The journal BMJ Global Health last fortnight withdrew an article questioning the ethics of trial of cervical cancer screening in India conducted by doctors from Tata Memorial Hospital in Mumbai. The lead author of the article is Dr Eric Suba, an American pathologist with the Kaiser Permanente Medical Center in San Francisco.

The journal published the article online on April 26, but withdrew the paper on May 18. “After acceptance, it was inadvertently published online while awaiting legal review,” the journal said. “It has now been withdrawn on legal advice.”

Retraction Watch, a website that tracks retractions in scientific papers, reported the email exchange between Suba and BMJ’s editors in which the editors told Suba that the “nature and severity of the allegations are such that if the journal were to publish it, BMJ would be at risk of defamation”.

The cervical cancer trials that Suba’s paper examines were those conducted on 375,000 poor women in Mumbai slums and in villages in Osmanabad district of Maharashtra and Dindigul in Tamil Nadu. There is a high prevalence of cervical cancer in these areas. The trials evaluated whether trained “non-doctors” such as health workers, auxiliary midwives and nurses could could use visual inspection with acetic acid to detect cervical cancer early.

The criticism was that these trials contained “no screening” control arm where 141,000 women were deliberately not offered any test for cervical cancer in order to compare the differences in outcomes between the screened and unscreened groups. Of the 548 women in the trial who died of cervical cancer, 254 were from the group that was not screened.

A public interest litigation in the Supreme Court on questioning the ethics of the research was dismissed in January.

Suba obtained documents under US Freedom of Information Act since the Mumbai trial was funded by the US government, to claim that the trial was unethical. Suba has made the retracted article public here.

Suba spoke to about the controversy over BMJ withdrawing the article. Here are the excerpts of that conversation.

How did you first hear about the trials cervical cancer screening in India?
I have been involved with cervical cancer prevention activities in Vietnam since 1994. In 1996, we documented that the burden of cervical cancer in Vietnam was associated with troop movements during the Vietnam War (Cervical cancer is associated with Human Papilloma Virus which is a sexually transmittable virus). Subsequently, the establishment of Pap screening in southern Vietnam was associated with 50% reductions in cervical cancer incidence between 1998 and 2003. The US Preventive Services Task Force has determined that the introduction of cervical screening to previously unscreened communities reduces cervical cancer rates by 60% to 90% within three years, and that these reductions of mortality and morbidity are “consistent and equally dramatic across populations”.

I have been invited to participate in global health conferences to discuss the experience in Vietnam. It was at such conferences that I first learned, during the late 1990s, about the cervical screening trials in India. The first scholarly opportunity I had to object to the India trials was in April 2002, in the medical journal Cancer. In correspondence with the Alliance for Cervical Cancer Prevention, published in that issue of Cancer, I had written, “The Alliance has incentives to focus on technological challenges and currently is conducting randomised trials comparing cervical screening with no screening among groups of Indian women at high risk for the development of cervical carcinoma. Because the Pap test remains the archetype of a successful preventive intervention, any negative results from such trials will not be generalisable to other settings.”

Later, in March 2004, in an article in the American Journal of Clinical Pathology, I wrote: “The Alliance offers the rationalisation that the no-screening arm is justified in India because its participants are receiving community education and because education alone reduces mortality from (though not incidence of) cervical cancer. We maintain that inclusion of a no-screening arm in such a randomized trial is problematic, not because education is unimportant, but because any negative findings from that arm will not be generalised to other settings, and any positive findings from that arm will be considered redundant… Correspondingly, women enrolled in no-screening arms of randomised trials comparing cervical screening with no-screening should be reassigned to screening arms without further delay.”

So my first published call to end these trials came in March 2004.

You have encouraged other cancer experts to consider debating the scientific basis of such no-screening arms in trials.
In April 2004, I joined a group of experts at a gathering at IARC in Lyon to co-author the IARC Handbook on Cervix Cancer Screening. This handbook was to include a section on the cervical screening trials being conducted in India. A few weeks before the April 2004 meeting, I emailed a copy of my March 2004 article to all of the handbook co-authors. At the meeting, I suggested to the co-authors that our handbook should include an explanation of the scientific justification for including no-screening control groups in the India trials, as some of the readers of our handbook would probably be interested in such an explanation. I knew the suggestion would require a debate among the gathered experts about whether any such justification actually existed, which was the point of my original suggestion. I was not surprised when my suggestion was summarily dismissed.

Did BMJ indicate that they had any doubts about publishing your paper? Were you surprised when they withdrew it?
Our paper was well-received by BMJ Global Health. The acceptance notification we received included the statement “Well done and many congratulations on an excellent paper!” BMJ selected our paper for press release, which further indicated that BMJ held our paper in high regard. Our paper was published on April 26. I was shocked when BMJ subsequently and abruptly decided to retract our paper.

They have cited legal reasons for withdrawing the paper. What do you think of that?
I wish to reiterate that, from a factual and scientific perspective, BMJ held our paper in high regard. I believe it is inappropriate to retract factually and scientifically valid papers out of concerns for the possibility of subsequent litigation.

I think it is unrealistic to believe that those whom our paper criticises would respond to such criticism with litigation, rather than scholarly debate. Our paper was critical of Tata Memorial Hospital, the US National Cancer Institute, the World Health Organisation’s International Agency for Research on Cancer and the Bill & Melinda Gates Foundation. It is implausible that any of these institutions would haul BMJ into the tabloid circus of English defamation courts to dispute the findings of a peer-reviewed, editorially-approved publication. These institutions would bring shame on themselves by doing so. These institutions would dispute such findings as they always have, through published scholarly debate.

One of the objectives of our paper is to stimulate such debate, especially in regard to the critically important issue of the health and welfare of the low-income women who had volunteered as study subjects.

Do you plan to take any steps to ensure this paper is published somewhere else?
My co-authors and I have appealed to BMJ’s ethics committee to reverse BMJ’s decision to retract. If BMJ’s ethics committee upholds BMJ’s decision to retract, we will re-submit our paper to another journal.

Are there any such similar trials elsewhere in the world, especially developing countries?
It would be highly unrealistic to assume that the cervical screening trials in India are the only problematic trials being conducted in developing countries.

Do you feel different standards are applied by aid agencies for developing countries while deciding on policies related to public health?
The unethical cervical screening trials in Mumbai, Osmanabad and Dindigul prove beyond the shadow of any doubt that aid agencies vigorously promote policies for developing countries that would be summarily condemned in the home countries of these same agencies.