Characteristics and predictors of falls in elderly patients

Abstract

Background: Poor performance on mobility testing is one of a number of factors associated with increased falls in community-dwelling elderly. The significance of these associations has not previously been tested in a sample drawn exclusively from a primary care practice.

Methods: This 1-year prospective study recorded falls, fall injuries, and related factors in 120 ambulatory geriatric outpatients of a family medicine practice. The association of mobility score, physician's estimate of mobility score, physician's estimate of likelihood to fall, and other fall risk factors was assessed with whether participants fell. Subjects recorded falls and injuries on weekly postcards. Follow-up by telephone was done to ensure compliance.

Results: Thirty-seven (36%) of the 102 participants who completed the study fell once or more. There were 56 total falls, of which 27 (48%) caused injuries. Thirty-six (64% of total) falls occurred in or around the subject's home, and significantly more (chi 2 = 10.93, P less than .001) of these falls had intrinsic causes compared with falls away from home. Prestudy history of subject's falls was significantly associated with subjects' falls during the study, although its sensitivity was only 41%. All other factors studied, including mobility score and the physician's estimates, were not significantly associated with fall status.

Conclusions: This study did not support the use of risk-factor determination to select primary care patients who should be assessed further for fall risk. The high prevalence of falls and injuries in this sample suggests that all elderly patients should be given fall prevention advice.