (p = 0.018). A medium HARP score was associated with a 2.5 fold increased risk of ED presentation or unplanned hospital admission compared to a low HARP score (95% CI: 1.193 to 5.061). Part A of the HARP score was significantly higher for hospitalised patients compared to non-hospitalised patients (13.5 ± 6.6 vs. 10.4 ± 5.1; p = 0.0082), whereas there was no difference in Part B scores (3.8 ± 3.9 vs. 3.5 ± 3.3; p = 0.9).
Conclusions: In patients with low and medium HARP scores, clinical features are more predictive of hospitalisation than psychosocial/self-management factors. Further studies are required to determine whether psychosocial/self-management factors are associated with unplanned hospitalisation in patients with higher HARP scores.