Correspondence

Which Research Results Should the Public Believe?

N Engl J Med 1995; 332:963-964April 6, 1995

Article

To the Editor:

With respect to Angell and Kassirer's editorial (July 21 issue)1 concerning the public's confusion about the results of clinical trials, I think the public and the media would be better able to assess research results if medical researchers themselves were more modest and careful when discussing their results with journalists. The Journal has led the way in embargoing research results until after publication, but such rules cannot prevent the ambitious or naive investigator (or the investigator's institution) from touting results and pushing conclusions beyond the limits of the data. This problem is likely to increase as corporate support for research grows, with all of its financial implications for the marketplace, researchers, and research institutions. We cannot and should not control the media, but we should be able to control ourselves and to place research results in their proper context.

Thomas A. Hazinski, M.D.
Vanderbilt University, Nashville, TN 37232

1 References

1. 1

   Angell M, Kassirer JP. Clinical research -- what should the public believe? N Engl J Med 1994;331:189-190
   Full Text | Web of Science | Medline

To the Editor:

As a full-time primary care practitioner, I find myself awash in a rising tide of recommendations regarding new medications, disease-management protocols, and the promotion of health. Unfortunately, it is often difficult to sort out what is important from what is not.

Much of today's medical advice is based on studies of large numbers of people demonstrating small but statistically significant differences attributed to specific interventions. Although statistically significant, these differences are often of questionable practical importance, particularly for a given patient. The determination of the practical importance of many clinical studies is confounded by the presentation of results in terms of relative risk or risk reduction. Information regarding incidence or survival rates would be more to the point.

The comments by Angell and Kassirer and by Keeney1 and Fitzgerald2 in the same issue address these very subjects and pose the question, “Why can't data be presented in a more meaningful way?” I urge the Journal to establish standards for the presentation of clinical data that would make it easier for patients to decide whether recommendations regarding disease prevention, evaluation, and treatment are important to them. Such standards would allow physicians to make more appropriate decisions about the allocation of medical resources within their practices. Policy makers would also find them useful.

Meeting these standards should be a prerequisite for the publication of clinical investigations as well as for the publication of any article concerned with the value of a given medical test or treatment. The information could be presented by the authors or provided by the editors in an abstract prefacing the article.

Howard T. Chatterton, M.D., Ph.D.
Marshfield Clinic, Ladysmith, WI 54848

2 References

1. 1

   Keeney RL. Decisions about life-threatening risks. N Engl J Med 1994;331:193-196
   Full Text | Web of Science | Medline

2. 2

   Fitzgerald FT. The tyranny of health. N Engl J Med 1994;331:196-198
   Full Text | Web of Science | Medline

Author/Editor Response

The editors reply:

Dr. Hazinski makes an important point.

We agree with Dr. Chatterton that the results of clinical trials are not always presented in a way that emphasizes the practical implications for individual patients. They are often instead presented from a public health perspective. But it is possible, of course, for a reader to derive one measure from the other — e.g., the approximate survival rate from risk reduction.

Marcia Angell, M.D.
Jerome P. Kassirer, M.D.