JMIR Publications

ABSTRACT

Background:

Effective fall prevention interventions in hospitals requires appropriate allocation of resources early in admission. Fall risk assessment tools have typically been inaccurate for prediction, ineffective in prevention, and time-consuming to complete.

Objective:

The objective of this study was to develop an accurate and dynamic prognostic model for inpatient falls inpatients using routinely recorded electronic medical records data.

Methods:

We used routinely recorded data from five Australian hospitals to develop and internally-externally validate a prediction model for inpatient falls using a Cox proportional hazards model with time-varying covariates. The study cohort included patients admitted during 2018-2021 to any ward, with no age restriction. Predictors used in the model included admission-related administrative data, length of stay and number of previous falls during the admission (updated 12-hourly up to 14 days post-admission). Model calibration was assessed using Poisson regression and discrimination using area under the time-dependent receiver operating characteristic curve.

Results:

There were 1,107,556 inpatient admissions, 6,004 falls and 5,341 unique fallers. The area under the time-dependent receiver operating characteristic curve was 0.899 (95% CI: 0.883 – 0.914) at 24-hours post admission and declined throughout admission (e.g., 0.765; 95% CI: 0.753 – 0.778 at 7th day post admission). Site-dependent over- and under- estimation of risk was observed on the calibration plots.

Conclusions:

We developed and internally-externally validated a prognostic model for inpatient falls with high discrimination. The model has potential for operationalization in clinical decision support for inpatient fall prevention. Performance was site dependent and model recalibration may lead to improved performance.