In 1998 the Lancet published a short case series of 12 autistic children admitted to the Royal Free Hospital (RFH) in London for treatment of bowel symptoms.1 The paper was one of the first to bring to the medical community's attention an unexplained association between autism and bowel disease, describing the children's clinical findings and noting that parents timed the onset of eight of their children's conditions from the MMR vaccination. Twelve years later the General Medical Council, the UK medical regulator, concluded that the way the three senior authors of the paper -- Dr Wakefield, and Professors Walker-Smith and Murch -- had researched and written the paper amounted to serious professional misconduct, especially in the way the children were selected and subject to invasive investigations, in the doctors' unethical research practices and Dr Wakefield's failure to disclose his conflict of interest. In the sanctions stage of the hearing that began in April 2010 the prosecution is recommending that Wakefield and Walker-Smith are erased from the register of medical practitioners and that Murch is suspended.
Following evidence from the prosecution and defence in the largest and most expensive hearing in the GMC's history, the fitness to practice (FtP) panel faced the apparently small issue of deciding which of two ethical approvals applied to the 1998 Lancet paper: approval 162-95 in 1995 or approval 172-96 in 1996. For the doctors, however, this one year's difference was momentous, even after 15 years of passing. The entire case would be won or lost on this one decision.
The GMC prosecution claimed that 172-96, granted by the RFH Ethics Committee (EC), was the only ethical approval for the Lancet case series. The defendants, however, referred to 162-95 as the appropriate approval. Accepting approval 172-96 would shift the entire findings of fact, in effect changing the goalposts.
Ethical approval 162-95 was a short letter from the EC to Walker-Smith allowing him to take two additional mucosal tissue samples for research from the bowel of children undergoing a colonoscopy to investigate their clinical condition. The team could thereby obtain further research data on a child's condition in addition to clinical data. The approval was generic and applied to children being treated by Walker-Smith. This included the Lancet case series of 12 children with autism and gastroenterological symptoms, a study describing clinical symptoms for which no research protocol was needed as the children were admitted for treatment -- ie they would be subject to the same tests whether or not they were included in the research study. The case series did not research the impact of new treatments or diagnostic procedures which would require more rigorous ethical approval. The use of colonoscopy and other procedures to investigate clinical conditions was a matter for the doctors' clinical judgment. The approval set the start date 5 September 1995.
In cross examination the chair of the Ethics Committee in 1995 and 1996 conceded that no EC approval was required to write up a retrospective case series for a journal. The children were admitted consecutively on the basis of clinical need, their clinical conditions investigated and written up in accordance with approval 162-95. The 12 children were referral by their GPs or local consultants. That ethical approval was granted was stated at the end of the paper.
The prosecution acting on behalf of the GMC, who brought the case against the doctors, argued that 172-96 (and the associated research protocol) was the only approval governing the Lancet case series. Ethical approval 1972-96, for a study of 25 children, provided a different set of conditions: the start date for including the children, 18 December 1996; the illnesses the children must have and the vaccinations received to qualify; and the bureaucratic procedures to be followed -- consent forms lodged, who had responsibility for ordering investigations, who was the lead researcher, etc.. These conditions of approval represented the armoury the prosecution would use to knock down the defendants' simpler conditions, with the start date the most powerful piece taking first shot. If the Lancet paper couldn't stand up to the start date -- that the children could only be admitted after 18 December 1996, not after 5 September 1995 -- then everything else would fall, including the contention that the children were selected for treatment on the basis of clinical need.
The summary of the panel's conclusions on findings of fact shows that the panel chose approval 172-96 ‘in the light of all the available evidence’ (p.3),2 rejecting the defendants' claim that the Lancet case series was done under approval 162-95. Most surprisingly the panel gives no reasons for rejecting 162-95. In one move the Lancet series became a research 'project' and not a treatment exercise, in which clinical interventions on severally disabled children (eg colonoscopies and lumbar punctures) were done for research purposes even though clinically indicated for the treatment of sick children with serious and unexplained symptoms. In its only reference to 162-95 in the summary, the panel misleadingly refers to it as a 'project' (p.3), whereas the approval was generic covering the clinical practice of taking additional samples from colonoscoped children for research purposes such as the Lancet case series and other descriptive case series were they to be undertaken.
2. Key findings based on 172-96
Start date and ethical approval
By selecting the start date of 18 December 1996, the prosecution could show that 7 of the 11 children (one of the 12 was a US patient and excluded from the proceedings) were accepted for the research before the start date and so without ethical approval. For example, Wakefield was charged 'that Child 2 was investigated under a project without the approval of the EC in that it was not research covered by any EC application other than that for Project 172-96 and...contrary to the conditions of approval for Project 172-96 Child 2 had been enrolled into the project before 18 December 1996' (p.14) - a phrase repeated ad nauseam for seven children and in almost all cases for the same set of charges against each of the three doctors. Using 172-96, the panel was able to frame the facts so that Dr Wakefield's statement in the Lancet paper that the study was ethically approved was 'dishonest, irresponsible, contrary to [his] duty to ensure that the information provided by [him] was accurate' (pp.48-9). Had 162-95 been duly acknowledged these charges would have fallen.
Applying the inclusion criteria of 172-96 meant that the 11 children did not qualify for the project as they failed to meet them, namely 'enteritis and disintegrative disorder following measles/rubella vaccination' (p.8). Instead they were described in the Lancet as autistic and vaccinated with MMR. By failing to adhere to these inclusion criteria for each child, Wakefield was admonished for having 'failed to comply with [his] duties to the EC as a named Responsible Consultant' (eg see p.14). Again these criteria were replicated for each of the three doctors.
However, look closely at the changing diagnoses among the children's doctors as reported in the summary and you realise how precarious the diagnosis of autism is even among experts. For example, child 6's GP offers an initial diagnosis of autistic syndrome in August 1996; whereas three months later the psychiatrist working with the research team suggests Asperger's Syndrome as the most likely diagnosis (pp. 24-5). It is well known that autistic spectrum disorders are diagnosed differently by different doctors. Yet true to their ideological ends, the GMC prosecution forced the facts when the reality was far from clear.
Research not treatment
By accepting approval 172-96, the Panel was asked to judge that the children were being admitted for research and not treatment. Placed within the 1996 framework, the Lancet paper was no longer a case series of children undergoing clinical treatment but a research project subject to far stricter approval. Consequently each doctor was charged with carrying out interventions on the children for research purposes alone for which they had no approval.
Interventions not clinically indicated
The most serious implication of judging the doctors' actions according to research criteria alone was that, in a UK climate of heightened child protection, they were found to have 'caused' interventions not clinically indicated and 'contrary to the clinical interests of the child'. Yet, although the doctors claimed 172-96 was not the appropriate approval, the answer to question 11 on the 172-96 protocol and pro forma stated that 'all the procedures and the majority of the samples are clinically indicated'.
The consequence of using 172-96 was that facts allowable under the 1995 approval were excluded, specifically testimony from parents of eight of the Lancet 12 saying that their children were being treated for serious clinical needs by the three doctors who exercised deep professional concern throughout.3 Their evidence submitted by letter to the GMC was disallowed.
2. Selecting the children for the research
The prosecution used a conventional referral model in deciding how the doctors should have selected the children for the Lancet study. Any departure from this model was denounced as unethical, thereby institutionalising, 14 years after the 1996 approval, a 'correct' process for which there was then no authority. For the Panel a case series is 'a routine process in which the investigators had played no active part' (p.45). Ethical research involves the researcher playing a passive role in the referral process, as patient is passed from GP to consultant and finally to the medical researcher, with the parties necessarily sharing an understanding of the patients' condition. The researcher is involved in the final but not the initial selection of patient/research subjects.
The summary gives examples of parents initiating contact with Dr Wakefield and his willingness to talk to them. However, far from his behaviour being held up as a model of the caring professional, the panel concluded that these communications constituted 'a biased selection of patients in the Lancet paper' (p.47), in which, to take the example of child 12, Dr Wakefield 'was actively involved in the referral process' in that he was 'in written and telephone contact with Mrs 12' (p.34). The panel contrasts these findings with the Lancet case series' own account of 'children consecutively referred to the department of paediatric gastroenterology'4 and concludes that 'The description of the referral process in the Lancet paper was therefore irresponsible, misleading, contrary to [Wakefield's] duty to ensure that the information in the paper was accurate' (p.46).
The Lancet paper admits the possibility of 'selection bias in a self-referred group' that characterised part of the referral process.5 Nonetheless, despite the element of self-selection, the referral process for the case series conformed to the conventional methods. For example, the GMC summary of findings describes the referral process in the 11 cases where GP's or local consultants consistently wrote referral letters to Professor Walker-Smith. Each child was processed down the referral line in the established way, even though a study of new symptoms would require that a process of pre-referral negotiation should take place, whereby Dr Wakefield communicated with local GPs, consultants and parents to establish the nature of the child's symptoms and to explain to all parties what the research entailed. The GMC denounced the negotiation that the parties to the referral exercised in order to reach agreement on issues such as the clinical conditions of the patient, what treatment is or is not available and what benefit the research might bestow on the patient.
The GMC applied an outmoded model of medical research to Wakefield et al's work, with yet again profound implications for the three doctors and for research ethics more widely, in ways that impede future research into autism, bowel disease and vaccine safety. To describe the ethically responsible search for patient subjects for a research programme into unexplained symptoms as 'fishing expeditions', to use the prosecutor's words, is a travesty of the research ethics involved, a demeaning and cynical account of the lengths to which caring doctors went in their concern for patients and parents, and a cruel impediment to research that could benefit seriously ill children.
3. The Lancet paper controversy: ask no questions
The GMC used approval 172-96 to censure the doctors for having written the Lancet paper because of its subsequent impact on MMR take-up. The panel accused Wakefield thus: 'You knew or ought to have known that your reporting in the Lancet paper of a temporal link between the syndrome you describe and the MMR vaccination had major public health implications, would attract intense public and media interest'.
Wakefield is blamed for having raised questions of the MMR that have been central to the lives of the autism community since the 1990s, and the question many parents have asked, whatever conclusions reached: did the MMR cause my child to become autistic? The GMC sought to censure Wakefield et al's independent and critical scientific research into the association between autism and bowel disease which raised legitimate questions and recommended further research about the role of MMR in the onset of these condition.
Further, 172-96 was used to conflate two studies, different in conception and funding: the Lancet case series and the Legal Aid Board study on behalf of autistic child litigants claiming damages from MMR manufacturers. The Panel was 'satisfied that the project 172/96 document is substantially the same as the protocol sent to the LAB by Mr Barr [the solicitor representing the child litigants alleging MMR damage] in June 1996' (p.11). By framing the two studies in this way, the panel could claim that Wakefield's LAB funding 'constituted a disclosable interest' that should have been disclosed to the EC (p. 11) and stated in the Lancet paper (p.49).
The potential consequences for research of the GMC’s judgement are that:
doctors will revert to a referral process for patients of unexplained symptoms which could impede research into much needed new treatments
medical scientists will avoid researching autism, related bowel diseases and especially vaccine safety, for fear they will be treated like the three doctors and endanger their careers
independent research into vaccines could become a no-go area in medical science, apart from research undertaken or sponsored by commercially-driven pharmaceutical manufacturers and government. Parents considering vaccinating their children would, more than ever before, have only official product information to rely on with no access to independent advice
Conflict of interest
The GMC argued that Wakefield should have disclosed a conflict of interest in that he was also receiving money from the LAB to research the role of the MMR in the onset of autism and bowel disease in 1500 sick and disabled children claiming damages against three manufacturers for injuries allegedly caused by the MMR. The GMC used a third party test, whereby conflict of interest rests on what other people would perceive as a conflict, so that Wakefield 'had a duty to disclose to the Editor of the Lancet any disclosable interest including matters which could legitimately give rise to a perception that [he] had a conflict of interest' (p.44), claiming that this understanding of conflict of interest prevailed in the scientific community in 1997. The GMC failed to report the Lancet's guidelines that: 'The conflict of interest test is a simple one. Is there anything . . . that would embarrass you if it were to emerge after publication and you had not declared it?'.6 Here a subjective test rather than third party test is used which places the onus on the author(s) to judge if they would be embarrassed were a conflict of interests to emerge.
If approval 162-95 had prevailed, the Lancet paper would have been seen as a descriptive case series, as the authors contended, entirely different from a legal aid funded study seeking to establish if a causal link exists between the MMR and autism/bowel disease.
Whilst the GMC has made much of the ethical failings of Dr Wakefield, which in their view would support a case for serious professional misconduct against him, no acknowledgement is given to the fact that in the case of Dr Wakefield, his legal aid funded study was done in support of litigation for severally sick and disabled children. From the GMC's point of view there is no difference between conflict of interest in this case and conflict where doctors fail to disclose funding sources from profit-making bodies such as pharmaceutical companies or fail to disclose that an academic publication is part of a programme in support of the development of commercial products.
Conclusion: ethics and ideology
That the panel reduced a complex multitude of facts to a core set of facts that accorded with 172-96, in the longest and most complex GMC hearing ever, should have caused alarm bells to ring. This reduction alone should signal that the process of reaching conclusions was driven by an ideological agenda which rendered complexity a choice between two ways of ordering the facts, namely between ethical approvals 162-95 and 172-96. For the defence, 162-95 appealed to the panel’s understanding of how researchers conducted a small case series which reported findings of an association between autism and bowel disease; and for the GMC prosecution and panel, 172-96 appealed to an ethical argument that simplified complexity into pre-digested and highly selective findings of fact. The reliance on the 1996 ethical approval as the exclusive means of framing the facts points to the presence of an ideological agenda in the hands of the GMC, prosecution and panel, and shows how in the end the 3 doctors were framed according to this agenda.
This agenda, like all ideology, points to what the panel expressly denied, that the hearing was orchestrated by the government, who recommended the case to the GMC, the GMC and sections of the press, in particular the Sunday Times journalist Brian Deer who submitted the original complaint and evidence in 2004, which served as the template for the GMC's preliminary investigation, the prosecution's case and the panel's conclusions - a template which they followed to the letter.