The Stoke CNEP Saga – the Government enquiry in retrospect*

Why was there a review of research governance in North Staffordshire?

Sustained media coverage of allegations about the CNEP trial created pressure on the Department of Health. The responsible minister quite rightly stuck to the line that appropriate mechanisms already existed to investigate matters of individual care (the NHS complaints process) or individual professional misconduct (the GMC and other professional regulators) but this was not enough to make the problem go away. Eventually the minister came to the conclusion that one aspect that was not being investigated was the framework through which research was supervised in the North Staffordshire NHS Trust. The review thus covered research governance, though many have mistakenly assumed that it was about misconduct. Had it been about research misconduct I might well have refused to lead it.

Why was it called a review?

The minister believed that the issue did not justify the cost and complexity of a legally-based enquiry, under any of the available protocols. This lack of a legal framework had the side-effect that witnesses could not be subpoenaed, and there was no sanction if they make false statements.

The terms of reference

Following the minister's announcement there was a widely communicated general invitation to submit written evidence. A research consultancy was commissioned to interview all those who appeared to have something relevant to say, and a group of witnesses were selected for oral hearings. This wide process led to the submission of massive amounts of material related to child abuse, and this led to that topic being included in the review. In many ways coupling these two issues together makes little sense, but it was a pragmatic solution to the problem and less complex than having two reviews. The review thus examined a number of research studies in North Staffordshire, as well as issues related to child abuse. This paper discusses only aspects related to the CNEP trial, but all the recommendations made by the government review are included in Box 1.

Criticisms of the review

Some months after publication, Hey and Chalmers¹ made a number of criticisms, some of which are dealt with below. They wrote, ‘Almost everything that the review said about the CNEP trial was wrong’. This was hardly surprising because every witness gave a different version of events. The job of the panel was not to make judgements as to what was right and wrong, it was to examine the governance issues. Furthermore, the lack of legal or forensic resources meant that the panel could not determine in a secure way whether some of the witnesses were mistaken, or did not understand events, or may have been malevolent. From a governance point of view it does not matter who is correct, each can be viewed as a test of the governance system and process.

We concluded that careful thought given to the governance process might strengthen both supervision of research and learning, to the benefit of everyone. That is why we recommended that a better system of research governance be developed.

Why was the review not in public?

Several potential witnesses indicated that they were only prepared to speak or write in confidence. These included the NHS Trust and several parents of children who had been involved with either CNEP or child protection issues. A small number of people refused to appear in person, some because they were abroad and others because they had considerable fears. All the latter were parents who had in some way interacted with the child protection system. They feared that giving evidence in person would lead to their children being taken away from them. Whether these fears had any basis in fact we were not able to determine, because they remained anonymous, but it does give some idea of the emotional intensity that surrounded parts of the review. I find it distressing that some of those interviewed found the process intimidating. The panel did everything it could to avoid that – we allowed any witness who wished it, to be accompanied by friends or trade union representatives, and every witness was interviewed before the oral hearing by the research consultants and was allowed to edit the record of their first interview. No-one could have been in any doubt as to the process. An obvious lesson to learn is that any process is probably intimidating to someone who has not encountered it before. A public, legal hearing would probably be the most intimidating of all.

It was quite clear that if we did not agree to confidentiality then we would have a biased and incomplete review. If we had been able to subpoena witnesses and take evidence under oath then we would not have been in this position.

A further issue was that ministers wanted ‘to avoid a circus’. If we had taken evidence in public then there is no doubt that it would have been a much more complex and expensive process. Many of those who did submit material to us also sent it to the media at the same time. Several gave press conferences. If we had worked in public then there would have been significant costs in security and in providing facilities for the media, and the process would have taken much longer.

I have written elsewhere³ that the ideal might be a public process. I support the notion that things need ‘to be seen to be done’, and I think that there would have been less controversy afterwards if everyone had heard what the review panel had heard. There are disadvantages, of course: witnesses can feel intimidated and there is a risk that innocent researchers might be smeared.

Why was the report not shown to those it affected?

We were criticized because those who had given evidence were not shown the final report before it was published. I asked for permission to do this, but it was denied. To some extent that was unfair, but in the months before the report was published, and indeed before it was even drafted, some of the witnesses claimed to have seen the report and made press statements announcing what they said it contained. None of these had any basis in fact, but it is easy to see why officials at every level were reluctant to feed this ‘rumour mill’.

Attempts to raise the stakes

The process was inherently more transparent than the processes used by the GMC and the Trust, both of which, for most of their course, were out of sight of the public and media. As soon as the minister's office announced the review, however, there was media interest. Unfortunately, interest from a number of quarters made the review appear more important than was intended. Various lobby groups saw the review as a means of claiming that what they had to say was important. The more important that the review appeared to be, the greater their status and credibility appeared to be. A mirror image of that was seen in the attitude of the press department within the region, which seemed to regard the review as a means of enhancing their reputation. I doubt if this phenomenon is unique to our review. I imagine that all processes with a public profile may suffer in this way.

One question leads to another

The original research question addressed in the CNEP trial was perfectly sensible and, seen in the context of supporting newborns in intensive care, it is easy to see why outcome measurements ceased when the children were eight months old. That limited follow-up carried a risk, however: from the viewpoint of the parents, life usually goes on after eight months. If survival is associated with handicaps, particularly neurological handicaps, these may not become obvious until later. Neurological handicaps are not uncommon among children born prematurely, and such children will often attend specialized facilities. It was inevitable, therefore, that there would be a cohort of children who had been in the CNEP trial who would meet regularly, along with their parents as they progressed through nursery and infant school. There is no escape from chronic handicap; parents are reminded of it every day. It is not difficult to imagine the conversations that may have gone on alongside the daily routine of taking children to nursery. If one parent wondered if the research trial had affected their child, then it would not be long before they all did.

These questions could only be resolved by extending the follow-up of survivors of children who had participated in the CNEP trial. One of the advantages of trial methodology is that it is theoretically possible to re-contact trial participants. Accordingly, one of our recommendations was that a follow-up of trial participants should be undertaken. Obviously there would be a risk that participants could not be found or that those who could be found might be a biased sample of the original cohorts, but that could be tested in a follow-up study. Interestingly the cost of the follow-up study was three times that of our review. The Department of Health in Whitehall refused to provide funds for this, but I eventually persuaded West Midland Regional Office to support the follow-up. The report by Kate Telford and her colleagues⁴ did not detect any statistically significant advantage or disadvantage of CNEP compared with intra-tracheal positive pressure ventilation.

One lesson now easily seen is that if funding had been available from the outset to support to school age the whole sad tale might have been avoided. The cost of such follow-up would have been substantial, but far less than all the costs incurred by the various enquiries. It is a difficult dilemma for any researcher when faced with limited funds. Is a limited study better than no study at all? In this particular case, the initially limited follow-up resulted in costs in the long run.

The need to protect researchers, patients and the public interest

Research is about the unknown, and going into the unknown involves risks. Neither the clinicians prescribing the treatments compared, the researchers studying them, the patients receiving them, nor the public who pays for them, knows for sure what will happen when a new treatment is compared with an existing treatment. Each of them has something different at stake, and each must be protected, as far as is possible, from misunderstandings or even mistakes that might occur. A governance system should provide that protection and it should have a means to mount an investigation when something appears to have gone wrong.

We examined the various statements about research governance that were current at the time. It was clear that some professional bodies had produced recommendations that were in advance of Department of Health policy. We tried to identify Trusts similar to North Staffordshire which would be prepared to allow us to compare the system we had been asked to review with their own. To our surprise, all the Trusts we asked refused. That seemed to be a further indication that change was needed. We briefed the Research and Development Directorate of the Department of Health in depth as we developed our thinking, and everything that we said about research governance in the report was approved by them.

A political failure in research governance

I was disappointed with the way that research governance developed subsequently. I thought it was too bureaucratic and seemed to have ‘lost the plot’.⁵ I think this was partly because, although the Research and Development Directorate developed the policies, people with little experience of research often implemented them. At a local level it sometimes appears that the simplest way to avoid trouble is to make it as difficult as possible to carry out research. In such circumstances it is hardly surprising that horror stories began to appear. Lengthy documentation was required, even for very simple studies, and this in turn incurred delays and costs that made research more difficult.

If so many people now say that the current system is inadequate why was it not possible to do it better? Ultimately the failure was political. At the time the NHS was undergoing repeated reorganizations at every level, and research governance, and the training and education that should have underpinned it, were simply not a high enough priority. As a result there was inadequate leadership and resources, at every level, for ensuring that the governance process was fit for purpose.

What can we learn?

Research governance should have included three things: a framework for the conduct of research, a system for learning and training, and a mechanism for investigating suspected or alleged misconduct. Some of this is now in place, but I believe that the learning and investigation elements are still weak.

The establishment of the UK Research Integrity Office (UKRIO)⁶ followed publication of the second edition of the Department of Health's Research Governance Framework. UKRIO should at least provide a repository of expertise. It recommends that a named local individual should investigate allegations of research misconduct. Unfortunately, this means that investigations will usually be carried out by someone who has never conducted such an investigation before, and who therefore lacks legal and forensic expertise. A better solution might be for UKRIO to make available a panel of independent experts who know what they are doing.

What were the outcomes of the review?

The main recommendations from the review are set out in the Box 1. They contain no criticisms of individuals, nor do they criticize the CNEP trial. Research governance was introduced. Guidance on consent was issued.⁷ The idea of a system to report an adverse event from non-drug treatments was later endorsed by the government document entitled An Organisation with a Memory.⁸ This was published in the year following our report, and responsibility for implementing it now resides within the National Patient Safety Agency. At a system level I believe that our recommendations were sensible and the UK health system is better for the actions taken. I am particularly pleased that we found the resources that made possible the extended follow-up of the children who had participated in the CNEP trial.⁴

Box 1. Review main recommendations.

1. That formal guidance on research governance within the NHS be developed and issued to the NHS and to partners whose research it hosts.

2. That the Department of Health, professional and regulatory bodies cooperate to consult on and produce agreed guidance clarifying issues of consent for participation in clinical trials.

3. That Department of Health considers the establishment of a surveillance system for unexpected outcomes from non-drug treatments.

4. That a substantial audit of the use of CNEP in North Staffordshire Hospital NHS Trust be carried out to see if claims of significant benefits or damage can be substantiated; and that North Staffordshire Hospital NHS Trust considers carefully the use of CNEP, and defines the scope of its use by a strict protocol until the evidence base is stronger.

Unfortunately there is more to say. Despite our insistence, reiterated at several press conferences and in interviews, that we were looking at system issues, much of the report was interpreted as criticism of individuals. Sometimes people see what they want to see and read what they expect to see. Because our report was the first to appear (long before those conducted by the Trust or the GMC), it became the vehicle for a number of groups to continue to air their particular views. An Organisation with a Memory would later say how important it is to try to avoid blame and learn lessons, but when we reported we were probably the first to try that line. It needs repeating many times.

What did I learn about report writing?

Writing a report as a civil servant responding to a request from a minister is not the same as writing a report as an independent investigator. Our report was a team effort, and was in two parts. The first contained a set of recommendations, which were accompanied by an explanatory text, following, in spirit, the legal model of judgement and obiter dicta (‘said by the way’, words introduced by way of illustration, or analogy or argument). The second part was the evidence submitted to the minister, which was not published.

The draft report was circulated for comment to the relevant divisions in the Department of Health. The recommendations remained unchanged, but publication of the report was delayed because there was disagreement between the public relations people, who thought it read better in a certain order, and the legal team, who had a different view. The published report eventually followed the advice of the legal team, which coincided with the original draft.

It is interesting that many subsequent criticisms of the report related to the explanatory material, which was derived from the conflicting views that had been presented to us. I sometimes wish we could have simply published our recommendations, with a short statement noting that everyone had told us something different. But that is not the way these things are done.

Footnotes

DECLARATIONS —

Competing interests None declared

Funding None

Ethical approval Not applicable

Guarantor RG

Contributorship RG is the sole contributor

*Disclaimer The events to which this article relates occurred a long time ago. The organization for which I worked at the time has since been reorganized several times, and I left it five years ago. I have no access to a comprehensive set of papers. This article is based on my memory of events and is written with the intention of assisting learning.I have avoided using names of individuals in order to try to depersonalize the content, and to focus instead on understanding the issues