PHOENIX — In early June, an Indian medical researcher traveled to Chicago to report seemingly astounding results from a decadelong study of cervical-cancer screening in the slums of Mumbai.
Using a simple, inexpensive vinegar test, researchers had cut the rate of death from cervical cancer by 31 percent, Dr. Surendra Shastri told more than 5,000 doctors at the American Society of Clinical Oncology's annual meeting.
"He has become one of my heroes," Dr. Sandra Swain, ASCO's president, told the gathering, hailing Shastri for results he said potentially could save 22,000 lives a year in India and 72,000 lives a year around the world.
But in widely reported videos and press releases trumpeting the results, neither the society nor Shastri delved into ethical and other questions that had been raised behind the scenes before the conference, in part as a result of earlier reporting by The Arizona Republic.
These included concerns about the findings, their broader relevance, and how they were achieved — and, in particular, why researchers put half of the women in their study, more than 76,000 — in a "control" group.
Those women were not screened or routinely treated for cervical cancer but were monitored as their cancers progressed, and at least 98 of them died. That is up from 15 deaths reported by The Republic six months ago.
In addition, the clinical-oncology society featured the study despite findings a year ago that women in the control group had not been properly informed and had not given adequate informed consent.
"It's very troubling," said Elizabeth Woeckner, president of Citizens for Responsible Care and Research,a New York non-profit that seeks better safeguards for human research subjects.
Why, Woeckner asked, would any woman agree to continue in an unscreened control group for a decade if the benefits of cervical-cancer screening were properly explained?
"If someone offers me the opportunity not to be treated, why am I going to go there?" she asked.
The Republic first reported in February on questions about the Mumbai study and another cervical-cancer study in rural India that included 31,000 women who were tracked but not screened or routinely treated.
Sixty-four women in the second study's control group had died of cervical cancer as of 2009, the most recent figures released.
Both studies were U.S.-funded: the Mumbai study by the National Cancer Institute, the other study by the Gates Foundation. The foundation declined to comment.
As The Republic reported, investigators for the federal Office for Human Research Protections determined last year that researchers at Mumbai's Tata Memorial Hospital didn't adequately inform the women in their control group about alternatives for getting cervical-cancer screening and that the women didn't give adequate informed consent — a keystone requirement for medical research involving human subjects.
The Republic noted that 15 women in the Mumbai control group had died of cervical cancer by 2009, the most recent data reported at the time. But new data Shastri reported in Chicago show 6 1/2 times that many deaths by the end of last year.
Medical researchers raised other questions about the study:
• Whether having an unscreened control group was scientifically necessary
• Why the NCI and Tata included an unscreened control group when similar cervical-cancer studies in Pakistan, Guatemala, South Africa, Zimbabwe, the Philippines, China and other developing countries did not rely on unscreened control groups
• Whether all women in the Mumbai control group should have been screened after a vinegar-screening study in India's Dindigul district published results in 2007 showing that such screening cut cervical-cancer death rates by 35 percent.
• How relevant the results of the Mumbai trial are for other developing countries, given that a score of other acetic-acid studies from such countries showed highly variable results.
Stakes are enormous
An estimated 72,000 women die of cervical cancer in India each year, more than a quarter of the estimated 276,000 annual deaths from the disease, said a May report by the Cervical Cancer Free Coalition, a U.S.-based advocacy group.
India's cervical-cancer death rate is seven times that of the United States, where Pap tests and screening for human papillomavirus, which causes nearly all cervical cancers, are routine.
Some researchers, including Shastri, argue that the costs of Pap and HPV tests are a barrier in India.
"We do not have the funds, cytopathologists and laboratories to provide Pap smear screening to these poor women," Shastri wrote in an email to The Republic. He wrote that the purpose of the Mumbai study was to see whether the method known as visual inspection with acetic acid, or VIA, could provide a viable alternative in developing countries.
In VIA screening, a trained worker swabs an acetic-acid solution on the cervix, then looks for abnormalities (such as precancerous lesions), which turn white when exposed to vinegar. The Mumbai study divided more than 151,000 women into two groups. Women in the study group underwent four rounds of screening with acetic acid over eight years. They were provided free treatment at Tata for any abnormalities detected.
Shastri said the women in the control group were offered cancer education at the time they were recruited. All the women signed or placed thumbprints on consent forms.
The original grant application said that although researchers would follow-up and track cancer cases among women in the control group, "no further education on prevention or early detection will be imparted."
Tata provided an English-language version of the model consent form the women signed to the National Cancer Institute in a 2005 grant request.
The Republic obtained it through a Freedom of Information Act request. The form stated: "The standard screening procedures for cervix and breast cancers are Pap smear and Mammography, in developed countries. Such facilities are also available in some centres in India and you may choose to undergo these tests on your own, if you do not wish to participate in this study."
But in a July 2012 letter to Tata, the federal Human Research Protections Office director of compliance oversight, Kristina Borror, noted that, among other issues, "a translated copy of the informed consent document actually used in the study that we were provided did not include this language."
A consent form from 2000 obtained by The Republic through a FOIA request also lacks that language.
Tata researchers said women were educated andinformed verbally, but Borror wrote: "We believe that subjects were not provided with adequate information to understand the differences between the research procedures and mammography and Pap smears. Understanding what their alternatives are is key to the subjects' ability to make an informed decision about whether or not to participate in a research study."
Dr. Eric Suba, a San Francisco-based pathologist who runs a Pap-based cervical-cancer screening program in Vietnam, filed the complaint in May 2011 that led the Human Research Protections Office to investigate the Mumbai study.
"How do you maintain an unscreened control group for 15 years without keeping these women in the dark about basic biological facts?" he asked in a telephone interview.
"Once you tell women about what cervical cancer is and what screening can do for them, they demand screening services."
In his complaint, Suba raised the informed-consent questions and argued that using women's deaths as the yardstick for success was scientifically gratuitous and unethical — that extensive results from studies around the world already demonstrated that screening is better than not screening.
But the research-protections office declined to present the mortality-rate allegationsto NCI and Tata, Borror said in a July 18 letter, "because there did not appear to be violation of the HHS human subject protections regulations." HHS is the Department of Health and Human Services.
Women not screened would likely seek treatment for cervical cancer only once their symptoms, such as pelvic pain or vaginal bleeding, become obvious. By that point, cervical cancer has reached an advanced stage.
As The Republic reported, Sarah Kobrin, NCI's program director for the study, said that it wasn't sufficient to show that acetic-acid screening detected cancer sooner than in women referring themselves because of symptoms and that the study needed to continue "until we could see a statistically significant difference" in mortality rates between the control group and the screened group.
In a recent interview, Shastri told Global Oncology, a Harvard-based cancer-research non-profit, that "since the standard of care for cervical-cancer screening in India is no screening, we were ethically justified in having a no-screening control" group.
In an email to The Republic, Shastri said women in the control group were made aware that they could leave the study and receive free screening at Tata. It isn't clear how many understood. Study data reported that fewer than 5 percent of the women recruited were literate; in the control group, fewer than one in 500 sought screening.
Shastri also noted the Human Research Protections Office's final determination: Borror said corrective actions taken by Tata, including offering screening now to all subjects in the control group and informing them about other options such as Pap smears, adequately addressed the department's concerns.
On Friday, the Indian Parliament's committee on health and human welfare, in a report criticizing a separate U.S.-funded HPV vaccine study, reiterated that obtaining proper informed consent from study subjects in any medical studies "is a fundamental requirement … to ensure that the human rights of the study subjects" are protected.
Relative standard of care
That Human Research Protections Office determination played a key part in discussions at the clinical-oncology society, said Dr. Douglas Yee, an oncologist and the chairman of the society's scientific program committee. "We were aware of the concerns brought to OHRP," he said. "We spent a lot of time delving into that particular matter."
In the end, the committee agreed that not offering screening was the standard of care because "this was a population that doesn't have access to health care."
Standard of care helps define researchers' obligations to their human test subjects.
In the United States or Europe, where screening is routine, including an unscreened control group would be barred by regulation. Some researchers argue that ethical standards shouldn't be relative.
But the oncology society committee agreed with Shastri that researchers' obligation was merely to provide the women the local standard of care, even if that was no care.
"I agree that no screening in most parts of the world would not be appropriate," Yee said.
"Dr. Shastri made it clear in his presentation that no screening is the standard of care for a huge part of the population," so the oncology society agreed with "the NCI peer review of the original grant that the control arm of no screening was appropriate."
Yee said the committee was aware of the lack of adequate informed consent but decided that because the Human Research Protections Office said the issue had been addressed,it was no longer an impediment.
He said the committee considered the 2007 Dindigul study, which found a 35 percent drop in mortality after one round of acetic-acid screening, but they decided that while "the Dindigul study was another important piece of data to support VIA screening … it was also important to further demonstrate the magnitude of mortality reduction as done in the Mumbai study."
Statistically, the results Shastri reported in Chicago in June, after four rounds of VIA screening, differ markedly from the study's results after one, two or three rounds, as published in the International Journal of Cancer in 2010.
After two rounds of screening vs. monitoring the control group, more cancers had been confirmed through self-referrals in the control group, 28, than in the similar-size screened group, 12.
After three rounds, more women had died of cervical cancer in the screening group, 17, than in the control group, 15. But after more than eight years and four rounds of screening, as more of the women in the control group succumbed to their cancers, cervical-cancer deaths rose to 98 in the control group vs. 67 in the screening group.
Shastri, in an email, said the study held "the highest level of merit in terms of both science and medical ethics," and it shows that acetic-acid screening is a feasible, affordable strategy in places where other forms of screening aren't affordable or available.
But Woeckner, of Citizens for Responsible Care and Research, said she's not convinced.
"Asking, 'Does it work in this city or town?' — that's a marketing trial, not a genuine scientific question," said Woeckner, whose mother died of cervical cancer.
She said she keeps thinking about the women in the control group being monitored as they sickened and died.
"You can't let people die to show something you already know," she said.