Reader Comments

Post a new comment on this article

Data Sharing: Easier Said than Done

Posted by plosmedicine on 31 Mar 2009 at 00:29 GMT

Author: Jonathan Leo
Position: Professor
Institution: Lincoln Memorial University
Submitted Date: September 09, 2008
Published Date: September 10, 2008
This comment was originally posted as a “Reader Response” on the publication date indicated above. All Reader Responses are now available as comments.

Dear Editors,
In their article, Towards a data sharing culture: Recommendations for leadership from academic health centers, Piwowar et al. (2008) make an excellent argument for sharing data, however there is a perception, or misperception, that needs elaboration. According to the authors, The National Institutes of Health’s (NIH) guidelines on data sharing are a model that Academic Medical Centers could emulate. However, while NIH does have a published policy statement, in several important cases, NIH has not followed their own guidelines. Here are two examples of some issues in data sharing that involve NIH funded researchers. 1) In 2003, a NIMH funded neuroimaging study reported that compared to controls, ADHD children had smaller frontal lobes (Sowell et al). The implication is that ADHD is associated with brain pathology. An alternate hypothesis is that that psychostimulants prescribed to ADHD children cause frontal lobe atrophy. In order to test this hypothesis, I requested data comparing medication-naïve ADHD children to those previously prescribed pscychostimulants. Because many of the ADHD neuroimaging studies are confounded by prior medication use by the ADHD children, this seemed like a logical comparison. However, Sowell et al would not provide this information even though an earlier NIMH sponsored study had reported on this same comparison (Castellanos et al 2002). I subsequently filed a Freedom of Information Act request, which was denied. On the NIH website, and cited in the paper by Piwowar et al. NIH states: “The NIH endorses the sharing of final research data to serve these and other important scientific goals. The NIH expects and supports the timely release and sharing of final research data from NIH-supported studies for use by other researchers.”Apparently NIH’s actions are not as strong as their words. I have been prevented from making a simple statistical comparison that could shed light on the potential iatrogenic effects of medications prescribed to millions of children. 2) Eli Lilly submitted two studies to the FDA as part of the approval process for the use of Prozac in children. The first study, which was funded by the National Institute of Mental Health (NIMH) and published in 1997, reported that “Side effects, as a reason for discontinuation, were minimal” (Emslie, Rush et al., 1997, p. 1033). There is no mention in the paper of any children attempting suicide. However, in a related document, FDA’s “Medical Review of Prozac,” written in 2001, there is a discussion about two children on Prozac attempting suicide (Center for Drug Evaluation and Research, 2001; Leo, 2004). This was pointed this out to Dr. Thomas Insel, Director of NIMH, but did not provide the information. I am simply seeking clarification, as the published study and the FDA report seem to be at odds.
In the case of clinical trials, in regards to the discrepancies between the published data and the actual data, many news organizations have focused on the pharmaceutical companies. However, the problems with data disclosure are not limited to industry, but extend to publicly-funded research projects, where the data is available but access is simply refused. A data-sharing paradigm for Academic Medical Centers that promotes scientific progress and the public health must include provisions that actually facilitate the sharing of such data, not in theory, but in actual practice.

Castellanos, F.X., et al. (2002). Journal of the American Medical Association, 288, 1740–1748.
CDER (2001). Application 18-936-SE-064. Medical review of fluoxetine,July2001.Available:
Emslie, G. J., A. J. Rush, et al. (1997). Archives of General Psychiatry 54: 1031-1037.
Sowell, E.R., et al. (2003). The Lancet, 362, 1699–1707.

No competing interests declared.