The use of RCT methods
The surgeons expressed positive views of the use of RCTs for examining the effects of medical interventions. Nearly all said that the SST addressed an important question, involved an experienced and respected research team and used the 'gold standard' research method:
It was well thought out. You know a lot of work was put into it. I mean we had two collaborators meeting; we had a hundred statisticians ferreting about so yes I think the design was good. Surgeon with low personal recruitment
There were however reservations about the SST design, despite positive views of RCTs in general. For example one surgeon wondered if there are better ways to gather reliable data about surgery in a shorter time frame and suspected that the RCT was used because it was 'politically correct'.
The surgeons were asked what they understood by 'equipoise' and 'uncertainty principle' in relation to the SST. Nearly all were unclear on the meaning of 'equipoise'. Although uncertainty was a familiar term, surgeons often related it to the outcomes of any surgery for an individual rather than to the specific comparison of interventions in a trial, here spinal fusion and the FRP (Functional Rehabilitation Programme).
(Mis) understandings about trial design
It became clear that the surgeons held different views of the purpose and design of the SST. Surgeons were often unsure of the trial aims, were unclear about the nature of the comparison, and expressed concerns about flexible entry criteria.
The purpose of the SST was to compare surgery to an intensive functional rehabilitation programme (FRP) for treatment of chronic low back pain. However, most of the surgeons we interviewed saw it as a trial of surgery, some thought it was a trial of rehabilitation, and others suggested a much broader remit.
I just thought it was a way of sort of trying to work out what's the best treatment for patients, with back problems, that you didn't know what to do with. Surgeon with low personal recruitment
[They] had the great [pause] foresight to actually ask a question which is you know, it's the big question that is in front of us every day when we come to work. Does surgery work? Is it worth doing? You know, that was, what, what, what bigger question can you ask? Surgeon with medium personal recruitment
None of the surgeons interviewed stated that it was designed to compare two 'equivalent' treatments, although this was a key feature of the trial. Many eligible patients had already received extensive physiotherapy and the surgeons described the need to 'dress up', 'talk up' or 'sex up' the FRP arm as something different or new. Some, in contrast, said that they presented FRP as the 'control' or 'conservative arm' in their recruitment 'spiel'.
The SST involved a pragmatic design, comparing interventions as used in clinical practice, rather than in the rigid and artificial circumstances created by explanatory trials. Broad eligibility criteria reflected the fact that surgeons, within and without the trial, vary in their views of which patients might benefit from surgery (several commented on the large difference in the numbers of fusions performed by UK spinal surgeons). The pragmatic design was not understood by the surgeons. Several expressed concern what they perceived as an unfortunate variability in the SST sample and expected that colleagues' perceptions of eligible patients would be different from their own. They talked about psychological and social factors in selecting patients for spinal fusion and the 'art', 'instinct' and 'eye' of the surgeon. Although only consultant-grade specialist surgeons identified candidates for the SST, variations in skill and preferences caused concern. The inclusion of patients with a range of clinical presentations left some feeling that the SST used a flawed design which would render the results unsound or irrelevant to their own practice. One surgeon argued that the broad inclusion criteria loaded the trial against spinal fusion and that, counter to standard trial procedures, they should have been highly selective:
[If] you want to come out with an answer which says 'spinal fusion works for a selected group of patients', you select the patients very, very carefully for those patients whom you have the least uncertainty for offering the fusion. .... I think that if you actually bias the selection of patients going into the trial to the patients who, in whom the outcome was seen at the beginning to be most uncertain, then you are going to end up with a load of bad apples in the trial who aren't going to do terribly well and I think that's why a lot of surgeons weren't – were fairly diffident about going into it at the beginning. Surgeon with medium personal recruitment
Others also thought the trial was biased against spinal fusion because the 'uncertainty of outcome principle' meant that the most 'promising' patients would have been excluded by their surgeon i.e. they would be given spinal fusion outwith the trial. As one put it 'It's not the test of spinal fusion. It's a test of spinal fusion in a group of patients nobody knows what to do with'. One consultant suspected that others, like himself, would also take the psychological, social and intellectual profile of potential participants into account when recruiting. Several surgeons felt that these difficulties and problems with recruitment and retention affected the validity of the trial.
These views may have affected the SST. Some surgeons indicated that when recruitment difficulties became public, they had wondered if there was any point in maintaining their recruitment efforts if the trial was destined to 'close early'.
Making sense of the trial results
The SST allowed recruiting surgeons to exercise clinical judgement. The variation in the rates of fusion surgery amongst surgeons, differences in their perceptions of appropriate surgical candidates and in preferred operative techniques, were all accommodated in the trial design. This inherent flexibility did, however, leave many unsure how to interpret the results. Some of the accounts suggested that there had been considerable discussion within the orthopaedic community about the study and the danger of 'bias' in the design. One surgeon commented: "We all thought the trial would rather go against spinal fusion because you are recruiting patients with a high degree of uncertainty." Surgeon with medium personal recruitment
As the SST results had not been published at the time of the interviews, surgeons had not had an opportunity to examine them in detail, but most had received some information about the findings. The principal investigator, Jeremy Fairbank, had presented the findings at a conference and to some colleagues individually. Everyone we interviewed felt they knew the main findings, however, only one surgeon was able to demonstrate a good understanding of the results:
Rehabilitation is more or less as effective as surgery in the treatment of chronic lower back pain in a particular group of patients. That's number one. Number two, recommendations for the future are that people with chronic lower back pain in this situation go and have a rehabilitation programme. If they fail that then it may be appropriate to treat as surgery. We don't know the answer to that. Surgeon with high personal recruitment
Those who accepted that the short-term follow-up showed no real difference between the treatment arms concluded that rehabilitation was a viable and cheaper option. Some of these surgeons were keen to use the results of the study to support their campaigns for better rehabilitation programmes within their Hospital Trusts, although there were also concerns that the results might be misinterpreted or 'misread' as suggesting that spinal fusion should not be funded.
I suspect that some of the PCTs (primary care trusts) will look at it and say 'Well we should stop spine surgery'. But on the optimistic side, some people may look at it and say 'Well look spine surgery should therefore be used for the physiotherapy failures', which begs the question of 'Why should surgery in that sense be any better?' So like all research you have some questions answered that throws up some more. Surgeon with low personal recruitment
The results could also be used to persuade patients that surgery might not be the best option, for example, one surgeon said that the results would help him to recommend his patients complete FRP before considering surgery. However, some stressed that the results would not change their practice, because they had always presented the outcomes of spinal fusion as 'very uncertain'.
Concerns about the design and conduct of the SST, largely based on misunderstandings about the nature of the trial, meant that not all participating surgeons accepted the results. A low recruiting surgeon, clearly unconvinced that the trial was adequately powered, suggested that a bigger trial was needed to find a significant difference between the two treatments. Some were disappointed that the trial did not come down firmly in favour of one treatment or expressed disappointment that the trial did not identify which patients would be likely to benefit most from spinal fusion (something which was outside the remit of the SST). A common belief was that the results did not apply to their practice. A surgeon who saw the trial as 'not that valid' described treatment for back pain as 'a personal journey between you and the patient' and suggested that:
[As] long as you have audited your own practice, and you can show that in your own practice that patients who have fusions, in general, do get very good outcomes, then it's perfectly ethical to continue to do. Surgeon with medium personal recruitment
Collaboration with the trial had motivated many of these surgeons to reflect upon and compare aspects of their practice. Some thought it might be more relevant to real patients and practice to pool routine, anonymised, national and international audit data that would enable them to compare and evaluate their practice.