Comparison of two fall risk assessment tools (FRATs) targeting falls prevention in sub-acute care

Abstract

FRATs are designed to identify both persons at high risk of falls and to allow for cost-effective targeting of fall prevention strategies. This study compares two FRATs (BHS FRAT and TNH-STRATIFY) for accuracy of predicting falls and targeting of fall prevention strategies in a sub-acute hospital. Comparisons of retrospective audit data over two periods (use of the BHS-FRAT; post TNH-STRATIFY implementation) were used in the evaluation (n=362). Inter-rater reliability of the TNH-STRATIFY was evaluated from independent assessment by two nurses for 30 sub-acute patients and using intraclass correlation coefficient (ICC2,1). Event rate (ER) and standard measures of predictive accuracy were calculated for both FRATs. The proportions of patients with documented fall prevention strategies addressing identified fall risk factors were compared between audit phases. The TNH-STRATIFY had high inter-rater reliability (ICC2,1=0.96). The BHS-FRAT and TNH-STRATIFY demonstrated poor predictive accuracy using recommended high risk cut-off scores, with low specificityER (0.07 and 0.13 respectively) and very low Youden IndexER (0.04 and 0.07 respectively), although these measures improved using modified cut-off scores. Positive and negative predictive values were moderate for the BHS-FRAT (0.51, 0.64) and TNH-STRATIFY (0.52, 0.61). The falls rate and proportion of recurrent fallers did not change between audit phases. Implementation rates for prevention strategies for key risk factors were higher following implementation of the TNH-STRATIFY. The results indicated that the TNH-STRATIFY, combined with associated nursing care plan falls documentation, improved the targeting of prevention strategies for key risk factors such as cognitive impairment, incontinence and mobility impairment.