Improved accuracy and efficiency of primary care fall risk screening of older adults using a machine learning approach.

BACKGROUND: While many falls are preventable, they remain a leading cause of injury and death in older adults. Primary care clinics largely rely on screening questionnaires to identify people at risk of falls. Limitations of standard fall risk screening questionnaires include suboptimal accuracy, missing data, and non-standard formats, which hinder early identification of risk and prevention of fall injury. We used machine learning methods to develop and evaluate electronic health record (EHR)-based tools to identify older adults at risk of fall-related injuries in a primary care population and compared this approach to standard fall screening questionnaires. METHODS: Using patient-level clinical data from an integrated healthcare system consisting of 16-member institutions, we conducted a case-control study to develop and evaluate prediction models for fall-related injuries in older adults. Questionnaire-derived prediction with three questions from a commonly used fall risk screening tool was evaluated. We then developed four temporal machine learning models using routinely available longitudinal EHR data to predict the future risk of fall injury. We also developed a fall injury-prevention clinical decision support (CDS) implementation prototype to link preventative interventions to patient-specific fall injury risk factors. RESULTS: Questionnaire-based risk screening achieved area under the receiver operating characteristic curve (AUC) up to 0.59 with 23% to 33% similarity for each pair of three fall injury screening questions. EHR-based machine learning risk screening showed significantly improved performance (best AUROC = 0.76), with similar prediction performance between 6-month and one-year prediction models. CONCLUSIONS: The current method of questionnaire-based fall risk screening of older adults is suboptimal with redundant items, inadequate precision, and no linkage to prevention. A machine learning fall injury prediction method can accurately predict risk with superior sensitivity while freeing up clinical time for initiating personalized fall prevention interventions. The developed algorithm and data science pipeline can impact routine primary care fall prevention practice.

Cost of Inpatient Falls and Cost-Benefit Analysis of Implementation of an Evidence-Based Fall Prevention Program.

Importance: There is insufficient research on the costs of patient falls in health care systems, a leading source of nonreimbursable adverse events. Objective: To report the costs of inpatient falls and the cost savings associated with implementation of an evidence-based fall prevention program. Design, Setting, and Participants: In this economic evaluation, a matched case-control study used the findings from an interrupted time series analysis that assessed changes in fall rates following implementation of an evidence-based fall prevention program to understand the cost of inpatient falls. An economic analysis was then performed to assess the cost benefits associated with program implementation across 2 US health care systems from June 1, 2013, to August 31, 2019, in New York, New York, and Boston, Massachusetts. All adults hospitalized in participating units were included in the analysis. Data analysis was performed from October 2021 to November 2022. Interventions: Evidence-based fall prevention program implemented in 33 medical and surgical units in 8 hospitals. Main Outcomes and Measures: Primary outcome was cost of inpatient falls. Secondary outcome was the costs and cost savings associated with the evidence-based fall prevention program. Results: A total of 10 176 patients who had a fall event (injurious or noninjurious) with 29 161 matched controls (no fall event) were included in the case-control study and the economic analysis (51.9% were 65-74 years of age, 67.1% were White, and 53.6% were male). Before the intervention, there were 2503 falls and 900 injuries; after the intervention, there were 2078 falls and 758 injuries. Based on a 19% reduction in falls and 20% reduction in injurious falls from the beginning to the end of the postintervention period, the economic analysis demonstrated that noninjurious and injurious falls were associated with cost increases of $35 365 and $36 776, respectively. The implementation of the evidence-based fall prevention program was associated with $14 600 in net avoided costs per 1000 patient-days. Conclusions and Relevance: This economic evaluation found that fall-related adverse events represented a clinical and financial burden to health care systems and that the current Medicare policy limits reimbursement. In this study, costs of falls only differed marginally by injury level. Policies that incentivize organizations to implement evidence-based strategies that reduce the incidence of all falls may be effective in reducing both harm and costs.

Classification of Injurious Fall Severity in Hospitalized Adults.

BACKGROUND: Many hospital systems in the United States report injurious inpatient falls using the National Database of Nursing Quality Indicators categories: None, Minor, Moderate, Major, and Death. The Major category is imprecise, including injuries ranging from a wrist fracture to potentially fatal subdural hematoma. The purpose of this project was to refine the Major injury classification to derive a valid and reliable categorization of the types and severities of Major inpatient fall-related injuries. METHODS: Based on published literature and ranking of injurious fall incident reports (n = 85) from a large Academic Medical Center, we divided the National Database of Nursing Quality Indicators Major category into three subcategories: Major A-injuries that caused temporary functional impairment (eg, wrist fracture), major facial injury without internal injury (eg, nasal bone fracture), or disruption of a surgical wound; Major B-injuries that caused long-term functional impairment or had the potential risk of increased mortality (eg, multiple rib fractures); and Major C-injuries that had a well-established risk of mortality (eg, hip fracture). Based on the literature and expert opinion, our research team reached consensus on an administration manual to promote accurate classification of Major injuries into one of the three subcategories. RESULTS: The team tested and validated each of the categories which resulted in excellent interrater reliability (kappa = .96). Of the Major injuries, the distribution of Major A, B, and C was 40.3%, 16.1%, and 43.6%, respectively. CONCLUSIONS: These subcategories enhance the National Database of Nursing Quality Indicators categorization. Using the administration manual, trained personnel can classify injurious fall severity with excellent reliability.