Local adaptation and evaluation of a falls risk prevention approach in acute hospitals

Abstract

Objective: To determine whether locally adapting a falls risk factor assessment tool results in an instrument with clinimetric properties sufficient to support an acute hospital's falls prevention program. Design: Prospective cohort study of predictive validity and observational investigation of intra- and inter-rater reliability. Setting Acute wards in two large hospitals in Melbourne, Australia. Participants: One hundred and thirty acute hospital inpatients participated in the predictive accuracy evaluation, with 25 and 35 inpatients used for the intra-rater and inter-rater reliability analyses, respectively. Intervention(s): To develop a falls risk screen and assessment instrument through local adaptation of an existing tool. Clinimetric property analysis of new instrument (Western Health Falls Risk Assessment, WHeFRA) and comparison with ‘gold standard tool’ (STRATIFY). Main Outcome Measures: Fallers, falls and falls per 1000 bed days. Sensitivity (Sens), specificity (Spec), Youden Index (YI) and these three statistics based on event rate of falls (SensER, SpecER and YIER), were calculated to determine predictive accuracy. Reliability was determined using intraclass correlation coefficient (ICC), weighted kappa and signed rank test. Results: Seven participants (5.4%) fell, with 14 falls (fall rate: 10.7 falls per 1000 patient bed days). The WHeFRA instrument was significantly more accurate at predicting fallers and the rate of falls than the STRATIFY. Intra-rater reliability ICC (95% confidence intervals) for WHeFRA screen was 0.94 (0.86–0.97) and inter-rater reliability was 0.78 (0.61–0.88). Conclusions: Local adaptation of an existing tool resulted in an instrument with favorable clinimetric properties and may be a viable procedure for facilitating falls prevention program development and implementation in acute hospital settings.