research were women (NIH, 2011). Pinn pointed out that the overall percentage of female participants in NIH studies has varied “over the years from 66 down to 49 percent depending upon what large studies may be underway.” That presents a somewhat different picture of women in clinical trials from that presented in the report of Jagsi and colleagues (2009),5 which found that women made up less than 40% of participants in a set of cancer clinical trials.
Workshop participant Nancy Lee, deputy assistant secretary of health for women’s health and director of the Office on Women’s Health of the U.S. Department of Health and Human Services, said that issues of sex-specific reporting vary among disciplines, for example, in experimental studies, epidemiologic studies, and surveillance reporting. There has not been enough separation by sex in surveillance reporting, she said, and she stressed the importance of holding surveillance and epidemiologic publications to the same standards as experimental and clinical reports. Lichtman agreed that it is important to report data on men and women in surveillance reports and noted that the challenge is to draw comparisons between men and women, making sure that there is adequate power to ensure confidence in conclusions. Consideration of sex-based differences is important not only in peer-reviewed literature but in grant applications, Lichtman added.
Journal of the National Cancer Institute: Trailblazing the Way on Sex-Specific Reporting
Barnett Kramer, editor-in-chief of JNCI,6 noted that JNCI was the first journal to include instructions for addressing the effects of sex as
5See discussion by Adler on page 2.
6JNCI is owned and published by Oxford University Press and is not affiliated with the National Cancer Institute or the federal government.
part of its manuscript-preparation policy. Specifically, the JNCI instructions for authors state that “where appropriate, clinical and epidemiologic studies should be analyzed to see if there is an effect of sex or any of the major ethnic groups. If there is no effect, it should be so stated in Results.”
The inclusion of that statement, Kramer explained, was the result of a telephone call from ORWH director Pinn. By the end of the conversation, he said, he viewed the lack of such a statement as an oversight (not something that needed to be debated), and JNCI simply developed language and updated the instructions. Five in-house PhD-level senior editors at JNCI edit all manuscripts for content and ensure that the quality of the science meets journal standards, including appropriate sex-specific reporting.
Peer reviewers look at content and statistics and may also note whether there are sex-specific analyses, but the responsibility for ensuring rigor and determining what falls under “where appropriate” in the policy for sex-specific reporting falls to the in-house editors at JNCI. It is their interpretation of the instructions for authors that determines when sex-specific reporting is appropriate, Kramer clarified. It is not left to the authors to decide.
The JNCI policy emphasizes human studies, Kramer said, because it is sufficiently challenging to extrapolate data from in vitro and animal models to human applications. Authentication of cell lines used in in vitro studies is a particular concern. For example, the cross-contamination of cell cultures with HeLa cells is a substantial problem for laboratories, and researchers may not even know that they have female-derived HeLa cells in their cultures. JNCI tries to ensure that the lineage of cell lines is accurately identified in manuscripts; it asks authors to provide evidence that a line was obtained from a reliable source or that DNA “fingerprinting” has confirmed the lineage. Another issue is the predictive value of animal studies, and Kramer noted that animal models have not served the field of cancer research very well.
Another concern is analysis. As discussed by Wong, the more groups compared, the greater the likelihood of false-positive or false-negative findings and of spurious statistical interactions. If the number of events is small, the results may be less reliable, and authors or editors will often state this as a caveat.
There have been no challenges to implementing the JNCI policy, Kramer said. The editorial board agreed to the changes, and authors were made aware; there has been no pushback since the institution of the policy.
The International Committee of Medical Journal Editors: Many Editors, One Policy
Over the last decade or so, effective changes in editorial policies have generally been instituted journal by journal, Davidoff explained. Actions by associations of editors have been minimal. One exception is the International Committee of Medical Journal Editors (ICMJE), which started as the Vancouver Group, a small group of journal editors with no budget and no staff but a common interest. They began to consider some of the issues that editors face and together, as the ICMJE, developed and announced the Uniform Requirements for Manuscripts Submitted to Biomedical Journals.7 For a manuscript of a clinical trial to be considered for publication in a journal that adheres to these requirements, for example, the authors must declare that all investigators had access to all the data. Another requirement is that the trial have been registered in a public registry. The ICMJE is an example of how the editorial community can function and have an impact as a community. Davidoff stressed that policy changes created by a representative group should not be mandatory. The Uniform Requirements are voluntary; journals choose to follow them and state in their instructions for authors that they do so. In addition, policies should be authoritative rather than authoritarian.
Jerome Kassirer, former editor-in-chief of the New England Journal of Medicine, agreed, noting that the clinical journals by and large adhere to the policies of the ICMJE, but they can and do diverge if they decide that their own policies are more appropriate for them. He added that although the ICMJE developed solid, evidence-based policies for how medical journals should operate, there is no comparable group of basic-science journal editors.
Kramer noted that sex-specific reporting is also part of the ICMJE Uniform Requirements, which state that “where scientifically appropriate, analyses of the data by such variables as age and sex should be included.”
Early on, the push to gather more information about differences between men and women through enrollment of women in clinical trials was thought of as a feminist issue, said Marianne Legato, founding editor
of Gender Medicine. Some hospitals compounded that notion by using women’s health as a marketing tactic. Decades later, there is still pushback, and Legato opined that the academic community is still not completely convinced that sex-specific medicine is relevant and necessary. One of the issues that remains for providers is whether sex-specific practical medicine makes a difference in outcome for a given patient sitting in the office. Another issue is whether it is economically feasible not only to delineate the differences between men and women but to apply them to a clinical population.
Objections to including women are still rampant, Legato said. Many investigators and practitioners believe that “gender medicine” is a politically motivated and scientifically indefensible discipline. Including both sexes increases the expense of studies and strains on already restricted research funding. The trajectory of a given disease (such as coronary arterial disease) is often quite different in men and women in timing, characteristics, and symptoms, and inclusion of both sexes is fraught with such issues as concomitant conditions in older patients, hormonal differences, and differences in the physiology of aging. Many researchers think that including premenopausal woman in clinical studies is potentially dangerous to their reproductive function and is potentially dangerous to any child conceived during the course of the trial, and women of childbearing age are often fearful and difficult to recruit.
Despite the resistance to studying women, there is growing acknowledgment that men and women have substantial and widespread biologic differences. There is more awareness of the extent and complexity of the sex-specific and gender-specific properties of living organisms.
In addition, there is a growing understanding of the human genome and increasing incorporation into research of ever-more effective ways of measuring the genome’s effect on human biology. Some of the important questions to be addressed, given the complexity of how phenotype is determined (not just by genes but by environment and experience), are: How useful will delineation of a person’s genome be in predicting disease and choosing therapy? Is it ever possible to separate what is hard-wired into the organism by virtue of biologic sex? What is the result of the effect of other factors on the phenotype? What are the effects of biologic sex on gene expression?
Finally, the emerging field of synthetic biology is transforming our understanding of what constitutes life and raising questions about what it means to be human. To what extent will the human phenotype be augmented or changed? Legato asked.
Legato explained that the first iteration of Gender Medicine, then called the Journal of Gender-Specific Medicine, was not successful and was discontinued. As the editor, Legato faced the challenge of explaining to researchers that gender-specific medicine is not women’s health, but a genuine comparison between the sexes. The relaunched journal, Gender Medicine, is now in its 5th year of publication and is more successful in impact and content.
Kramer said that it is not possible to have identical guidelines for preclinical animal studies and (human) clinical studies, but JNCI strives to maintain the same level of rigor in every field in which it publishes, namely, the entire spectrum of research that is fundable in the National Cancer Institute. About one-third of the papers report bench-oriented research in the laboratory, one-third epidemiologic research and observational studies, and one-third clinical studies of therapeutics, prevention, or screening. Every study must be reported in such a way that someone else could verify the results by repeating the study exactly as reported. Thus, one must know the sex of the animals used in a study, and reporting of sex should be an absolute requirement, Kramer said.
Jeffrey Blaustein, editor-in-chief of Endocrinology and a neuroendocrinologist, commented on the numerous aspects of animal studies that one could monitor: randomization; blinding; adverse environment, such as nearby construction; age of animals at shipment; interruptions in circadian cycles; and so on. It is not now possible to monitor submitted manuscripts for discussion of all variables in animal studies, and in many cases one has to assume that randomization or blinding, for example, has been done.
Kramer responded that the “rules of engagement” are more mature for clinical studies, and there are standard checklists for required elements, such as randomization. The JNCI in-house editors have checklists for both preclinical and clinical studies. In the clinical literature, when editors are rating the quality of reporting in a manuscript and, for example, how the randomization was done is not reported, the manuscript loses points on the checklist. There is no assumption that the authors did it if they do not report it. The same thing should be done for reports of animal studies, Kramer asserted. If blinding and randomization are not mentioned, he said, he suspects that the author did not know enough about the process and about where biases can be built in to report it.