In May, the Lancet published a peer-reviewed study claiming to analyse data from nearly 15,000 patients who had received the drug hydroxychloroquine for treatment of Covid-19. The study concluded that patients who received the drug were dying at a higher rate than 81,000 controls who did not receive the drug and this led to the WHO suspending its clinical trials of the drug.
The paper was later retracted by its academic authors after they were unable to gain access to the study dataset, supposedly held by a private company, Surgisphere. There are doubts whether this dataset even exists, given it is not consistent with publicly available data and the hospitals from which the data supposedly came say they know nothing about it.
The Lancet recently updated its authorship policy to prevent this situation occurring again, so that it now requires that for papers which are the result of an academic and commercial partnership, all authors will be asked to sign author statements to confirm that they had full access to the data reported in their paper. Here is the relevant paragraph of the Lancet announcement of this change.
“Changes to the signed declarations by authors in the author statements form will require that more than one author has directly accessed and verified the data reported in the manuscript. We will require that the authors who have accessed and verified underlying data are named in the contributors’ statement. For research Articles that are the result of an academic and commercial partnership, one of the authors named as having accessed and verified data must be from the academic team. In addition, all authors will be asked to sign the author statements form to confirm they had full access to the data reported in their Article, and accept responsibility for submitting the Article for publication.” Lancet, Learning from a Retraction, 17 Sep 2020
This issue of access to the study data was exactly the issue which led to the withdrawal of WHO and several external academic members from the Core Team of the IHME Global Burden of Disease Study (GBD) in 2012. The WHO subsequently declined to endorse the results of the first GBD 2010 results, published in the Lancet in December 2012, as it was unable to access the data used.
As the head of the unit responsible for WHO global health statistics, I was closely involved in this issue of WHO engagement with GBD work and the IHME, funded by the Bill and Melinda Gates Foundation to carry out Global Burden of Disease work, and recently published a paper summarizing WHO work on global burden of disease statistics over the last thirty years (see also earlier post here).
Subsequent updates of the IHME GBD have been published in the Lancet, with author lists including many hundreds of collaborators. For example, the most recent GBD cause of death paper has 1,020 authors. On advice from IHME, I signed up as a collaborator in order to obtain regular updates on IHME GBD activity and prior to submission of GBD papers to the Lancet would receive emails requesting me to state whether I wished to be listed as an author. From discussion with various academics who did sign up to be listed as author, I know that the IHME authorship criteria did not require that the collaborator had necessarily made a substantive contribution to the work, or even necessarily made a comment on the draft paper that was acted on, let alone had access to all the study data. There are clear incentives for “collaborators” to sign up as authors, particularly authors without an already strong publication record, as they receive authorship for highly cited papers in one of the highest impact medical journals. In turn, they lend the IHME GBD project an image of wide collaboration, particularly with country-level academics and experts, which may give the impression that the country-specific data has received scrutiny from in-country experts.
The IHME is unusual in not allowing access to data by collaborators. In my time working for WHO, I collaborated with many other academic groups, and never had an academic group refuse access to study data when I requested it as part of a collaborative effort. Will the new authorship policy of the Lancet address this practice? I doubt it. Although the IHME is essentially a privately-funded research group, it is affiliated with the University of Washington and considered an academic institute. It is quite unclear why the policy that all authors should have had access to study data would not apply to large multi-organizational studies in general, rather than just those where the study data is held by a commercial organization.