I found this article fascinating and informative. The discussion appropriately explores the reasons for making data publicly available and the barriers. One of these is the need for industry - a sponsor of most of the RCTS of new compounds and owners of the resultant data - to protect its proprietary interests and to shield itself form unwanted further analyses of data that might reveal damaging results (a less effective compound, or a greater frequency of adverse events, for examples), and to protect itself from lawsuits.

I was reminded of the discussions around the diet drugs in the early 1990s. A good example is the diet drug fenfluramine which was withdrawn from the market in 1997 after it was shown to be associated with pulmonary hypertension and later, in a combination tablet with phentermine (fen-phen), to be associated with mitral valve disease. Both of these results needed to be confirmed by independent studies, particularly the association with mitral valve disease, the diagnosis of which was made on the basis visual interpretation of echocardiograms.

One withdrawn from the market, tens of thousands of patients who took these medications, and their physicians would have benefited from further analysis of the data by independent investigators.

Having been withdrawn from the market, there was no other way to develop data and do the analyses that would confirm the original reports. An opportunity lost.

The success of the effort to make registration of clinical trials compulsory (by many funding organizations and in the US Senate with the FDA bill that will move forward this fall and which requires registration of all clinical trials) is a model that could be used to encourage efforts to make data publicly available. Perhaps the International Committee of Medical Journal Editors, a major catalyst for the trial registry legislation in the US, should issue a similar statement regarding data sets. (competing interest, I was a member of ICMJE)

It will take, however, legislation to level the playing field for pharmaceutical and device manufacturing companies and enable them by compulsion to post the data used in their clinical trials.

There are other important problems to overcome, as noted by the authors, and perhaps the most important of these is to establish easy to use and standardized protocols for data deposition. This will depend on the type of data, as pointed out by the authors.

May I also note that in addition to the efforts described by the authors, Equator (Enhancing the Quality and Transparency of Research) a UK based group under the leadership of Doug Altman, is attempting to both improve methods of reporting research and of establishing collaborations between investigators and others interested in promoting these goals. If the essential elements need for research of a particular type can be defined, then the data storage and retrieval elements needed can also be established and hopefully implemented. (competing interest - I am on the steering cte. of Equator)


Abenhaim L, Moride Y, Brenot F, Rich S, Benichou J, Kurz X, et al. Appetite-suppressant drugs and the risk of primary pulmonary hypertension. N Engl J Med 1996;335:609-16.

Connolly HM, Crary JL, McGoon MD, Hensrud DD, Edwards BS, Edwards WD, et al. Valvular heart disease associated with fenfluramine-phentermine. N Engl J Med 1997;337:581-8