The $8.5-million, 12-year study designed to settle the question of the birth control pill's safety has instead reignited the controversy, and has provoked attack from prominent members of the scientific community.
Scientists inside and outside the government have questioned the federally funded study's validity, the conduct and objectivity of the researchers, and the relationship of pill manufacturers to the research, which is intented to evaluate their product.
Starting in 1968, a medical team in Walnut Creek, Calif., recorded and analyzed each scrap of data on the health of more than 16,000 women, seeking to define the pill's effects on every organ in the body.
Last October, seven months after submitting their final report to the government, the California researchers declared the pill's hazards "negligible," at least for healthy, white, middle-class women. Contradicting the findings of most other research, they said women who took birth control pills had no increased risk of heart attacks, blood clots, strokes or death from any cause.
But even as the 250-page report is readied for publication at the Government Printing Office, serious questions have been raised about it. The following are some of the concerns:
The validity of the conclusions: Some scientists say the study proves nothing, because it did not examine enough women who were actually taking the pill. They say too many of the women studied were past users, or went off the pill soon after the study began.
The analysis of the data: Experts at the National Institutes of Health and elsewhere say that in their final report the researchers analyzed some statistics in a way that obscures the pill's possible side effects. For instance, scientists working on the project reported in 1978 that they had found an eightfold increase among pill-users in the risk of blood clots in leg veins. But in the final report, a new analysis of the same data concluded that the risk of such clots was not higher in pill-users.
The behavior of the scientists: The study's authors, Dr. Savitri Ramcharan and Dr. Frederick Pellegrin, summarized their unpublished findings last October at a university symposium sponsored by G.D. Searle and Co., a major birth control pill manufacturer.The researchers then cooperated with American and Canadian public relations campaigns financed by two manufacturers, which resulted in newspaper, television and magazine reports that declared the pill's risks "negligilbe."
National Institutes of Health officials who had overseen the study took issue with the sweeping verdict, and were shocked by the scientists' connection with the campaign. "Her [Ramcharan's] conclusions are unjustifiably sanguine," said Dr. Philip Corfman, director of the Center for Population Research at the National Institutes on the pill was misleading because ". . . it's a shift so much away from what is reasonable to conclude from this study. . . ."
The role of pill manufacturers: In the mid-1970s, five manufacturers contributed $149,000 to finance a part of the Walnut Creek project that the government had decided to drop. The companies did not request changes in the study's design, and the government later renewed funding for the work. But the fact that manufacturers had helped finance an evaluation of the pill raised the question of conflict of interest. "It didn't look very good," said Dr. Nicholas Wright, who until Jan. 1 headed the contraceptive evaluation branch at the center.
It tooked worse when Searle publicized the project's favorable findings, without mentioning the manufacturers' mentioning the manufacturers' financial support. "For them [the drug companies] to tout this as a U.S. government study and say the pill risk is negligible . . . is the old Madison Avenue technique," Corfman said. "It's not illegal, but I find it improper."
Ramcharan, who was in charge of collecting and analyzing the data and who wrote the final report, defended the validity and significance of her conclusions. She said that if pill-users indeed had five times the death rate of other women -- as a British study found -- the Walnut Creek study would have detected it.
"In general I stand by the conclusions from the Walnut Creek study," she said, although she feels they have been taken out of context by the media. In particular, she said she believes the study shows that the pill does not significantly raise the risk of heart and blood vessel diseases in women who do not smoke or have other risk factors.
She denied that the drug companies' financial contribution represented a conflict of interest, since the companies did not influence the study's design or findings. She said her cooperation with attempts to publicize her conclusions arose from a belief that the public had a right to know her results, since the study was in the public domain.
Ramcharan's career includes contraceptive testing in India and research at the Worcester Foundation from 1957 to 1959 on the prototype of the pill. In 1962, she changed fields from biochemistry to public health, and in 1971 became director of the Walnut Creek study.
Despite her early work on the pill's development, she said, ". . . I consider myself a scientist and medical researcher. As such, I do not think that I have biases that would interfere with my objectivity in evaluating the risks of . . . oral contraceptives."
Pellegrin, who said he administered the project but was not responsible for its scientific execution or analysis, said the findings "clarified that there are people who are high-risk for taking the pill and who shouldn't take it. But by far the majority of women under 30 are healthy" and do not increase their risk of serious diseases by pill use, he said.
However, he said some reports of the stuy in the press had given the false impression that the pill was safe for everybody. "I do not think it [the publicity] has done any damage," he added.
The controversy over the Walnut Creek project began last October, when Ramcharan and Pellegrin presented their findings at a University of Pennsylvania symposium financed by the pill's original producer, G.D. Searle.
According to a Searle spokesman, the company gave the university a $25,000 grant to support the symposium after it was approached last spring by a group of gynecologists "who knew about the new research findings and felt they had not been communicated properly." Ramcharan and Pellegrin agreed to present their still-unpublished conclusions, after first clearing the proposal with NIH.
The Searle spokesman said the symposium had nothing to do with the drug companies' earlier financial support of the study. "We took a symposium and helped communicate those results to people," he said. "The allegation that $150,000 in 1975 influenced those results in just categorically untrue."
At the conference, Ramcharan reported the Walnut Creek team's major findings:
Past and present pill-users had not higher death rate than those who had never taken it.
There was "no increase in risks from circulatory diseases [heart attacks, strokes and blood clots] such as had been ascribed to oral contraceptive use by studies . . . in other populations."
Pill-users showed no increased risk of cancer of the breast, ovary or endometrium (the lining of the womb).
Pill-users had a higher risk of melanoma, a maligant form of skin cancer, and of lung cancer. Ramcharan theorized that the former finding might be because they sunbathed more, and the latter because they smoked more.
Pill-users had a higher risk of cancer of the cervix (the mouth of the womb). According to her analysis it was explained by increased sexual activity, which has previously been established as a risk factor for the disease.
Commenting on the study's implications as a guest on "Good Morning, America" last Dec. 10, Ramcharan said "the pill in our population [the group studied] was not as risky as had been reported in other studies . . . and where we did find risks . . . they were related to factors other than pill use.
"I don't think you can conclude from our study or any that has been done that the pill is absolutely safe or absolutely dangerous. . . . But particularly the younger, healthier women could feel more reassured," she said.
Corfman objected to her assessment.
"What's 'young'? And what do the words 'young, adult, white' mean?" he said. "It's so fraught with conditions and value judgments that I find it's not a useful public health statement."
The Walnut Creek project's fatal flaw, according to scientists at the population center, was that it studied too few women who were actually on the pill. Although those who entered the project in 1968 comprised equal groups of current users, past users and women who had never taken the pill, 70 percent of them were 30 or older.
As the study progressed, many went off the pill because of side effects or age. By 1978, only 2,000 of the original 16,638 women were still taking birth control pills, according to Dr. Heinz Berendes, formerly the project officer in charge of the study and now director of epidemiology and biometry at the National Institute of Child Health and Human Development.
Although little was known about the pill's risks when the study was designed, its worst side effects -- diseases of the heart and blood vessels -- are now thought to occur in women who are on the pill, not in past users. For this reason, the Walnut Creek study "certainly does not disprove anything that has emerged" about such side effects, according to Dr. Martin Vessey, a prominent British public health researcher who reviewed the study for NIH.
Vessey said the study's conclusions were based on too few cases. For example, regarding heart attacks, he said, "In saying the result of the Walnut Creek study are different from . . . others, we're arguing about the difference between two deaths and four deaths."
Ramcharan acknowledged that the study's finding that the pill did not increase the death rate was based on "two few [women] to rule out the effect of chance on the results." But she said the numbers would have been sufficient to detect a fivefold increase in the death rate, if such a difference had existed. The impact of other risk factors, such as smoking, on deaths from heart disease did show up.
She added that heart attack is such a rare disease in women under 40 that no research project can study it adequately. In the Walnut Creek study, only one heart attack occurred in a woman under 40, a heavy smoker and past user of the pill. The two women over 40 who had heart attacks while on the pill were also smokers, she said.
"We concluded that the findings in our study gave no support for an increased risk of [heart attack] in [pill] users that was unassociated with smoking," she said. The researchers reached the same conclusion for other kinds of blood vessel disease, she said, "even though the number of case in each specific disease category was small. . . ."
Berendes said the failure to find a higher death rate among pill-users might have been because the women studied, members of a Kaiser-Permanente health plan, were healthier than average.
"It's a middle-class population. It's a healthy population," he said. "The mortality . . . of these women [in the study] is lower than that of average California women of the same age group."
Ramcharan commented that any drawbacks of conducting the project on the health plan members were balanced by the advantages of obtaining reliable data from a stable, homogeneous group. She added that white, middle-class suburban women represent "rather a large segment of the female population of the United States."
Scientists who had reviewed the study said it treated statistics in ways that obscured the pill's possible side effects. Vessey noted that an article published by Walnut Creek researchers in 1978 showed that pill-users had eight times as high a risk of developing blood clots in leg veins. But in the final report, the researchers analyzed the same numbers differently -- combining cases where clots happened spontaneously with others where a clot developed following surgery or an injury -- and concluded that the risk was not increased.
Another scientists noted that Ramcharan, in her final report on the Walnut Creek study, differed from her earlier published conclusions on the pill's effect on cancer of the cervix. In the final report, Ramcharan said pill-users had more cervical cancer because they had more sex. Vessey's analysis showed that sexual activity could explain only part of the increase.
"She's gone into print three times saying there's an increased risk of cervical cancer, and now she publishes a paper where she says, 'I don't believe any of that bad staff I said before,'" the scientist said.
Ramcharan called the two differing reports on the risk of blood clots "essential aspects of the same study." She said that because of technical limitations of the statistical method used for the final analysis, the earlier report, which found an eight-fold increase in blood clots among pill-users, considered a number of factors not analyzed in the final report. She agreed that failure to analyze such factors "would be likely to mask the [pill] risk that might be present."
But she cautioned that the high risk of blood clots reported in the earlier paper might in turn be exaggerated by other statistical problems.
She said her early reports of an increase in cervical cancer due to pill use were tentative, since at the time she had no information of users' sexual behavior. However, Walnut Creek researchers obtained that information and adjusted for the known effect of sexual behavior on cancer risk. "The apparent increased risk of cervical cancer in pill users was accounted for by their higher sexual activity," she said.
A scientist familiar with the final report said Ramcharan seemed to strain for improbable explanations of her findings when the probable explanation was that the pill heightened the risk of a disease. "When one is standing on the American continent and hears hoofbeats, it is more likely to be horses than zebras," the scientist said. "I think every time Savitri hears hoofbeats, she sees zebras."
At the time of the October, conference, the Walnut Creek study's favorable verdict on the pill received widespread publicity, because Searle had arranged a press conference and press releases through its public relations consultant, Hill and Knowlton. News stories all over the country bore headlines such as "Birth Control Pills Are Called Safe" and "Pill Risks Appear Negligible," and Ramcharan appeared on television and in People magazine.
"Nobody at Searle is trying to say the pill is safe," said a Searle spokesman, when asked about the publicity. He said the company felt the pill's hazards had been exaggerated in the past, and wanted to make the Walnut Creek study's findings known. "The allegations that there is a continuing media campaign are not true," he said.
However, a month after the Pennsylvania symposium, Ramcharan reported her findings at a meeting of the Canadian Fertility Society in Toronto. The cost of her visit to Toronto. and of a press conference and press releases prepared for the meeting by Reidcorp Advertising, were paid for by another pill manufacturer, Weyth Ltd., according to Dr. Patick Taylor, the society's president.
Taylor said his society invited Ramcharan to speak because it members had read news reports of the Walnut Creek study and wanted a chance to evaluate her still unpublished results for themselves. He said the society, like other medical associations, relies partly on drug companies to support its meetings. He asked Wyeth to pay the society's cost of bringing Ramcharan to Toronto and the company agreed. Then, because he was trying to organize the meeting from his home in Calgary, he asked Wyeth to take care of informing the Canadian press.
Taylor said Weyth contacted Reidcorp Advertising, and when he arrived in Toronto, he was handed a press release that proclaimed, "Pill users healthier than non-users."
"Redicorp went ape," he said. "I'm not at all pleased, because my outfit is a reputable scientific outfit."
A Weyth official denied that his firm had paid Ramcharan's expenses, and would not comment on whether Weyth had paid for the public relations effort.
Pellegrin said that he and Ramcharan had been aware that the press would report their study's findings, but that he had been astonished by the degree of attention it received.
"There is big interest in the pill -- 6 or 7 million [American] women are on it, so obviously they're interested," he said. "Everbody just went bananas, and that was a surprise to me."
Asked about the media's reporting of the study, he said, "It may have overly emphasized the safety of it . . .
The fact that we were talking about people who were young and healthy wasn't covered."
NIH officials said that promoting the findings of an unpublished study was not generally accepted behavior in the scientific community. But Ramcharan said that since 1975, when her study's initial results were published, she had become accustomed to dealing with reporters, and had come to believe that consumers had a right to hear her findings as soon as they were available.
Some NIH officials lamented the fact that the manufacturers' contribution and the public relations campaigns had affected the credibility of the Walnut Creek project. They said it contributed detailed data that may yield significant information about the pill or other aspects of women's health.
Officials at the Center for Population Research, the NIH branch which controlled the funding, ended the project in early 1980 because it contained so few women who were still on the pill.
Pellegrin, who began the project at Kaiser-Permanente's health center in Walnut Creek, is employed by the Kaiser-Permanente health plan and will continue there now that the project is over. But Ramcharan, who directed the study from 1971 on and wrote the final report, was paid out of the NIH grant. A Kaiser-Permanente spokesman said she is no longer associated with the health plan, and is looking for funding to continue the study.
Ramcharan said she had not approached pill manufacturers about extending the Walnut Creek project.
The manufacturers have come to the rescue before. In late 1976 or early 1977, according to Pellegrin, five pill manufacurers -- Searle, Mead Johnson, Syntex, Ortho and Parke Davis -- contributed a total of $149,000 to pay for a part of the Walnut Creek project which the government had dropped. Later the government resumed funding for the week, although it was never completed, Pellegrin said.
Berendes, who was the NIH project officer when the contribution was made, said the government cannot prevent its researchers from accepting outside support, even from manufacturers of the product that is being evaluted. o
"I may have been responsible for setting up the first contact between the drug companies and Walnut Creek," he said."You know the real world. In the real world the investigator always wants more money than the sponsor's willing to give."