Letter: Honorary Authorship in Neurosurgical Literature: A Cross-sectional Analysis

To the Editor: Authorships in peer-reviewed literature reflect credit given for substantial contributions to scientific research. These authorships can help researchers furthering their career, increasing their academic reputation, and acquiring funding. With these same credits, responsibility is expected from all authors concerning the integrity of the work. To provide researchers some tools to determine who should be an author, the International Committee of Medical Journal Editors (ICMJE) published a guideline1 recommending that authorships should be based on 4 criteria: "Substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work"; "Drafting the work or revising it critically for important intellectual content"; "Final approval of the version to be published"; "Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved." All authors should meet all 4 of these criteria in order to qualify for authorship. Authors not fulfilling these criteria and still listed as an author on an article are considered to be given honorary authorship (HA). The ethical issues that arise, such as unfairly giving credit for work, are worrisome. To date, some studies have evaluated the prevalence of HA in different areas of the medical sciences.2-6 However, an evaluation of HA and factors affecting these rates in the neurosurgical literature is yet to be made. Therefore, we examined the prevalence of HA in 2015. An online survey was performed based on previously conducted studies.2,3,5 Email addresses were identified by screening all papers published in 2015 in 4 renowned journals focused on general neurosurgery, namely Neurosurgery (NS), Journal of Neurosurgery (JNS), World Neurosurgery (WNS), and Acta Neurochirurgica (ANC). Studies were included if they had at least 2 authors and if an email address was available of the corresponding author. The survey included questions regarding (1) characteristics of respondents, (2) the awareness of authorship-guidelines, (3) how coauthorships were decided, and (4) who performed certain tasks during the preparation of the manuscript. Furthermore, respondents were asked if they believed that according to the guidelines a coauthor did not make sufficient contributions to be included as coauthor (ie, self-perceived HA). Furthermore, respondents were to select items from a list of tasks, defined as nonauthorship tasks by the ICMJE. If coauthors only performed these tasks and nothing else, the manuscript was considered to have an "ICMJE-defined" honorary author. The full methods section and questionnaire is included in Supplemental Digital Content 1 and 2. Multivariate logistic regression was used to calculate odds ratios (ORs) with their respective 95% confidence interval (95% CI). A P-value < .05 was considered statistically significant. A total of 1612 articles were screened, of which 1181 eligible email addresses were available (Figure 1). Eventually, 1143 surveys were sent and 48 were not deliverable. After the final reminder, a total of 373 surveys were completed, leading to a response rate of 32.6%; 127 responses (34.0%) from authors of JNS, 97 responses (26.0%) from ANC, 82 (22.0%) from WNS, and 67 (18.0%) from NS. Table 1 gives an overview of respondents' characteristics.FIGURE 1.: An overview of the study procedures. NS, Neurosurgery; JNS, Journal of Neurosurgery; WNS, World Neurosurgery; ANC, Acta Neurochirurgica. TABLE 1. - Survey Respondent Demographics (n = 373) Characteristic n (%) Gender 354 Male 313 (88.4%) Female 41 (11.6%) Continent 373 Africa 5 (1.3%) Asia and Oceania 118 (31.6%) Europe 134 (35.9%) North America 106 (28.4%) South America 10 (2.7%) Primary profession 354 Neurosurgeon 272 (76.8%) Other MD 50 (14.1%) PhD/researcher 24 (6.8%) Statistician 1 (0.3%) Other 7 (2.0%) Academic title 347 Professor 110 (31.7%) Associate professor 83 (23.9%) Assistant professor 47 (13.5%) Instructor/lecturer 24 (6.9%) Fellow/resident 44 (12.7%) Other 39 (11.2%) Length of professional experience 356 1-2 yr 43 (12.1%) 3-5 yr 55 (15.4%) 6-10 yr 90 (25.3%) >10 yr 168 (47.2%) Peer-reviewed manuscripts coauthored 373 <5 33 (8.8%) 6-10 47 (12.6%) 11-15 33 (8.8%) 16-20 17 (4.6%) 21-25 16 (4.3%) >26 227 (60.9%) The first author mostly decided on the order of the authors (38.2%), followed by a group-made decision by all authors (32.1%). In 1 case (0.3%) the study funder decided the authorships order (Table 2). Almost 75% of the respondents were aware of the ICMJE-guidelines on authorships and 65.6% were aware of the general issue of HA. Yet, 20.1% of the respondents currently work at a department where a senior member is automatically listed as coauthor on all submitted manuscripts. However, more than half of the respondents felt this was either never (33.6%) or rarely (26.4%) justified. TABLE 2. - How Authorship is Determined and Awareness of Authorship Guidelines Characteristic N (%) Primary role in actual writing of the article 354 Wrote all or most of the article 245 (69.2%) Wrote minor parts of the article 9 (2.5%) Only revised and made corrections 23 (6.5%) Supervised writing of others 42 (11.9%) Performed majority of data collection/analysis 16 (4.5%) Other 19 (5.4%) Position on 2015 paper in NS/JNS/WNS/ANC 355 First author and corresponding author 208 (58.6%) First author but not corresponding author 5 (1.4%) Corresponding author but not first author 58 (16.3%) Senior author and corresponding author 79 (22.3%) Senior author but not corresponding author 5 (1.4%) Decided the order of authorship 355 First author 136 (38.3%) Senior author 89 (5.4%) Authors decided as a group 114 (32.1%) Funding source of this study 1 (0.3%) Other 15 (25.1%) Criteria used to determine authorship: authors listed… 355 In the order of the amount each contributed 141 (39.7%) In the order of the amount each contributed, except the last author, who is the most senior in the group but didn't contribute to the study 29 (8.2%) In the order of the amount each contributed, except the last author, who provided the concept, supervision, and responsibility for all working steps of the project 170 (47.9%) In alphabetical order 1 (0.3%) Other 14 (3.9%) Aware of ICMJE-guidelines 355 Yes 263 (74.1%) No 92 (25.9%) If not, aware of other guidelines 132 Department/institution guidelines 87 (65.9%) No guidelines are followed 38 (28.8%) Other 7 (5.3%) Aware of the general issue of honorary authorships 349 Yes 229 (65.6%) No 120 (34.4%) Senior member of the department, who is automatically listed as an author on all submitted manuscripts 354 Yes 71 (20.1%) No 260 (73.4%) Don't know 23 (6.5%) If so, do you feel this is justified? 250 Never justified 84 (33.6%) Rarely justified 66 (26.4%) Sometimes justified 64 (25.6%) Most of the time justified 22 (8.8%) Always justified 14 (5.6%) Did anyone suggest to include a honorary author 354 Yes 30 (8.5%) No 324 (91.5%) NS, Neurosurgery; JNS, Journal of Neurosurgery; WNS, World Neurosurgery; ANC, Acta Neurochirurgica; ICMJE, International Committee of Medical Journal Editors. Nonauthorship tasks performed most frequently by coauthors were collecting data and performing cases of the study (145 studies), reviewing the manuscript (133 studies), and approving the manuscript before submission (112 studies). Other less frequently performed tasks were: recruiting study subjects (75 studies), supervising coauthors (74 studies), contributing illustrations (52 studies), and obtaining funding or material support (37 studies). Of all the studies, 121 did not have any coauthors performing nonauthorship tasks, leading to a prevalence of 62.9% of ICMJE-defined HA overall. When asked if the respondent believed that any of the coauthors did not make sufficient contributions to coauthor the manuscript, 24.2% answered yes (see Figure 2).FIGURE 2.: The prevalence of "self-perceived" and "ICMJE defined" honorary authorships for the surveyed journals and totally. NS, Neurosurgery; JNS, Journal of Neurosurgery; WNS, World Neurosurgery; ANC, Acta Neurochirurgica.Table 3 shows the results of the multivariate logistic regression models. Respondents working in a department where a senior member was not automatically enlisted as an author on all submitted manuscripts, were significantly more likely to have a self-perceived HA (OR: 3.68, 95% CI: 1.80-7.54) but significantly less likely to have an ICMJE-defined HA (OR: 0.18, 95% CI: 0.10-0.34). TABLE 3. - Multivariate Logistic Regression on Parameters Associated With ICMJE-Established or Perceived Honorary Authorship Variable Perceived honorary authorship ICMJE-defined honorary authorship Odds ratio 95% CI of OR Odds ratio 95% CI of OR Journal (ref. ANC) World Neurosurgery 1.04 (0.48-2.30) 1.33 (0.64-2.75) Journal of Neurosurgery 1.01 (0.49-2.08) 1.75 (0.90-3.39) Neurosurgery 0.82 (0.34-2.03) 1.82 (0.85-3.92) Number of peer-reviewed studies authored (ref. <5 published studies) 6-10 0.40 (0.12-1.35) 1.22 (0.36-4.18) 11-15 0.22* (0.05-0.96) 3.00 (0.87-10.42) 16-20 0.78 (0.18-3.48) 1.65 (0.37-7.38) 21-25 0.74 (0.16-3.32) 1.68 (0.38-7.38) >26 0.53 (0.20-1.43) 2.33 (0.84-6.48) Length of professional career (ref. 1-2 yr) 3-5 yr 0.78 (0.26-2.33) 1.74 (0.64-4.68) 6-10 yr 1.71 (0.68-4.29) 1.05 (0.42-2.59) >10 yr 1.09 (0.45-2.67) 1.29 (0.55-3.02) Primary profession (ref. medical personnel) Academic 1.46 (0.52-4.10) 0.66 (0.25-1.76) Gender (ref. male) Female 0.85 (0.33-2.17) 0.60 (0.25-1.41) Senior member with automatic coauthorship (ref. yes) No 0.18* (0.10-0.34) 3.68* (1.80-7.54) Don't know 0.67 (0.23-1.98) 4.36* (1.35-14.10) n 332 332 Pseudo R-square 0.13 0.09 −2 Loglikelihood (df = 2) −163.09 −199.94 *P < .05. This study presents the results of an online survey administered to corresponding authors of 4 neurosurgical journals. To the best of our knowledge, this is the first study to attempt to address the issue of HA in the neurosurgical literature. Our aim was to report the prevalence of HA of 2015. A total of 24.2% of the corresponding authors felt that according to the ICMJE-guidelines, their coauthors did not make sufficient contributions to an article. When taking authors in account who only performed nonauthorship tasks as defined by the ICMJE, a higher prevalence (62.9%) was found. Awareness of the ICMJE-guidelines, however, was high with only 25.9% of the respondents stating not to know these guidelines. Having a senior member at the department automatically enlisted as coauthor on all manuscripts was the only significantly associated factor with honorary authorships, however with opposite effect sizes. Departments with no automatic coauthorships for the senior member were more likely to have ICMJE-defined HA, but at the same time less likely to have self-perceived HA. This survey shows that despite a high awareness of the ICMJE-guidelines on determining authorships, the prevalence of HA in the neurosurgical literature is high. Further research should focus on why authorship guidelines are not applied in some cases and, thus, how to decrease the amount of HA. Disclosure The authors would like to thank Ms. Amrita Tejiram for critically evaluating the manuscript. The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.