Unwanted results: the ethics of controversial research

The recent publication of a study on psychiatric admission rates among low-income women after abortion and childbirth1 has elicited a barrage of letters to CMAJ of a pitch that we do not frequently encounter (page 101).2 We are chided for publishing flawed research and told that we should be ashamed of publishing the “opinions” of self-evidently biased researchers. We are accused of doing a disservice to women, medicine and the Journal, of failing to conduct proper peer review, and of not adequately scrutinizing the credentials of the authors.

The abortion debate is so highly charged that a state of respectful listening on either side is almost impossible to achieve. This debate is conducted publicly in religious, ideological and political terms: forms of discourse in which detachment is rare. But we do seem to have the idea in medicine that science offers us a more dispassionate means of analysis. To consider abortion as a health issue, indeed as a medical “procedure,” is to remove it from metaphysical and moral argument and to place it in a pragmatic realm where one deals in terms such as safety, equity of access, outcomes and risk–benefit ratios, and where the prevailing ethical discourse, when it is evoked, uses secular words like autonomy and patient choice.

Hence, perhaps the thing that is most offensive to some of our correspondents is the apparent co-opting of the medical view by persons they believe to be unqualified — or disqualified. The attack in our letters column is largely an ad hominem objection to the authors' ideological biases and credentials. There are two questions here: first, does ideological bias necessarily taint research? Second, are those who publish research responsible for its ultimate uses?

The answer to the first must be that opinion can of course cloud analysis. In light of the passion surrounding the subject of abortion we subjected this paper to especially cautious review and revision. We also recognized that research in this field is difficult to execute. Randomized trials are out of the question, and so one must rely on observational data, with all the difficulties of controlling for confounding variables. But the hypothesis that abortion (or childbirth) might have a psychological impact is not unreasonable, and to desist from posing a question because one may obtain an unwanted answer is hardly scientific. If we disqualified these researchers from presenting their data, we could never hear from authors with pro-choice views, either.

The phrase used by Deborah Stone3 in her classic text on public policy — “No number is innocent” — might be read as saying that every statistical analysis is guilty of serving one political agenda or another. A softer interpretation is that quantitative analysis is always subject to contextual influences. Biases and expectations, pre-existing knowledge and methodologic habits all influence what kinds of hypotheses are tested and how. Ultimately, our measurements are delimited by what can be measured, and by what we choose to measure. Feminist critiques have frequently pointed out the failure of mainstream health research to “count women in,” whether in randomized controlled trials or in economic analyses of health reforms.4 When researchers do attempt to amass quantitative evidence in women's health the ideological stakes are high: evidence is trailing rather far behind politics. Thus, the values of self-care and patient empowerment exemplified by breast self-examination gave rise to a similar outcry against unwanted results published in this journal.5^,6^,7 But if it is true that more explicit research into women's health issues will point the way to better care, better outcomes and more equity in access, we cannot toss out data any time we don't like their implications. Nor can we leap from a single observational study to public policy. We must allow the gradual and honest accumulation of further evidence to confirm or contradict what we think we know.

Which brings us to the second question: Should we deny the publication of a study because it might be applied by one or the other side of a factionalized debate? It strikes us that the results of the study by Reardon and colleagues are neutral: they could be “used” to further the argument that abortion is undesirable; or to support arguments for better post-abortion counselling and support. We cannot second-guess such interpretations without unfairly imposing our own values on the research we choose to publish. — CMAJ