Predictive Performance of a Fall Risk Assessment Tool for Community-Dwelling Older People (FRAT-up) in 4 European Cohorts.
Pierpaolo Palumbo, Jochen Klenk, Luca Cattelani and 5 others
PMID 27594522WHAT IT FOUND
FRAT-up modestly predicted falls in community-dwelling older adults across four European cohorts, but accuracy varied widely.
It overestimated risk in two cohorts, so it gives only a modest risk signal.
Key findings
01FRAT-up AUCs were 0.562, 0.699, 0.636, and 0.685 in ActiFE, ELSA, InCHIANTI, and TILDA, and the pooled mean AUC was 0.646.
02FRAT-up overestimated risk consistently in ELSA and TILDA, and in ActiFE and InCHIANTI low-risk participants had more falls than expected while high-risk participants had fewer.
03Most single risk factors were significantly associated with subsequent falls in ELSA, InCHIANTI, and TILDA, but only 6 odds ratios were significant in ActiFE.
STILL TO COME
How it was doneWhat they foundWhat it means for PTsWhat it means for RNs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The four cohorts did not measure falls the same way. ActiFE used prospective fall calendars for 12 months, InCHIANTI used retrospective reports about the previous 12 months, and ELSA and TILDA used retrospective reports over 2 years. Harmonization was imperfect. Five risk factors could not be constructed in ELSA and three in TILDA, and the fall outcome was imperfectly harmonized in all datasets except ActiFE. Missing data were imputed, and totally missing variables were replaced with FRAT-up prevalence rates. There was very high heterogeneity among cohorts, with I2 95.1%, so a single pooled performance estimate may not apply to all populations. FRAT-up was not well calibrated. It overestimated risk in ELSA and TILDA and showed low resolution in ActiFE and InCHIANTI. The study validated prediction accuracy. It did not test whether using FRAT-up changes care or reduces falls.
The easy way to misread this
Do not conclude that FRAT-up prevents falls or should replace clinical judgment. This was a prediction-tool validation, not an intervention trial, and the pooled AUC was 0.646 with risk overestimated in two cohorts.