I just found a SOSORT paper with a discussion about Lori Dolans Metanalysis. They mention the ethics of a Random Controlled Trial and in a roundabout way, accuse Dolan of cherry picking. Very interesting.
2009 SOSORT paper
In discussing the meta analysis they say (emphasis added is mine):
The authors concluded that "Based on the evidence presented here, one cannot recommend one approach over the other to prevent the need for surgery in AIS." Nevertheless, the authors elected to exclude from the study the groups with bracing plus exercises. Under the conditions of their analysis, therefore, according to the Material and Methods of the paper, their conclusion should have been as follows: "Based on the evidence presented here, according to the English literature and excluding the combined approach of bracing and exercises, one cannot recommend one approach over the other to prevent the need for surgery in AIS.". In fact, according to the same criteria used in the previously mentioned metanalysis [7], the papers published by some members of the international Scientific Society on Scoliosis Orthopaedic and Rehabilitation Treatment (SOSORT) [8-12], that are in the English literature but include also exercises, have yielded results that are in conflict with those of the reported systematic review [7] (Figure 1).
The world of treatment of scoliosis is gradually changing [2] and two main ideas are facing each other: one is more surgically oriented, with the prevalent idea that bracing is not an effective treatment [6]. This position has been used to justify the ethical approval of a Randomised Clinical Trial now underway in the US. The SOSORT is more conservatively oriented, and their members have presented a substantial body of data on the effectiveness of conservative treatment in general [8-12], and of exercises and bracing in particular [13-26]. Consequently, a formal debate among this Society concluded that a Randomised Controlled Study on brace efficacy would be ethically unacceptable [27].
I would attach their Figure 1 but the whole text is available at the link above. The bottom two bars on the chart are from the Dolan metanalysis paper showing no difference between braced and observation. The upper bars show the results from studies that were excluded because exercise therapy was used in conjunction with bracing. It seems to me that, even if we assume exercise therapy does nothing, we still have data from those studies showing braced patients and the point the SOSORT author makes is valid. Looks a bit like cherry picking to me.
Note: I earlier criticized Dolan for excluding the Nachemson data (the failed 1995 SRS bracing study) in her analysis. I later conceded I was dead wrong about that criticism because the Nachemson data did not include surgical rates. This is a different issue. And, of course, I could be wrong once again. Nevertheless, seeing the debate/argument discussed in the technical literature is fascinating.
2009 SOSORT paper
In discussing the meta analysis they say (emphasis added is mine):
The authors concluded that "Based on the evidence presented here, one cannot recommend one approach over the other to prevent the need for surgery in AIS." Nevertheless, the authors elected to exclude from the study the groups with bracing plus exercises. Under the conditions of their analysis, therefore, according to the Material and Methods of the paper, their conclusion should have been as follows: "Based on the evidence presented here, according to the English literature and excluding the combined approach of bracing and exercises, one cannot recommend one approach over the other to prevent the need for surgery in AIS.". In fact, according to the same criteria used in the previously mentioned metanalysis [7], the papers published by some members of the international Scientific Society on Scoliosis Orthopaedic and Rehabilitation Treatment (SOSORT) [8-12], that are in the English literature but include also exercises, have yielded results that are in conflict with those of the reported systematic review [7] (Figure 1).
The world of treatment of scoliosis is gradually changing [2] and two main ideas are facing each other: one is more surgically oriented, with the prevalent idea that bracing is not an effective treatment [6]. This position has been used to justify the ethical approval of a Randomised Clinical Trial now underway in the US. The SOSORT is more conservatively oriented, and their members have presented a substantial body of data on the effectiveness of conservative treatment in general [8-12], and of exercises and bracing in particular [13-26]. Consequently, a formal debate among this Society concluded that a Randomised Controlled Study on brace efficacy would be ethically unacceptable [27].
I would attach their Figure 1 but the whole text is available at the link above. The bottom two bars on the chart are from the Dolan metanalysis paper showing no difference between braced and observation. The upper bars show the results from studies that were excluded because exercise therapy was used in conjunction with bracing. It seems to me that, even if we assume exercise therapy does nothing, we still have data from those studies showing braced patients and the point the SOSORT author makes is valid. Looks a bit like cherry picking to me.
Note: I earlier criticized Dolan for excluding the Nachemson data (the failed 1995 SRS bracing study) in her analysis. I later conceded I was dead wrong about that criticism because the Nachemson data did not include surgical rates. This is a different issue. And, of course, I could be wrong once again. Nevertheless, seeing the debate/argument discussed in the technical literature is fascinating.
Comment