Slow Processing Speed Predicts Falls in Older Adults With a Falls History: 1‐Year Prospective Cohort Study

Jennifer C. Davis, John R. Best, Karim M. Khan, Larry Dian, Stephen R. Lord, Kim Delbaere, Chun Liang Hsu, Winnie Cheung, Wency Chan, Teresa Liu‐Ambrose

Journal of the American Geriatrics Society · 2017 · 56 citations · 51 references

Concepts

Abstract

BACKGROUND/OBJECTIVES: A previous fall is a strong predictor of future falls. Recent epidemiologic data suggest that deficits in processing speed predict future injurious falls. Our primary objective was to determine a parsimonious predictive model of future falls among older adults who experienced ≥1 fall in the past 12 months based on the following categories: counts of (1) total, (2) indoor, (3) outdoor or (4) non-injurious falls; (5) one mild or severe injury fall (yes vs no); (6) an injurious instead of a non-injurious fall; and (7) an outdoor instead of an indoor fall. DESIGN: 12-month prospective cohort study. SETTING: Vancouver Falls Prevention Clinic, Canada (www.fallsclinic.ca). PARTICIPANTS: Two-hundred and eighty-eight community-dwelling older adults aged ≥70 years with a history of ≥1 fall resulting in medical attention in the previous 12 months. MEASUREMENTS: We employed principal component analysis to reduce the baseline predictor variables to a smaller set of five factors (i.e., processing speed, working memory, emotional functioning, physical functioning and body composition/fall risk profile). Second, we used the extracted five factors as predictors in regression models predicting the incidence of falls over a 12-month prospective observation period. We conducted regression analyses for the seven falls-related categories (defined above). RESULTS: Among older adults with a falls history, processing speed was the most consistent predictor of future falls; poorer processing speed predicted a greater number of total, indoor, outdoor, and non-injurious falls, and a greater likelihood of experiencing at least one mild or severe injurious fall (all P values < .01). CONCLUSION: Poorer performance on the processing speed factor, a trainable factor, was independently associated with the most costly type of falls-injurious falls.

References

51