Purpose
To evaluate the rates of ghost and honorary authorship in ophthalmology and to determine risk factors associated with ghost and honorary authorship.
Design
Cross-sectional survey.
Methods
Corresponding authors of articles published in Ophthalmology, JAMA Ophthalmology , and the American Journal of Ophthalmology from June 2019 to December 2020 were emailed an electronic survey. The rates of ghost and honorary authorship, demographic characteristics of the corresponding authors with and without ghost and honorary authorship, and risk factors for ghost and honorary authorship were evaluated.
Results
Corresponding authors (n = 830) were emailed a survey and 278 total responses (34.1%) were received; 227 responses (27.9%) were complete and included for analysis. Most respondents (n = 206, 90.7%) believed that the International Committee of Medical Journal Editors (ICMJE) guidelines for authorship adequately address criteria for authorship. Twenty-seven corresponding authors (11.9%) reported characteristics of their articles that indicated the presence of both ghost and honorary authorship (95% CI, 7.7%-16.1%). One hundred fifteen (50.7%) reported honorary authorship (44.2%-57.2%), and 37 (16.3%) indicated ghost authorship (11.5%-21.1%). Being a resident or fellow corresponding author increased the risk of honorary authorship (OR 11.75; 1.91-231.57; P = .03). There were no factors that predicted articles having ghost authors.
Conclusions
While many authors believe the ICMJE guidelines for authorship comprehensively delineate fair authorship practices, listing authors on scientific publications honorarily and excluding authors who qualify for authorship are relatively common practices in ophthalmological research. Further investigation into the drivers of honorary and ghost authorship practices in ophthalmology, and the effectiveness of preventive measures are needed to ensure fair authorship attributions.
INTRODUCTION
A uthorship establishes contribution, accountability, and credit for scientific publications. In 1985, the International Committee of Medical Journal Editors (ICMJE) created their inaugural authorship criteria, which have frequently been updated, with the most recent version released in 2019. The ICMJE recommends 4 criteria upon which to base authorship and these are utilized by most major ophthalmology research publications, including Ophthalmology, JAMA Ophthalmology ( JAMA Ophth.) , and American Journal of Ophthalmology (AJO) . These criteria are specified as: i) substantial contributions to the conception or design of the work, or the acquisition, analysis, or interpretation of data for the work; AND ii) drafting the work or revising it critically for important intellectual content; AND iii) final approval of the version to be published; AND iv) 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.
Honorary authorship refers to an individual who does not meet ICMJE criteria for and yet is included in authorship, while ghost authorship refers to an individual who meets ICMJE criteria and yet is excluded from authorship. , In a recent editorial in AJO , it was noted that honorary authorship can include “guest” authorship of a well-known author; “gift” authorship to a mentor or to foster collaboration; “coercion” authorship, when an author intimidates the research team to include them on the publication; and “collaboration” authorship, where individuals generally of the same department include each other despite not meeting authorship criteria. Ghost authorship can involve industry partners, statisticians, professional medical writers, or trainees.
A seminal paper in JAMA in 1998 reviewed articles published in 6 high-impact journals spanning different medical disciplines and found that of 809 articles that 156 (19%) of articles had honorary authors and 93 (11%) had ghost authors, as reported by corresponding authors who were surveyed. Since then, numerous studies have evaluated the rates of honorary and ghost authorship throughout the scientific literature. , , A comprehensive meta-analysis across all scholarly disciplines reported rates of either honorary or ghost authorship, honorary authorship, and ghost authorship ranging between 10% to 89%, 1% to 63%, and 2% to 75%, respectively. The wide ranges in reported prevalence are likely due to the variation in the definition of honorary and ghost authorship used in the different studies. Given the prevalence of honorary and ghost authorship in the medical literature and limited data on these authorship practices in the field of ophthalmology, this study aimed to evaluate the prevalence of ghost and honorary authorship in ophthalmic journal articles and describe the characteristics of authors who reported these authorship practices for their published work.
MATERIALS AND METHODS
SUBJECT RECRUITMENT AND SURVEY DEVELOPMENT
This study was reviewed by the Johns Hopkins University Institutional Review Board and determined to be exempt. All study activities adhered to the tenets of the Declaration of Helsinki and to regulations outlined in the Health Insurance Portability and Accountability Act (HIPAA). Authorship information from journal articles published between June 2019 and December 2020 from 3 high-impact and comprehensive ophthalmology journals ( Ophthalmology, JAMA Ophth. , and AJO) were collected by 4 authors (GAJ, MJF, JZ, and SEG). The following types of articles were included as they represented either full-length or brief original research, or editorial articles: original articles, reports, and editorials in Ophthalmology ; original investigations, brief reports, clinical challenges, observations, and research letters in JAMA Ophth .; and original articles, editorials, and perspectives in AJO . In 2019 and 2020, these journals required that submitted articles followed ICMJE guidelines to determine authorship.
For each manuscript, the title, corresponding author(s) name(s), contact email address listed in the article, and complete PubMed citation were collected. If a single corresponding author had multiple articles, only the most recent article was included ( Figure 1 ). If the corresponding author had ≥2 articles in the same month, 1 was chosen at random. Corresponding authorship was shared between 2 authors in 3 instances in AJO, while this occurred 8 times each in both JAMA Ophth. and Ophthalmology. For these instances of dual corresponding authorships, if 1 of the authors had another article published in the study journals during the specified time, that article was selected for that author instead of the article for which they were 1 of dual corresponding authors. If not, the original article was used for both authors. One article in Ophthalmology did not provide the corresponding author’s email address, an alternate could not be found, and was thus excluded. Furthermore, there were 11 technical reports with the same corresponding author. In addition, some articles had generic email addresses listed for corresponding authors when a study group was referenced in the author line, such as for the Pediatric Eye Disease Investigator Group (PEDIG) studies. In these instances, when possible, an alternate email was found for the corresponding author; the article and corresponding author were removed if not, which was the case for 12 articles.
An online survey (Supplemental Document 1) was developed using Qualtrics survey software (Qualtrics XM) based on the mail-in questionnaire published in the JAMA in 1998 6 and a subsequent emailed questionnaire published in the British Medical Journal (BMJ) in 2011. The current questionnaire consisted of 25 questions pertaining to the demographics of the corresponding author, the roles and functions of all authors, the writing of the article and participation of professional medical writers, authorship order, the importance of journals publishing different types of authorship information, and corresponding author(s)’ views on the ICMJE criteria for authorship.
The corresponding authors of the selected articles were sent an email in January 2021 inviting them to participate in the online survey. The email contained the complete PubMed citation for their published article. The survey letter noted that the survey response was anonymous, their participation in the survey served as their informed consent to participate, no personal data were linked to the survey, and that no incentive or financial compensation was given to complete the survey. Up to 4 reminder emails were sent at 1-week intervals.
DEFINITION OF GHOST AND HONORARY AUTHORSHIP
For evaluation of honorary ( Table 1A ) and ghost authorship ( Table 1B ), the same definitions were used as previous surveys , that were based on the ICMJE criteria and recognized the specific contribution of assisting in writing the article. , In Table 1A for Survey Question 15, which was identical to a question asked in the BMJ survey in 2011, responses that identified >4 authors as fulfilling only 1 function were not counted as having honorary authorship (n = 3), unless the article met other qualifying criteria. This step was taken to eliminate responses that may have misunderstood prompts regarding co-author function.
| Articles were considered to have HONORARY authors if one of the following three criteria were met: | |
| 1. | The corresponding author did NOT perform ALL the following actions (Survey Question Q12). |
| a. Conceiving and designing the work, OR analyzing and interpreting the data, OR collecting data or other materials | |
| b. Writing the manuscript or part of the manuscript OR revising the manuscript to make important changes in the content | |
| c. Approving the final version of the manuscript | |
| 2. | The corresponding author did NOT feel comfortable explaining the major conclusions of the article for public presentation (Survey Question Q13). |
| 3. | The corresponding author indicated a co-author author performed ONLY ONE of the following functions (Survey Question Q15). |
| Supervising the work of any of the coauthors | |
| Recruiting coauthors | |
| Recruiting study subjects | |
| Analyzing or interpreting data | |
| Conducting literature search | |
| Analyzing or interpreting literature | |
| Reviewing the manuscript | |
| Communicating with journal editor(s) | |
| Signing the statement of copyright transfer to the journal | |
| Conceiving and designing the work | |
| Collecting data and/or other material | |
| Obtaining funding or material support | |
| Performing statistical analysis | |
| Writing the manuscript or part of the manuscript | |
| Approving the manuscript before submission to a journal | |
| Revising the manuscript or making important changes in content | |
| Reviewing page proofs or the journal’s edited version of the manuscript | |
| Articles were considered to have GHOST authors if: | |
| the corresponding author confirmed that any author met AT LEAST ONE of the following criteria. | |
| a. Someone not listed as an author or in the acknowledgements assisted in writing the article ( Survey Question Q17 ) | |
| b. Persons not listed as an author on this article made contributions the corresponding author felt merited authorship ( Survey Question Q21 ) | |
| c. Persons not listed as an author provided other contributions (eg, data collection, analysis, writing or editing assistance, or technical assistance) and were not listed in the acknowledgments ( Survey Question Q23 ) | |
| Articles were considered to have POSSIBLE GHOST authors if: | |
| the corresponding author confirmed that any author met AT LEAST ONE of the following criteria. | |
| a. A non-author was listed in the acknowledgements of assistance in writing the article ( Survey Question Q17 ) | |
| b. Persons not listed as an author provided other contributions (eg, data collection, analysis, writing or editing assistance, or technical assistance) and were listed in the acknowledgments ( Survey Question Q22 ) | |
STATISTICAL ANALYSIS
All data were exported from Qualtrics into Microsoft Excel (Microsoft Corporation). All statistical analysis was completed in Excel and R 3.6.1 (R Core Team). The prevalence rates and adjusted Wald 95% confidence intervals (CI) of honorary authorship, ghost authorship, and both honorary and ghost authorship together were then calculated. Chi-squared (χ 2 ) testing comparing the prevalence of these authorship practices between the current study and a previous study was then undertaken.
Further analysis assessed for risk factors for ghost and honorary authorship. Student’s t tests were used to compare continuous outcomes and χ 2 tests were used to compare categorical outcomes. Key contributory factors were identified based on univariate comparison. Logistic regression analyses were used to study the association between authorship outcomes and covariates of interest: academic degree, academic appointment, past publications, assistance in writing, age, race, and primary profession. A χ 2 test was used to check the collinearity between Q6 (academic titles the corresponding author holds), Q9 (age of the corresponding author), and Q3 (primary profession of corresponding author). Further analysis into the impact of age on ghost authorship analysis and primary profession on honorary authorship was performed due to significance in univariate analysis. Odds ratio (OR) and 95% CI were reported. Results were considered statistically significant at P < .05.
RESULTS
A total of 1185 corresponding authorships from original research articles in Ophthalmology, JAMA Ophth . and AJO met criteria for inclusion ( Figure 1 ). After removal of articles as outlined in Figure 1 , 830 unique corresponding authors were emailed an invitation to complete the survey. Fifteen authors had email addresses that were undeliverable, and an alternate could not be found. In the end, 815 survey invitations were sent and 278 responses (34.1%) were received, of which 227 (27.9%) were complete and included for analysis.
RESPONDENT CHARACTERISTICS
Of the 227 respondents, the majority were male (n = 146; 64.3%) and Caucasian (n = 145; 63.9%) ( Table 2 ). Most were attending physicians (n =158; 69.6%) and had an academic appointment (n = 197; 86.8%). The majority had published >20 articles between 2015-2020 (n = 134; 59.0%) and 206 (90.7%) believed that the ICMJE guidelines for authorship adequately address criteria for authorship, while 16 (7.0%) specifically stated that the ICMJE guidelines do not. Questions on the importance of publishing specific authorship details were asked ( Table 3 ). Publishing institutional affiliation (55.1%), conflicts of interest (77.5%), and funding sources (66.5%) had a higher percentage of “very important” ratings than publishing author contributions (37.4%). The corresponding authors’ methods of determining authorship order were also examined (Supplemental Table 1). In 109 articles (48.0%), the corresponding author alone chose the order, while the order was decided by the study group in 78 articles (34.4%). The most common methods for determining the author order were by the amount each contributed (98; 43.2%) or by the amount each contributed with the last author being the most senior individual (80; 35.2%).
| Respondent Demographics | N = 227 |
|---|---|
| Sex | |
| Male | 146 (64.3%) |
| Female | 77 (33.9%) |
| Prefer not to say | 1 (0.4%) |
| Race/ethnic background | |
| Caucasian/White | 145 (63.9%) |
| Asian | 64 (28.2%) |
| African American/Black | 1 (0.4%) |
| Hispanic/Latin | 4 (1.8%) |
| Other | 7 (3.1%) |
| Prefer not to answer | 4 (1.8%) |
| Age (years) | |
| <30 | 4 (1.8%) |
| 30-39 | 60 (26.4%) |
| 40-49 | 65 (28.6%) |
| 50-59 | 44 (19.4%) |
| 60-69 | 36 (15.9%) |
| 70-79 | 16 (7.0%) |
| Respondent Characteristics | N (%) |
| Respondent professions | |
| Attending physician | 158 (69.6%) |
| Researcher | 32 (14.1%) |
| Resident or Fellow physician | 9 (4.0%) |
| Statistician | 2 (0.9%) |
| Attending physician and researcher | 9 (4.0%) |
| Resident physician and researcher | 3 (1.3%) |
| Nurse | 0 (0.0%) |
| Researcher and statistician | 2 (0.9%) |
| Researcher and other | 2 (0.9%) |
| Attending physician and other | 1 (0.4%) |
| Other | 9 (4.0%) |
| Degrees held by respondents | |
| MD/DO | 103 (45.4%) |
| MD/DO, PhD | 44 (19.4%) |
| MD/DO and other advanced degree (ie, MS/MA, MPH, MBA) | 29 (12.8%) |
| PhD and other degree (ie, MS/MA, MPH, MBA) | 12 (5.3%) |
| MD/DO, PhD and other advanced degree (ie, MS/MA, MPH, MBA) | 8 (3.5%) |
| PhD | 22 (9.7%) |
| Other | 6 (2.6%) |
| Not reported | 3 (1.3%) |
| Academic appointments of respondents | |
| Professor | 80 (35.2%) |
| Associate professor | 47 (20.7%) |
| Assistant professor | 40 (17.6%) |
| Adjunct professor | 7 (3.1%) |
| Instructor | 5 (2.2%) |
| Professor emeritus | 1 (0.4%) |
| Other | 12 (5.3%) |
| Do not have academic appointment | 30 (13.2%) |
| Not reported | 5 (2.2%) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
