Babies and consent: yet another NHS scandal

But it should lead to improvements in research governance within the NHS

News p 1291

The chief medical officer of England has a slide that he often shows of a long list of scandals within the NHS. He must now add to his list the failures in the paediatric department in North Staffordshire Hospital (p 1291). A review of what happened in the hospital shows that the case is complex, but the story will probably play out in the media and in public consciousness as babies being entered into a dubious trial of a new treatment without their parents' consent.1 The worst single accusation is that consent forms were forged, and the police and the General Medical Council are investigating that possibility. At best, the North Staffordshire episode will lead to an overdue improvement in research practice throughout the NHS. At worst, it will further undermine public confidence in the NHS and doctors and lead to a proliferation in bureaucracy that will increase the difficulties of doing research.

The North Staffordshire review started out by examining complaints about consent for a trial of continuous negative extrathoracic pressure (CNEP) in premature babies, but the panel found that it had to consider other issues. These included other research, the routine use of CNEP to treat bronchiolitis, and the use of covert video surveillance to detect Munchausen syndrome by proxy. The panel was thus struggling with several issues at once, but its greatest difficulty was working out what was acceptable at the time the CNEP research was done in the early 1990s. The world has since changed dramatically. In 1990 people were mostly happy to believe that “doctor knows best.” Obtaining consent was often a formality rather than an informed, two way discussion of choices, risks, and benefits.

At theBMJ we can see how the world has changed by looking at the uncontrolled trial we published from North Staffordshire Hospital on the use of CNEP in premature babies in 1989.2 We would reject such a study now, insisting on a randomised controlled trial. We would now expect the paper to make clear that the ethics committee approved the trial and that patients gave fully informed consent. In 1989 we didn't have such expectations. I labour this point because it has been a central difficulty for the review panel in allocating blame. By the standards of 2000 much of what has happened in North Staffordshire is clearly unacceptable, but it may be nothing like as egregious viewed through 1990 eyes. Forgery, however, was a crime in the time of Babylon.

The CNEP trial (a randomised controlled trial) was approved in 1990 and eventually published in Pediatrics in 1996.3 The trial can easily be criticised, not least for having an unvalidated outcome measure. “Failure to use a validated instrument” is now one of our standard reasons for rejecting studies. But most studies published in medical journals can be easily criticised. Considerably less than 5%, for example, meet the stringent scientific criteria of the journal Evidence-Based Medicine.4

The review panel examined the work of the local research ethics committee in approving the CNEP study and found that it generally followed the guidelines of the time but might have done better in assessing the quality of the research—perhaps by using external peer review. (Ironically, the panel criticises the ethics committee for not following the latest thinking while itself placing an old fashioned faith in peer review and failing to note research that finds severe deficiencies in peer review.5) Again we have lots of evidence on the highly variable performance of ethics committees, not least because they are usually overworked, undertrained, and underresourced.6

Informed consent is at the heart of the issue, and many of the parents have no memory of giving consent. The panel was, however, presented with evidence from another well done trial where patients did not remember giving consent even though they clearly had. The panel thus concludes somewhat lamely: “It is not possible now to be sure who completed some of the consent forms or to be sure that all of them were completed as intended.” Further investigations are needed, but again research shows that the standards of obtaining consent are poor.7

Another part of the report looks at the current use of CNEP to treat bronchiolitis. It found “that there is no substantial evidence base in support of the use of CNEP in management of bronchiolitis.” This might seem shocking until we realise that there is little good evidence to support the conventional management of the condition, namely positive pressure ventilation.

The aim of showing how the deficiencies identified in the North Staffordshire review are widespread in the NHS (and all healthcare systems) is not to excuse what has happened but to explain it. The main outcome of the review is a series of recommendations on improving “research governance” within the NHS, and these are to be welcomed.

There are some basic principles that should underpin improvement. The first principle should be transparency. Patients must be fully informed of what is happening, not only in research but also in clinical practice. Similarly processes of approving research must be open to public scrutiny, and the NHS and doctors should be honest about the lack of evidence for so much of routine practice. Problems may have arisen in North Stafford because the doctors wanted to “protect” the patients: they thus “sold” CNEP as “a kinder, gentler treatment.” But it is neither kind nor gentle to deceive patients about the reality of their predicament.

A second principle should be partnership. Patients should be involved at all stages of designing, approving, and carrying out research. Finally, we should promote the principle that it is good for everybody, including participants, to conduct research. The worst outcome from this tragedy would be that it becomes increasingly difficult to do research in the NHS. Then we will never know how best to treat bronchiolitis.