Background Falls are a major cause of disability in older age and fear of falling (FOF) may limit physical activity. Associations between falls, fear of falling and objectively measured physical activity (PA) has been little studied, particularly in community-dwelling older adults.

Methods The study sample was men participating in a prospective, population-based cohort study in 24 British towns provided questionnaire information on a history of falls in the previous year, fear of falling, and other medical history and wore an Actigraph GT3X accelerometer for 7 days. Uniaxial accelerometer data were analysed in 60s epochs, excluding bouts of >60 minutes of complete inactivity. Days with >=600 minutes valid wear time were included. PA intensity was categorised as sedentary: <100 counts/minute, light: 100–1952 counts/minute and moderate to vigorous PA (MVPA):>1952 counts/minute. Associations between activity levels and falls, or fear of falling, were estimated using hierarchical linear regression models accounting for clustering and controlling for confounders.

Results A total of 1543 men with a mean age of 77(range 71–91y) wore the Actigraph; 1100 (72%) had >600 minutes wear time on 5–7 days and questionnaire data. 21% (n=229) had fallen, 30% (n=68) received medical attention. Men who fell took fewer steps/day than men who did not: mean difference (adjusted for age, day order, month, wear time and town) –504 (95%CI –216, –792) and spent more minutes sedentary 13(95%CI 4, 22), and less in light –10(95%CI –1, –19) and MVPA –3.5(95%CI –1.3, –5.8). In relative terms, the reduction in MVPA was greatest: 22%(95%CI 7, 37). Differences were explained by exercise self-efficacy and mobility limitations, but not by exercise outcome expectation. 13% (n=141) men reported FOF, of whom 53% (n=74) had fallen in the past year. Men with FOF took markedly fewer steps than men without:–1325(95%CI –1646, –1005), spent more minutes in sedentary 29(95% 18, 40), and less in light activity –21(95%CI –32, –11) or MVPA –9(95%CI –11, –7). In relative terms, the reduction in MVPA was greatest:–72%(95%CI –93, –52). Differences were partly explained by exercise self-efficacy, mobility limitations, and to a lesser extent, exercise outcome expectation. There was no evidence that FOF had a greater impact on PA levels among men who had fallen compared to those who had not fallen.

Conclusion PA levels and in particular step counts and MVPA levels, were lower among men who had fallen in the past year, and even more markedly, among those who reported FOF.