Clinical Impact of an Analytic Tool for Predicting the Fall Risk in Inpatients: Controlled Interrupted Time Series

Abstract

Background Patient falls are a common cause of harm in acute-care hospitals worldwide. They are a difficult, complex, and common problem requiring a great deal of nurses’ time, attention, and effort in practice. The recent rapid expansion of health care predictive analytic applications and the growing availability of electronic health record (EHR) data have resulted in the development of machine learning models that predict adverse events. However, the clinical impact of these models in terms of patient outcomes and clinicians’ responses is undetermined. Objective The purpose of this study was to determine the impact of an electronic analytic tool for predicting fall risk on patient outcomes and nurses’ responses. Methods A controlled interrupted time series (ITS) experiment was conducted in 12 medical-surgical nursing units at a public hospital between May 2017 and April 2019. In six of the units, the patients’ fall risk was assessed using the St. Thomas’ Risk Assessment Tool in Falling Elderly Inpatients (STRATIFY) system (control units), while in the other six, a predictive model for inpatient fall risks was implemented using routinely obtained data from the hospital’s EHR system (intervention units). The primary outcome was the rate of patient falls; secondary outcomes included the rate of falls with injury and analysis of process metrics (nursing interventions that are designed to mitigate the risk of fall). Results During the study period, there were 42,476 admissions, of which 707 were for falls and 134 for fall injuries. Allowing for differences in the patients’ characteristics and baseline process metrics, the number of patients with falls differed between the control (n=382) and intervention (n=325) units. The mean fall rate increased from 1.95 to 2.11 in control units and decreased from 1.92 to 1.79 in intervention units. A separate ITS analysis revealed that the immediate reduction was 29.73% in the intervention group (z=–2.06, P=.039) and 16.58% in the control group (z=–1.28, P=.20), but there was no ongoing effect. The injury rate did not differ significantly between the two groups (0.42 vs 0.31, z=1.50, P=.134). Among the process metrics, the risk-targeted interventions increased significantly over time in the intervention group. Conclusions This early-stage clinical evaluation revealed that implementation of an analytic tool for predicting fall risk may to contribute to an awareness of fall risk, leading to positive changes in nurses’ interventions over time. Trial Registration Clinical Research Information Service (CRIS), Republic of Korea KCT0005286; https://cris.nih.go.kr/cris/search/detailSearch.do/16984

Evaluation of an approach to clinical decision support for preventing inpatient falls: a pragmatic trial

“…The fall-prevention CDS intervention was more significantly associated with decreasing injurious falls for patients aged 65 and older. Following previous studies, 4 , 13 the present trial demonstrated that integrating machine-learning predictions and guideline-driven CDS functions into the EMR could help to reduce the considerable physical and cognitive burdens experienced by nurses when they are ensuring patient safety. To improve the effectiveness of this intervention further, combining already-known effective strategies might help to decrease fall rates, such as routine patient and family engagement and other quality improvement tools.…”

“…This was the first fall-prevention clinical study to provide evidence for combining a machine-learning technique and CDS functions for EMR systems. A previous evaluation 13 of the feasibility of fall-risk prediction using machine learning found that changes in nurse behavior resulted in significant improvements in risk-targeted, preventive activities. In the current study, we added more-sophisticated CDS functions to the EMR system to remind nurses about guideline-driven activities tailored to patient-level risk factors.…”

“…Our research team developed a fall-risk prediction model based on electronic nursing data and tested its clinical feasibility and user acceptance in previous studies. 4 , 13 Before moving to expensive large-scale trials, it is necessary to accurately evaluate the performance of the prediction model and coordinate its practice workflow. This study aimed to assess whether the fall-prevention CDS approach using machine learning that stimulates risk-targeted interventions is associated with reductions in the rates of falls and injurious falls among hospital inpatients.…”

Predicting Falls in Long-term Care Facilities: Machine Learning Study

“…Supervised learning trains algorithms based on labeled training data, whereas the unsupervised learning approach does not require labeled training and can find structures within the data. Several EHR-based MLAs have been developed for fall risk predictions in hospitalized patients [ 14 - 18 ]. Few studies have explored the utility of ML approaches for senior residents in community-dwelling or long-term assisted living facilities [ 19 - 23 ].…”

Artificial intelligence for falls management in older adult care: A scoping review of nurses' role

“…Interestingly,Cho et al (2021) was the only study where nurses in clinical practice used an AI-based system across six hospital units for real-world clinical validation. The AI-based fall prediction tool was trialled alongside a standardized falls risk assessment to determine which was more accurate in identifying older patients at risk of falling and to encourage nurses to implement falls prevention strategies, with the intervention group implementing more falls risk targeted interventions.…”