Emergency Department-Based Physical Function Measures for Falls in Older Adults and Outcomes: A Secondary Analysis of GAPcare.

This is a secondary analysis of a small trial: 110 people at two hospitals in one US state. The measures were not randomly assigned, so an association with where someone was sent does not show the test caused or predicted that decision. The groups being compared were small, especially the 14 people sent to a nursing home, so the differences reported rest on few patients. Participants were mostly women (67%) and white (94%). The authors say the findings may not apply to other sites and call for more Hispanic and Black participants in future work. Only patients the ED team expected to go home or to assisted living could join, and people who were homeless, in a nursing home, without a phone or with altered mental status were excluded. The study says nothing about those patients. Half the participants were in the GAPcare intervention, with physical therapy and pharmacy input that may have improved their outcomes; the authors acknowledge this and say their reported effect sizes are likely underestimates. The TUG could not be completed by everyone and needs a standard arm chair, which is not in every ED room; patients may also lack confidence in walking after a recent injury. The AM-PAC six clicks needs professional training to administer and its rater reliability has not been tested. Only 46 of the 55 people assigned to the intervention actually received the physical therapy consultation, so the intervention was not delivered as planned to everyone. Larger, longer studies are needed to test whether using these measures actually changes outcomes. This study only found associations, and it was not designed to test whether acting on the scores helps. Half of the participants who were expected to go home were unexpectedly admitted, so the ED clinicians' own expectations were not a reliable guide.

The authors declare that they have no conflicts of interest to report. The text provided does not state who funded the study.

Do not read these results as showing these tests predict a future fall. None of the Timed-Up-and-Go, the AM-PAC six clicks or the Barthel ADL separated people who came back to the ED for another fall from those who did not. The link to where someone was sent is an association in 110 people, not evidence that scoring a patient changes their destination.