Five senior editors at the Croatian Medical Journal published a tidy randomized study of two methods for assessing authorship contribution in the Journal of General Internal Medicine. The study involved 865 authors of 181 manuscripts submitted to the journal (CMJ, that is) between January and July 2005 who were randomized to provide either binary (yes/no) or ordinal (0-4 scale) rating of their contributions. Each author on each paper had to complete and sign the form; the first or corresponding author could not serve as a proxy for the others. The qualitative values for the ordinal rating were “none” (0), “small” (1), “moderate” (2), “large” (3), and “full” (4).
One set of questions addressed contributions used by ICMJE to define authorship eligibility:
Conception and design of the study
Analysis and interpretation of data
Collection and assembly of data
Drafting of the article
Critical revision of the article for important intellectual content
Final approval of the article
A second grouping of responses do not reflect ICMJE authorship requirements:
Provision of study materials or patients
Statistical expertise (not analysis of data)
Administrative, technical, or logistic support
Guarantor of the study
One hopes the first 4 would not be used to claim authorship … though I am confused their inclusion of the last bullet in this grouping versus on its own. ICMJE guidelines acknowledge the potential role of a study guarantor, though they stop short of requiring it. However, it is not in the same non-authorship class as providing secretarial assistance or referring patients.
Specifically, ICMJE observes:
Some journals now also request that one or more authors, referred to as “guarantors,” be identified as the persons who take responsibility for the integrity of the work as a whole, from inception to published article, and publish that information.
Given the Hellinga case, perhaps ICMJE will revisit the guarantor issue in their next update.
Getting back to the CMJ editors’ study, they conclude, not surprisingly, that “ordinal scales for reporting authors’ contributions to manuscripts were more sensitive than tick boxes for assessing the appropriateness of authorship.” They analyzed the data with less strict (0 vs 1-4) and strict (0-1 vs 2-4) ordinal rating criteria compared with the binary rating. A shame they didn’t also run the numbers using even more strict criteria (i.e., 0-2 vs 3-4).
The authors go on to suggest that “final approval of the article” be dropped as an ICMJE authorship requirement and instead be viewed in the same light as the conflict of interest statement or copyright transfer. Part of their rationale is that prior research (their work and that of others) indicates this criterion is the most poorly understood by authors. Hmm. Seems like a teachable moment to me.