Do Corresponding Authors Take Responsibility for Their Work? A Covert Survey

Abstract

Background: Publication of a manuscript does not end an author’s responsibilities. Reasons to contact an author after publication include clarification, access to raw data, and collaboration. However, legitimate questions have been raised regarding whether these responsibilities generally are being met by corresponding authors of biomedical publications. Questions/purposes: This study aims to establish (1) what proportion of corresponding authors accept the responsibility of correspondence; (2) identify characteristics of responders; and (3) assess email address decay with time. We hypothesize that the response rate is unrelated to journal impact factor. Methods: We contacted 450 corresponding authors throughout various fields of biomedical research regarding the availability of additional data from their study, under the pretense of needing these data for a related review article. Authors were randomly selected from 45 journals whose impact factors ranged from 52 to 0; the source articles were published between May 2003 and May 2013. The proportion of corresponding authors who replied, along with author characteristics were recorded, as was the proportion of emails that were returned for inactive addresses; 446 authors were available for final analysis. Results: Fifty-three percent (190/357) of the authors with working email addresses responded to our request. Clinical researchers were more likely to reply than basic/translational scientists (51% [114/225] versus 34% [76/221]; p < 0.001). Impact factor and other author characteristics did not differ. Logistic regression analysis showed that the odds of replying decreased by 15% per year (odds ratio [OR], 0.85; 95% CI, 0.79–0.91; p < 0.001), and showed a positive relationship between clinical research and response (OR, 2.0; 95% CI, 1.3–2.9; p = 0.001). In 2013 all email addresses (45/45) were reachable, but within 10 years, 49% (21/43) had become invalid. Conclusions: Our results suggest that contacting corresponding authors is problematic throughout the field of biomedical research. Defining the responsibilities of corresponding authors by journals more explicitly—particularly after publication of their manuscript—may increase the response rate on data requests. Possible other ways to improve communication after research publication are: (1) listing more than one email address per corresponding author, eg, an institutional and personal address; (2) specifying all authors’ email addresses; (3) when an author leaves an institution, send an automated reply offering alternative ways to get in touch; and (4) linking published manuscripts to research platforms.