Predicting Falls in Nursing Homes: A Prospective Multicenter Cohort Study Comparing Fall History, Staff Clinical Judgment, the Care Home Falls Screen, and the Fall Risk Classification Algorithm

Abstract

ObjectivesTo evaluate and compare the predictive accuracy of fall history, staff clinical judgment, the Care Home Falls Screen (CaHFRiS), and the Fall Risk Classification Algorithm (FRiCA). DesignProspective multicenter cohort study with 6 months' follow-up. Setting and ParticipantsA total of 420 residents from 15 nursing homes participated. MethodsFall history, clinical judgment of staff (ie, physiotherapists, nurses and nurses' aides), and the CaHFRiS and FRiCA were assessed at baseline, and falls were documented in the follow-up period. Predictive accuracy was calculated at 1, 3, and 6 months by means of sensitivity, specificity, positive and negative predictive value, positive and negative likelihood ratio, Youden Index, and overall accuracy. ResultsIn total, 658 falls occurred and 50.2% of the residents had at least 1 fall with an average fall rate of 1.57 (SD 2.78, range 0-20) per resident. The overall accuracy for all screening methods at all measuring points ranged from 54.8% to 66.5%. Fall history, FRiCA, and a CaHFRiS score of ≥4 had better sensitivity, ranging from 64.4% to 80.8%, compared with the clinical judgment of all disciplines (sensitivity ranging from 47.4% to 71.2%). The negative predictive value (ranging from 92.9% at 1 month to 59.6% at 6 months) had higher scores for fall history, FRiCA, and a CaHFRiS score of ≥4. Specificity ranged from 50.3% at 1 month to 77.5% at 6 months, with better specificity for clinical judgment of physiotherapists and worse specificity for FRiCA. Positive predictive value ranged from 22.2% (clinical judgment of nurses' aides) at 1 month to 67.8% at 6 months (clinical judgment of physiotherapists). Conclusions and ImplicationsNo strong recommendations can be made for the use of any screening method. More research on identifying residents with the highest fall risk is crucial, as these residents benefit the most from multifactorial assessments and subsequent tailored interventions.