Scientific journals serve as an important conduit for the dissemination of new knowledge. Getting published in one of the big medical journals is major accomplishment. This doesn’t mean that every article published in, say, the New England Journal of Medicine is good, but it has undergone rigorous peer review and is at least somewhat likely to report a valid test of an interesting hypothesis. This peer review process is important in vetting the quality of medical research, and while it isn’t perfect, it’s better than most of the alternatives.

There are other ways of disseminating medical knowledge, many of which serve purposes other than scientific progress. If, for example, I wanted to hype a new medical procedure which I just happen to provide, I could write a blog or issue a press release. The former strategy is common at the Huffington Post, where a recent piecehyped “platelet rich plasma” injections, a procedure with little scientific validity. The author of the post provides a link to her institution, where the procedure is performed, without any disclosure of conflict of interest.

While this sort of informercial is pretty sleazy, there is a subtler way to hype your institution.  Science by press release is so common it’s hardly noticed.  Superficially, it’s in an institution’s best interest to control their message and to promote their research, but when the source of medical news is a PR department rather than a peer-reviewed journal the results can be unsettling, and can ultimately erode scientific credibility.

A case in point is a press release from SUNY Buffalo. The headline proclaims, “Hypnosis Can Help Control Pain Among Women with Metastatic Breast Cancer, UB Researcher Finds.”   Whenever I see a dramatic claim like this, I try to go straight to the primary source. In this case, the primary source is a journal (Health Psychology) that none of my parent institutions subscribe to, so all I have to go on is the abstract.  Still, an abstract can be pretty informative.

The abstract describes a study of 124 breast cancer patients:

This randomized clinical trial examined the effects of group therapy with hypnosis (supportive-expressive group therapy) plus education compared to an education-only control condition on pain over 12 months among 124 women with metastatic breast cancer.

Setting aside the problems I have with the study design and the plausibility of hypnosis as a phenomenon, the results were not very encouraging:

[T]he intervention resulted in significantly less increase in the intensity of pain and suffering over time, compared to the education-only group, but had no significant effects on the frequency of pain episodes or amount of constant pain…

I’m not sure what these results mean, other than hypnosis didn’t do that much. Cancer pain is usually treated with medications, and there was no arm in this study to compare hypnosis with this standard of care, so the extent of the benefit of hypnosis (if any) in entirely unknown. But the press release (and the study’s primary author) found a way to put these results in the best possible light:

“The results of this study suggest that the experience of pain and suffering for patients with metastatic breast cancer can be successfully reduced with an intervention that includes hypnosis in a group therapy setting,” according to Butler. “These results augment the growing literature supporting the use of hypnosis as an adjunctive treatment for medical patients experiencing pain.”

The study found a small effect on one measurement of cancer pain. To telegraph this into such a broad and strong statement is, generously speaking, premature.

Testing hypotheses and publishing an honest assessment of the results is hard work. But in the end, it’s much better for science than promotional blog posts and press releases.

References

Butler LD, Koopman C, Neri E, Giese-Davis J, Palesh O, Thorne-Yocam KA, Dimiceli S, Chen XH, Fobair P, Kraemer HC…. (2009) Effects of supportive-expressive group therapy on pain in women with metastatic breast cancer. Health psychology : official journal of the Division of Health Psychology, American Psychological Association, 28(5), 579-87. PMID: 19751084

To read more from Peter Lipson, visit White Coat Underground.