To Use Physical Restraints Or Not?

No one likes to use physical restraints. Physical restraints limit mobility and can have related adverse effects. However, they are utilized in the hope of preventing more serious adversities, particularly falls. The authors of a paper in this issue of the Journal have been leaders in increasing the sensitivity of nursing home providers and others to the undesirable side effects of restraints as well as to their limitations as a fall prevention strategy.12 Building on research by Tinetti and others, this article provides further compelling evidence that restrained residents still experience falls. The issue of restraint use was approached earlier by Tinetti et al., who, for a period of more than 1 year, studied 397 mobile nursing home residents who were not restrained at baseline.34 They found that 122 became restrained, most on an intermittent basis. Serious fall-related injuries were experienced much more commonly in the residents who became restrained compared with those who remained unrestrained (17% vs 5%, respectively; P < .001). Of the 607 falls that occurred, 57% were experienced by restrained residents; the mean number of falls increased progressively from 1.0 (±1.6) in nonrestrained residents to 1.9 (±2.0) and 3.6 (±3.7) in intermittently and continuously restrained residents, respectively. For the restrained residents who fell, only 10% of the falls occurred while the resident was actually restrained or immediately after removal of restraints. Most falls in restrained residents happened during unrestrained periods, suggesting that if restraints increase falls and related injuries, they do so indirectly. The authors point out that restraint use probably worsens deconditioning, gait, and balance abnormalities, thereby increasing their fall and injury risk. Another plausible explanation (which does not exclude the first) is that restraint use is a marker for residents at high risk for falls. In this issue of the Journal, the article by Capezuti et al. explores this question further.5 They followed 332 residents in three nursing homes over a period of 9.5 months, and examined correlations between use of physical restraints, falls, and fall-related injuries. Residents were classified on the two variables most likely to affect fall and safety risk: cognitive function and ambulatory status. Cognitive function was assessed using the Mini-Mental State Exam (MMSE), with residents dichotomized as nonconfused (18 or more of 30) or confused (17 or less). The importance of cognitive function in this context is how it affects safety awareness and judgment. This may not be adequately assessed by the MMSE, and an arbitrary cutoff score of 17 may also not capture critical differences in awareness and judgment. Similarly, the categorical measure of ambulatory status used in this study (ambulatory vs nonambulatory-chair fast) is limited in its ability to detect important differences in injury risk within categories. For example, it would not be surprising if the confused ambulatory residents who were restrained were more deconditioned and had more severe balance and gait problems than the confused ambulatory residents who were not restrained. Thirty percent of restrained residents were classified as ambulatory compared with 71% of unrestrained residents. Overall, 107 restrained residents (89.9%) and 84 nonrestrained residents (41.4%) were classified as confused. Mean MMSE scores were 5.1 for restrained and 17.6 for nonrestrained residents. As would be expected, the subset of confused and ambulatory residents (n = 89) experienced a significantly higher number of falls when compared with all other residents, including those who were not at risk for being restrained (n = 233). Restraints within the ambulatory confused group were associated with more falls and recurrent falls in this comparison. However, restraints were associated significantly with decreased falls and recurrent falls in nonambulatory residents in other comparisons that were done. Among all residents, restraints were used primarily in the confused nonambulatory (72/102 (71%)), confused ambulatory (35/89 (39%)), and nonconfused nonambulatory subgroups (11/39 (28%)). When the effect of restraints on falls was looked at in these three subgroups the only significant association was decreased recurrent falls in the confused nonambulatory subgroup. What conclusions can be drawn from this observational study? The author's main finding is that “restraint use is not significantly associated with a consistent increase or decrease in fall/injury risk among subgroups of nursing home residents likely to be restrained.” Clearly, the use of physical restraints did not eliminate falls and fall-related injuries. Based on these data, the argument is made that alternatives to restraints should be used. This argument is very easy to support, even if one believes that the restrained residents in the study were at higher risk for injury than the unrestrained residents with whom they were compared. We believe that most people, including nursing home providers, prefer not to use physical restraints even if they hold to the belief that physical restraints have some fall prevention benefits. The restraint-free movement and the publicity that it has received deserves credit for creating this new preference. We also believe that providers would act on this preference for alternative strategies more often if they believed that the suggested alternatives to restraints would work, that they were more satisfactory to residents and their families, or that they were described with sufficient specificity so that they could be implemented. The alternatives suggested in this article and by others need to be evaluated on these dimensions. Restraint alternatives that are often mentioned are reclining chairs, cushions, and other devices that limit mobility but that, presumably, are more comfortable, permit more freedom of movement, or are more satisfactory to residents. An industry has emerged to sell these products without evidence that residents or families find these alternatives more satisfactory than seat belts, which seem to be the most popular type of restraints. Evidence that residents “do better” with these devices according to any measure, and that residents have more freedom of movement, is needed before we continue to make investments in this area. Such evidence would increase providers' motivation to engage in this form of restraint reduction. The concept of individualized care, which involves an assessment of individual injury risk and rehabilitation, is also a highly attractive alternative to restraints. Unfortunately, clinical research has not demonstrated how to accomplish this type of rehabilitation, and it is unfair to expect that nursing home providers can create clinically effective assessment and rehabilitation protocols with their limited resources. A rehablitative intervention to improve injury risk will be particularly challenging to develop because of the severe degree of cognitive impairment and high injury risk that characterize restrained residents. Comparisons of injury risk between small samples of ambulatory restrained and unrestrained residents have shown that the restrained residents are significantly more demented, show more safety judgment problems, and have a much higher overall risk for injury.6 The issue of who will pay for these individualized rehabilitation efforts must also be addressed. For example, we do not think that physical or occupational therapy services will be approved by third party payers for the very demented and long-term stay residents who are typically restrained. We are also unsure that nursing departments within nursing homes have resources to implement what will likely prove to be very difficult and labor intense rehabilitative intervention protocols, even if the protocols developed are described specifically and proven to be clinically effective. We doubt that further efforts to prove that restraints are associated with falls will change nursing home care more than it has already been changed by articles such as the Capizio et al. article. What is needed to move care to the next higher level is hard evidence that the restraint alternatives that are suggested are better and can be done within the resource limitations of nursing homes. The organization of this evidence is the primary responsibility of the clinical research community.