- 1School of Nursing and Department of Epidemiology and Community Medicine, University of Ottawa
- 2Élisabeth Bruyère Research Institute, Ottawa, Ontario, Canada
The commentary by Griffiths regarding the Cochrane review on the effectiveness of hip protectors for reducing hip fractures published in the January 2005 issue of Evidence-Based Nursing1 highlights important methodological considerations that may account for the failure of this review to demonstrate a reduction in fractures among elderly people living in residential care. Although we concur with Griffiths’ conclusion that it is premature to advise patients to stop using hip protectors, we would like to advance 3 other important considerations.
Firstly, falls among older adults are often multifactorial, and therefore, prevention requires a multiple intervention approach.2 That is, multiple strategy interventions that target risk factors for falls at individual, health provider, and organisational levels should be offered. If hip protectors are offered as part of a multiple intervention programme for falls prevention, elderly people may be reminded that hip protectors alone do not prevent falls and perhaps dissuaded from developing a false sense of confidence when wearing hip protectors and taking undue risks. Secondly, a review of the studies included in the Cochrane review by Parker et al indicated that many of the staff training interventions described were brief and not informed by appropriate theory. The study by Meyer et al3 was an exception. These authors described a theory based, multilevel intervention. Their tailored education sessions to promote the use of hip protectors by staff illustrates the intense training that may be required to effect a change in staff behaviour and subsequently, patient use of hip protectors. Finally, when clinicians advise elderly people to use hip protectors, they need to be cognisant of several issues: hip protectors with an energy shunting mechanism provide the most protection;4 hip protectors must be worn correctly in order to provide protection from hip fractures; and reinforcing messages are needed to encourage their continued use. The review by Parker et al and the commentary by Griffiths remind us that introducing a new innovation into clinical practice is a complex undertaking that must be supported by well designed multiple intervention programmes.
When reviewing evidence, it is crucial to remember the maxim that absence of evidence of effect is not the same as evidence of no effect. Those who criticise evidence-based practice often remind us of the need to examine evidence in the round and not take too narrow a focus. These are important reminders. This wider view and considered judgment is reflected in the fact that neither the review by Parker et al,1 nor my commentary,2 simply dismissed the use of hip protectors based on an overall lack of evidence of benefit.
However, absence of evidence is absence of evidence. The important methodological consideration I noted, which is highlighted by Edwards and Fahey, is that in institutional settings, any benefit of hip protectors might be masked by contamination of any control group in individually randomised controlled trials. Thus, there is no a priori reason for setting the results of individually randomised controlled trials above cluster randomised controlled trials. Hence the conclusion, despite the absence of evidence in individual randomised trials, that “data from cluster randomised studies provide some evidence of effectiveness of hip protectors in reducing the risk of hip fractures in those living in nursing homes and considered to be a high risk of hip fractures”.1
This consideration does not apply in community settings, where the opportunity for contamination of the intervention is low. So when Edwards and Fahey suggest that “…if hip protectors are offered as part of a multiple intervention programme for falls prevention, elderly people may be reminded that hip protectors alone do not prevent falls and perhaps dissuaded from developing a false sense of confidence when wearing hip protectors and taking undue risks…” they might well be correct. They are free to speculate, and the mechanism seems at least plausible. But there is no evidence to support this assertion for community dwelling older people because there is no evidence of benefit. Furthermore, there is no evidence to support this as a mechanism for any benefit among those in institutions to whom (I speculate) it is least likely to apply. Specifically there is no overall evidence that the incidence of falls is reduced.1
The study by Meyer et al3 is singled out by Edwards and Fahey because of the way the intervention was delivered. In particular, they draw attention to the fact that there was a description of tailored educational support for nurses in the intervention sites. But use of protectors in this study, at 34%, was lower than in many of the studies reported by Parker et al. Even when considering the least generous estimate of compliance from other studies, this falls below the median. However, it should be noted that in the study by Meyer et al, a higher proportion of people who fell were users of protectors. Even so, only 58% of falls occurred when a protector was worn. Even after intensive training of staff, a significant level of non-compliance remains.
This brings me to the final point made by Edwards and Fahey. They state that when advising elderly people, clinicians must remember that “…. hip protectors must be worn correctly in order to provide protection from hip fractures…” and “hip protectors with an energy shunting mechanism provide the most protection…” Perhaps this is a reasonable “best bet” when advising those who choose to use them. However, clinicians must also bear in mind that the evidence of benefit is limited to laboratory and institutional settings and that there is considerable evidence that current devices are simply not acceptable to many clients.
Evidence-based health care has been accused in the past of ignoring patient preference in the face of evidence of effectiveness from trials. Surely then, it is hard to condone ignoring preference in the absence of such evidence? Thus, I would like to reiterate my main conclusion about hip protectors. There seems little evidence to justify their use outside of properly designed research programmes. Practitioners would do better to concentrate their efforts on implementing interventions known to prevent falls.