The sad thing is, it was a good paper about people with severe physical disability and how a coordinated approach can successfully meet their needs (BMJ 1993;306:95-8). One professor called it seminal. It was used as an empirical base to promote community care. I trumpeted the findings with gusto. But it has had to be retracted (retraction box p 1700) because my coauthor was struck off the medical register for research fraud and I am unable to vouch for the probity of the data used (28 February, p 647). It has been an unedifying experience putting my own scientific credentials into question.
Mark was a registrar and senior registrar in my department in Somerset when the survey was done in 1989. He was a reliable trainee and became a trusted friend and colleague. He was on the way to becoming accredited as a public health physician and was appointed lecturer and later senior lecturer at our local university department of social medicine. He was a success.
The survey was commissioned and designed jointly by our social services department and us. It was carried out in house. One hundred and eighty one severely disabled adults were interviewed in 1989 and underwent physical assessment. Mark supposedly carried these out. A year later he did without prompting what I believed was an imaginative addition. He reinterviewed by telephone those people who had been found to have unmet needs, to see whether their needs had been met in the intervening 12 months after the original assessment. The approach was able to show that many of the unmet needs identified were remediable at minimal cost and not based on some professional or patient wish list.
The results were fascinating. Those severely disabled people who were being seen only by their GP and community nurse had far more unmet needs than those who were also being seen by a social worker or who were in contact with a patient organisation. Here was empirical evidence of inadequate medicosocial assessment reported at just the right time to back the government’s community care proposals.
When I heard that Mark was to appear before the General Medical Council we searched for the original data records. The original questionnaires and data sheets could not be found. The health authority has moved since the survey was undertaken and old files had gone from basement to basement. Not only was I unable to find the completed questionnaires but the computer files containing the coded results could not be found either.
I then tried to contact as many as possible of the original people who had had the follow up telephone interview. My purpose was to be assured that at least some would recall the telephone interview undertaken by Mark six years before. Mark gave me a list of these individuals. This act gave me confidence of his innocence. There were 75 people on Mark’s list. They were those with unmet needs and who had responded to the telephone follow up survey out of the original 181 individuals. On the family health services authority’s register, 36 were still living, 14 had died, two were known to have moved out of the county, and 22 could not be identified on the register. Of the 36 alive, telephone numbers were found for 18. The 18 were telephoned. Three did not answer. Of the remaining 15, 10 could not recall the original interview and physical examination which was undertaken in their own homes, the other five did remember the home visit, and none remembered the follow up telephone survey. Two individuals, who were very alert and receptive to my inquiries, were quite convinced that they had never had a telephone survey undertaken one year after the original. So I could find no one who could remember being telephoned and only a third could remember the original home visit.
Before the GMC hearing I was happy to assume that everything that Mark had done was true and accurate. Now, the rules of trust have changed. After the GMC ruling it has to be assumed that all his work is fraudulent until proved otherwise. So I confronted Mark over a pint of beer. I asked about the survey and suggested that it was important to know the truth for the benefit of science. A confession would not result in further detriment to him. Were the survey results fabricated? No, certainly not, was the reply. He spoke with the sad eyes of someone losing the trust of another of his few remaining friends.
There are public as well as personal consequences if the paper is a fraud. If it is not but is so claimed there is a loss to science if the paper is retracted when it is in fact honest.
Nationally, it is difficult to say what the effect of the paper has been on health care. The paper was used in the part I examination of the membership of the Faculty of Public Health Medicine.
Who knows whether it had an effect on the introduction of community care? It should have done—it was pertinent to the new multidisciplinary assessment arrangements that were being put in place jointly by social services and health authorities.
From a local point of view, it had a powerful effect on my own professional work. The paper gave me confidence to seek radical solutions to the way we assess and care for disabled people. We were able to push through some innovative arrangements such as the use of specially trained nurses to undertake assessments, the establishment of a consultant post in physical disability, and fully integrated arrangements, such as joint budgets, between social and health services. I felt confident enough about the work to present it at a conference at the Faculty of Public Health Medicine.
Fraudulent research cannot be prevented completely but perhaps we should be less trusting, especially if the results are spectacular. I will not coauthor an original paper again without scrutinising the original data. There needs to be a repository for the original survey records and computer files of published research for reanalysis.
We all cut corners to get things done. There is no clear cut distinction between slipshod work and fraudulent research. Where do I stand, where do you stand, on this continuum of probity?