OUP user menu

Not a Placebo, but Is It Effective?

Nikolai Bogduk MD
DOI: http://dx.doi.org/10.1111/pme.12052 315-316 First published online: 1 March 2013

Randomized, placebo-controlled studies are a rare commodity in interventional pain medicine. Pain Medicine, is therefore, pleased to publish such a study, in which Kappural et al. report their results of transdiscal biacuplasty for the treatment of chronic low back pain .

Publication recognizes that a study was methodologically sound. Purists might argue that the study of Kappural et al. could have been more stringent. For instance, the sham treatment might have involved penetrating the disc with the electrodes but not delivering radiofrequency current. However, for practical purposes, the authors performed their interventions as well as might realistically be expected, for an invasive procedure in human, volunteer subjects. For the manner in which the treatments were conducted, the risk of bias was low; and this is corroborated by the results obtained.

To the credit of the authors, the results are presented in a transparent fashion. Readers should appreciate this and salute the authors for their integrity. Providing transparent data allows perspicacious readers to conduct their own analysis and draw their own conclusions. Furthermore, the authors provide both continuous outcome data and categorical data. The latter are essential for the estimation of success rates.

A well-conducted, placebo-controlled trial provides evidence on two counts. It shows if a treatment is effective; and it shows by how much the treatment is more effective than a sham.

In the study of Kappural et al., the continuous data and the categorical data both clearly show that active treatment was significantly more effective than sham treatment. For the outcomes of pain and function, mean scores were significantly better after active treatment, and the success rate was significantly greater. This constitutes strong evidence that the outcomes achieved by biacuplasty, in this study, cannot be dismissed as those of a placebo effect.

Some critics might be concerned that the success rate in the placebo group (3%) was too low, but this is not disturbing. It would be unusual for a success rate higher than this to persist at 6 months in patients with chronic back pain. Moreover, the rules of engagement require that in a prospective trial, we must accept the placebo response rate that was, indeed, encountered.

However, pivotal to the interpretation of the results of this study is the definition used of success. A patient was deemed to have a successful outcome if they exceeded the minimal clinically important change for pain and for physical function. Although it has become common to use such a definition in studies of spine surgery, it is a very generous definition.

The minimal clinically important change is the least improvement that patients, on the average, rate as rendering them better than they were. For back pain, it is an improvement by only 2 out of 10 points; for physical functioning, it is an improvement of 15 out of 100 points. Although they might consider themselves better, these patients nevertheless continue to have pain and continue to be disabled.

What patients desire from a treatment for chronic back pain is 90% relief of pain, or more . For marked improvement, they require at least 50% relief . Minimal clinically important changes fall considerably short of these targets.

The objectives and rewards of clinical scientists are different from those of industry. Clinical scientists pursue the truth by collecting data. Their reward is recognition of their skill and integrity. Kappural et al. consummately deserve this reward.

In contrast, industry desires evidence of efficacy so that it can promote its product. In this regard, the data of Kappural et al. describe only a very weak treatment for back pain applicable to very few patients. The selection criteria were strict. The results reported do not apply to overweight patients, to ones who smoke, or to patients with workers compensation claims or litigation. Yet even in ideal patients, the success rate is only 30% (±17%), and “success” means only a two-point change in pain coupled with a 15-point change in physical function. In marketing a treatment with these limited properties, industry should not overstate its claims. In responding to temptations to adopt this treatment, readers should recognize its limitations. By publishing the results of Kappural et al. and discussing them, the Journal serves to help readers do so and not be prey to exaggerated claims of efficacy.


View Abstract