Over 150 potentially low-value health care practices: an Australian study

To the Editor: We read the recent study by Elshaug and colleagues1 with interest, but were concerned by the inclusion of radiotherapy for spinal cord disease as a potentially low-value intervention. The lack of clinical review in the identification process appears to be the key driver in coming to this erroneous conclusion.

Elshaug and colleagues recognise the need for clinical expertise in assessing trials demonstrating no difference between a health care service and an active comparator by excluding these from analysis. However, we suggest that the application of this expertise should extend to all publications included in the analysis. Of the three publications identifying radiotherapy for spinal cord compression, none disputed its role in managing this condition.2-4 One randomised trial, assessing the addition of decompressive surgery to radiotherapy, supported the role of radiotherapy for this disease.2 The intervention being assessed was, in fact, surgery, but this was not recognised.

The study by Elshaug et al also presented a list of 13 services which were identified by more than one search strategy and flagged as being potential priority candidates for review. Radiotherapy for spinal cord compression was again identified by this “triangulation” strategy, but two review articles3,4 were flagged by referencing the same randomised trial,2 essentially “double dipping”.

We commend the authors on attempting to develop a strategy to identify potential low-value interventions, but this process cannot be divorced from clinical judgement. Indeed, this is the entire raison d’être for expert working groups and consensus guidelines, which evaluate the quality and validity of trials in a clinical context.