Published Papers - Abstract 316

Heesch K, Byles J & Brown W. Does physical activity decrease the risk of falls in older women? , 2006; : 250

Few studies have prospectively assessed among older women the role of physical activity (PA) in preventing falls. Studies that have done so have relied on small samples and short follow ups. PURPOSE: To prospectively examine in older women the dose-response relationship between PA and risk of falls. METHOD: Data were collected from a cohort of older women (69-77 years; N=10,882) who completed mailed surveys as part of the Australian Longitudinal Study on Women's Health. Data on PA and other potential risk factors were collected in 1996. In 1999, the women were asked if they had had a fall to the ground in the last 12 months (yes/no). PA was measured with National Heart Foundation questions and categorized based on weekly frequency as: none, very low; low; moderate; and high. RESULTS: In a univariate logistic regression model, PA at any level above “none” reduced the risk of falls (odds ratio [OR] = 0.84, 0.81, 0.76, 0.60, respectively, p<.05). When demographic variables (country of birth, area of residency, education) were added to the model, the association between PA and falls remained (p<.05). When potential confounders (alcohol use, number of chronic health problems, number of adverse life events, bladder incontinence, experiencing elderly abuse, medication use) were added, only “high” PA decreased the risk of falls (OR = 0.67, p=.02). When potential mediators (BMI, past falls, and SF-36 physical and mental health scores) were added, PA no longer reduced the risk of falls (p>.05). CONCLUSION: After adjusting for demographic and health-related variables, only a high level of PA reduces the risk of having a fall in older women. This association disappears once potential mediators are included. The indirect influence of PA on falls via mediators of this association needs to be examined to better understand the association between PA and falls among older women.