Henry Overman (http://spatial-economics.blogspot.com/2012/03/nhs-competition-bad-science-or-bad.html) makes a number of points about our blog on the LSE site (http://blogs.lse.ac.uk/politicsandpolicy/2012/03/05/bad-science-nhs-competition/) which I think are worth replying to.
His first general point is that our criticisms don’t reflect the various Cooper papers, and he cites our blog versus the Cooper papers.
This seems an odd strategy. Cooper has, in addition to his academic papers, published a range of pieces in the Guardian and FT (and in many other places), as well as in press releases and on Radio 4. He seems to have no problem with his main finding being represented as being ‘competition saves lives’. He therefore doesn’t claim, as in the first example on Overman’s blog entry that the ‘major improvements in primary prevention in general practice’ could be due to ‘primary prevention and hospital care’. Cooper is being very clear – he believes competition saves lives. To present the work in any other way is to misrepresent what Cooper has claimed from it.
Overman’s second point is that Cooper is clear on the mechanism for improving care – ‘Competition on elective care improves management which also happens to benefit AMI’. This appears to contradict his first point (that the major improvement might have come through primary prevention in general practice), and further muddy the waters by claiming that competition ‘also happens to benefit AMI’. What does that mean? Is that causal or not? Overman suggests AMI was chosen because patients have no choice in relation to it. This amounts to a causal chain whereby even the clinical areas which are not responsive to competition themselves, were affected by competition. Does this sound remotely credible to you?
Overman’s third and fourth points are that it isn’t necessary for patients to remember or make informed choices to provide ‘incentives for hospitals’. However, he doesn’t really explain what the incentive for hospitals actually were. Perhaps these incentives comprise of the additional payments that might accrue from more patients choosing a hospital. But, at a time when waiting lists were in place, there seems to be little incentive for clinicians to want to be chosen – in fact we might argue the opposite, that being chosen less would help clinicians get through their waiting lists. Because the paper does not actually work through any causal chain (despite Cooper calling himself an ‘empiricist’ in his FT piece), there is no actual data on how the incentives were actually received by any real clinicians or managers. That these incentives existed is simply inferred, not shown.
Overman’s fifth point is that the EJ is more than capable of reviewing work like this. I think my response would be that the paper has internal validity (which might have been what the reviewers were looking for), but simply does not capture the complexity of the phenomena it is meant to be investigating. If you are going to make claims that ‘competition saves lives’ – which is itself ridiculous – health workers save lives, not governance structures – then you’d better have some pretty strong evidence. In my view, Cooper and his co-authors don’t. They have massively over-generalised from narrow data, and timed press events publicising their more sensational findings to be taken up by politicians and journalists who haven’t even read their work. This isn’t good science.