Automated Registration of Falls among Older Adults in a Statewide Falls Prevention Evaluation

Background Falls are a central problem in old age. However, retrospective reports and diaries miss many falls. We implemented an interactive voice recognition system (IVR), in which people completed automated telephone interviews every 30 days to assess falls, physical activity, hospitalization, and emergency department admissions. Methods Falls Free PA enrolled 1857 older adults in 2010-2011 from 21 Pennsylvania counties. We examined compliance with the monthly falls registration interview as well as measures of construct and criterion validity. While the 12-month follow-up is ongoing, the sample has completed a median of 6 monthly follow-up interviews.

Results The mean age of the sample was 75 (± 8.5). Women were 72.9% of the sample, whites 88%, and Spanish speakers 2.9%. 61% had not attended college. The IVR calls, consisting of 6 questions, lasted a mean of 2.5 min. Compliance was excellent: 71% of participants completed more than 80% of scheduled calls and 60% completed all calls. Over follow-up, 27.1% reported at least one fall, suggesting higher incidence for falls than that established with other methods. People reporting falls were significantly more likely to report hospitalization (18.5% vs. 9.6%, p < .001), emergency department admissions (28.8 vs. 12.7%, p < .001), and lower rates of physical activity (4.7 vs. 5.0 days per week with 30+ min of moderate physical activity, p < .001).

Conclusions/Discussion The interactive voice recognition system is an effective way of ascertaining falls in a diverse sample of older Americans.