Physical restraint use and falls in nursing home residents

OBJECTIVE: To examine the relationship between restraint use and falls while controlling for the effect of psychoactive drug use among nursing home residents, including subgroups of nursing home residents with high rates of restraint use and/or falls. DESIGN: Secondary analysis of data from a longitudinal clinical trial designed to reduce restraint use. SETTING: Three nursing homes. PARTICIPANTS: Subjects (n = 322) were either restrained (n = 119) or never restrained (n = 203) at each observation point during a 9.5-month data collection period that preceded the intervention phase of the clinical trial. MEASUREMENTS: We evaluated restraint status (independent variable) three times during the data collection period by direct observation over a 72-hour period. Incident reports documenting falls and fall-related injuries (dependent variables) were reviewed. Cognitive status was measured using the Folstein Mini-Mental State Exam and functional status (including ambulation status) by the Psychogeriatric Dependency Rating Scale. Psychoactive drug use profile was obtained through record review. MAIN RESULTS: Using multiple logistic regression, we compared the effect of restraint use on fall risk between a confused ambulatory subgroup and the remaining sample and found a significant difference in the odds ratio for falls and recurrent falls (P = .02; chi-square = 5.24, df = 1; P = .003, chi-square = 9.12, df = 1). In the confused ambulatory subgroup, restraint use was associated with increased falls (odds ratio: 1.65, 95% CI: 0.69, 3.98) as well as recurrent fall risk (odds ratio: 2.46, 95% CI: 1.03, 5.88). Increased falls and recurrent fall risk was not observed in the remaining sample (falls odds ratio: 0.49, 95% CI: 0.28,0.87; recurrent falls odds ratio: 0.42, 95% CI: 0.20,0.91). One subgroup, the nonconfused ambulatory residents, were never restrained; after removing this subgroup, the confused ambulatory continued to be associated, though not significantly, with a higher risk of falls and injuries. Only nonconfused nonambulatory restraints were associated with a lower risk of all three outcomes: falls (odds ratio: 0.28, 95% CI: 0.05, 1.58), recurrent falls (odds ratio: 0.48, 95% CI: 0.05, 4.72), and injurious falls (odds ratio: 0.42, 95% CI: 0.04, 4.01); these results, however, were not statistically significant. There was no evidence that the effect of restraint use on fall risk depended upon the use of psychoactive drugs (chi square = 4.43; df = 2, P = .11). CONCLUSION: Restraints were not associated with a significantly lower risk of falls or injuries in subgroups of residents likely to be restrained. These findings support individualized assessment of fall risk rather than routine use of physical restraints for fall prevention. Researchers and clinicians should continue to focus efforts on developing a variety of approaches that reduce risk of falls and injuries and promote mobility rather than immobility.