Screening tool for predicting patient aggressive behavior and staff injury at a pediatric hospital

Few clinical tools predict patients at the highest risk for seclusion/restraint events (SREs) or staff injury due to aggressive patient interactions (APIs). We examined the utility of a "high-risk notification" (HRN) tool to proactively identify patients at admission at greatest risk of SREs and APIs. We also assessed how initial SRE events influence subsequent risk.

Using the HRN tool, 2166 patients (61.3% female; mean age 13.5 years) across 2969 admissions were classified as either HRN + or HRN-. We calculated SRE and API incidence rates per 100 patient-days and estimated the proportion of SREs and APIs attributable to HRN + status at admission. To examine the influence of SRE/API on subsequent events, we used logistic regression with random hospital-stay effects, modeling daily risk of SREs and APIs as a function of SREs or APIs on previous days.

HRN + patients made up 5.6% of admissions but accounted for >50% of APIs (9.55 per 100 days) and >45% of SREs (28.7 per 100 days). HRN- patients (94.4% of admissions) experienced significantly lower rates (1.56 APIs and 5.69 SREs per 100 days). Both HRN+ and HRN- patients had significantly increased risk of SRE and API after their first event. HRN + patients with prior events were 19 and 26 times more likely to experience another SRE or API, respectively. Among HRN- patients, over 90% of subsequent events followed an initial SRE/API event, while 67% of subsequent events for HRN + patients were attributable to a first event.

The HRN tool identified <5% of patients as responsible for ∼50% of SREs and APIs and analyses strongly indicate that preventing initial SRE/API significantly reduces repeated events for both high- and low-risk patients. The HRN tool shows promise in predicting and preventing SREs and APIs, allowing for targeted safety interventions that could reduce recurring events and improve staff and patient safety.