Clinical importance of an intervention must reside with the patient.

We thank Haines and colleagues for their interest in our Editorial. While we agree with much of what they say, we feel that their letter misses the point of our Editorial. So we are grateful for the opportunity to make our point more clearly.

We argued that an intervention is clinically important if it has effects that are large enough to make the costs, inconvenience, and harms associated with intervention worthwhile (Ferreira and Herbert 2008). Decisions about clinical importance need to be made by patients with the assistance of health professionals because they need to take into account how patients value both the costs and benefits of intervention. We reiterate that distribution- and anchor-based methods do not help physiotherapists to make clinical decisions because they do not provide a voice to the person who ultimately must agree to the intervention. Ultimately clinical importance must be assessed from the perspective of the person who is seeking health care.

Haines et al argue that decisions about whether, when, and how much to provide of a health care intervention have societal impacts, so they should be based on societal perspectives. The benefit-harm trade-off method cannot provide a societal perspective so it cannot be used in isolation to inform decisions about whether, when, and how much to provide of a health care intervention. Unlike the benefit-harm trade-off method, anchor-based and distribution-based methods can provide data that can be used to model societal impacts.

We acknowledge that it is necessary to consider societal impacts in many contexts and, when that is the case, distribution- and anchor-based methods can provide data complementary to information about clinical importance. These data can inform decisions by service providers about the allocation of health care resources. In other contexts it might not be necessary to consider societal perspectives. For example, private care providers arguably need to have little interest in the societal impacts of an intervention.

An important limitation of the anchor-based approach was not acknowledged by Haines and colleagues. Anchor-based approaches cannot inform decisions about whether the effect of an intervention is clinically important, because effects of intervention can only be understood in terms of between-group differences. In contrast, anchor-based approaches attach a meaning to the change in outcome patients experience over time. Using anchor-based approaches to make decisions about the clinical importance of effects of intervention involves using estimates derived from within-group changes to interpret between-group differences. This could be very misleading.

References

Ferreira ML, Herbert RD (2008) Aust J Physiother 54: 229-230.