Authorship Decisions in Economic Evaluations

Benjamin M Craig PhD, ISPOR CONNECTIONS Editorial Advisory Board member and Assistant Professor, University of Wisconsin, Department of Family Medicine, Madison, WI, USA

Among the research trades, authorship is a professional currency in addition to the trappings of intellectual self worth. Ambiguities concerning authorship needlessly damage collaborations, impede publication, and deter career development. While researchers are required to meticulously describe their budgets, some fail to equally clarify ownership of scientific spoils, proving detrimental for both the authors and the research community.This health care system model has been considered the one with the greatest financial solidarity on a global level [3]. Nevertheless, on a macroeconomic level, its sustainability depends on a complex set of structural variables, such as employment and economic growth. On a microeconomic level, the challenge is to obtain sufficiency of the premium in a social insurance model that will allow providing a benefit plan with integral guarantees in order to address the population’s health problems, on an epidemiological transition level and in an environment of fast evolution in medical technology.

To the best of my knowledge, no guidance has yet been published in pharmacoeconomics and outcomes research regarding authorship decisions. Naturally, our field requires the integration of clinical expertise and research training, most often fulfilled by interdisciplinary teamwork. Among the articles listed in 2006 Recently Published Works section of ISPOR Connections, 88% are multi-author publications, a proportion which has changed little since 2000. For our field, the mixing of expertise can be troublesome, because conventions on authorship vary greatly by discipline [1].

This paper outlines common guidance on authorship and establishes three rules for streamlining the decision process. Because these rules are based on firsthand experience in pharmacoeconomics and outcomes research, they should be taken with a grain of salt. Dilemmas regularly faced by seasoned and junior investigators are described not to provide a definitive statement on authorship; instead, their purpose is to introduce issues surrounding authorship for reference by junior researchers and for discussion among more seasoned ones.

Three Rules of Authorship
The person with the authority to determine authorship is typically the lead author or the principle investigator of the project. Regardless, these three rules are recommended:

Rule #1 Ask potential authors about their authorship expectations before the first meeting Before a job begins, employees typically have some sense of salary. “Everybody needs money. That's why they call it money!” (Danny DeVito as Mickey Bergman in "Heist”). In research, authorship is a commodity: it speaks to the intellectual caliber of investigators, and researchers who contribute to a project will expect to be paid. As with salaries, authorship negotiations are best handled one-on-one prior to the start of the research activities.

These negotiations entail a summary of the research project, emphasizing the contributions and expectations of each author. Every contributor should receive a list of expected tasks with an initial timeline for completion. Planning far in advance is often difficult, even impractical, but it begins a discussion of reasonable expectations. For example, tasks such as modeling, literature review, and editing can be assigned in trade for a particular rank in authorship based on the effort they require. Effort asked of a senior person is given more weight, because they typically have greater opportunity costs (i.e., higher wage rates), but they are often more productive because of their wealth of experience and training. A verbal description of other author contributions may help calibrate evaluations over rank. Consistent verbal repetition of these contracts helps to set precedent, so that later misunderstandings can be avoided.

How many authors does the paper merit? This number varies greatly by discipline. In economics, articles generally have up to three authors, but the average number of authors is increasing [2]. In the 2006 issues of the Journal of Health Economics, 23% of the articles were sole authored and only 10.5% had over three authors. On the contrary, epidemiology and clinical papers typically have over three authors. In the 2006 issues of New England Journal of Medicine, 98.1% of the original research articles had over three authors, and most of those with fewer authors were written by health economists.

Clinical articles increasingly use group titles (e.g., Clinical Trial Writing Group) instead of listing all authors, which may be in response to Journal requirements and referencing rules. Because the value of authorship may depend on whether the authors' names are listed, discussions of a group name may elicit varying responses. Health technology assessments do not require primary data collection, so the manuscripts typically have less than four authors. However, if an article will have more than six authors, a group name may be substituted for the sixth or more authors. Otherwise, these remaining coauthors may be left out of the reference section under the default convention of “et al.” in future references to the work.

Although clinical-economic trials are increasingly prevalent, health technology assessment are usually based on parameters taken from the literature or secondary data. When a primary researcher provides parameters or data to another, authorship may be expected in trade, even if the primary researcher does not contribute further to the paper. Journals frown on this bartering, but it is a known practice. Bartering over data must be clarified before the project is initiated; otherwise, the primary author may rightfully withdraw their data from the paper, essentially gutting the manuscript.

Some project staff and educators are paid for their efforts, and while they contribute to the project, their efforts may not merit authorship. Such staff includes interviewers, editors, professional writers, data entry personnel, and librarians. Difficulties persist when thesis advisors or supervisors expect authorship regardless of their contributions, because they contributed toward the funding of the project or supervised the career development of the lead author. While they should be rewarded for these contributions, in terms of salary and promotion, their support alone does not necessitate authorship.

Rule #2 Clearly define the authorship order at the first meeting Economic evaluations usually order the authors by amount of contribution, which is more similar to epidemiology than economics. Because the conventions of authorship order vary by discipline, open communication seems to be the best strategy [3]. In traditional economics, authors are arranged in alphabetical order, and arrangement outside of alphabetical order implies secondary authorship [4]. In epidemiology and clinical research, however, authorship order is based on contribution, except for the last author who may be the senior investigator on a multi-paper project [1]. Frank discussions with administrators concerning the “last authorship” position may prevent later confusion, particularly when two senior researchers might assume this privileged post.

Recently, a new pattern is emerging, where senior coauthors who once coveted the last position request second authorship. Journals, in the name of space, have shortened the list of authors in the reference sections. The 2004 Uniform Requirements of Manuscripts do not limit the number of authors on any submitted manuscript; however, when the article is referenced only the first six authors will be listed followed by “et al.” If the manuscript has seven or more authors, the last author may be removed automatically by reference software applications, much to the disappointment and detriment of the senior coauthor.

After introducing the preliminary order and contributions of each author, your collaborators may respond by asking to do more or less, changing authorship toward a preferred order. My experience suggests that such discussions work well as a team building exercise and help legitimize the distribution of tasks. This open formality may seem awkward, but it saves the group time and improves the likelihood of a successful project.

Rule #3 Once written, submit the paper to a journal that matches the authorship At the first meeting, it is prudent to discuss potential journals for the manuscript submission. Journals have particular requirements for authorship best known at the beginning of the project, such as the maximum number of authors, authorship requirements, and rules on conflicts of interest. Thus, early journal identification is good practice as it gives purpose to a research project and motivates authors toward a common goal and audience.

The maximum number of authors varies by journal. For example, the Lancet editorial board set a maximum of eight authors in 1997, which was heavily criticized at the time (Johnstone 1997). Journals such as the Journal of Health Economics do not list a formal limit, but the infrequency of more than three authors might indicate multi-author rejections or, at least, insistent requests for authorship reductions (e.g., dropping research assistants).

For most journals, every author must participate in writing the manuscript and insist that the specific contributions of each contributor be identified. Discussing these requirements at the very beginning of a project may simplify the later removal of authors who fail to contribute in a timely fashion.

Authors are also required to list all potential conflicts of interest at time of submission [6]. For example, the New England Journal of Medicine will not publish an economic evaluation if an author has financial interest in the study results. Authors may privately inform you of their potential conflicts, which may either affect the journal selection or, if caught early, the authorship decision.

Lastly, if the paper's authorship does not fit the journal, the paper should be submitted to a more appropriate publication. It is unfair to both the journal and the paper's authors to change authorship for the purposes of submission to a particular publication. For additional guidance, Harvard Medical School has posted further authorship advice (http://www.hms. harvard.edu/integrity/ authorship.html).

Common Dilemmas
All researchers experience dilemmas in authorship. These cases describe possible responses to three common dilemmas: Removal of a co-author, change in author order, and quid pro quo.

Removal of an author Every seasoned investigator has experienced a collaborative effort where a member of the original team is unable or unwilling to complete their pre-defined responsibilities in a timely fashion. Often the investigator remains interested, but has experienced a change in position or shift in work loads. After missing the predefined deadlines, the first step is to talk with the collaborator, so that you can identify the reason behind their lack of contribution. Options include an extension of the deadline, a change in authorship order, or the removal of the author.

If you have little interest in keeping the author, a cordial dismissal that preserves good feeling and the possibility of future collaboration is the next goal. As a result of this discussion, the author may bow out of the project, in hopes that they may preserve their reputation and collaborate with you sometime in the future. If this offer is not made, you may outwardly empathize with the author and ask to decrease their workload by removing the responsibilities (and authorship) of this project. Most removal cases in my experience have ended in this fashion.

Some cases are not as simple, and involve an author with insufficient motivation to complete the work in a timely fashion, but expects to maintain authorship. As lead author, you must balance the welfare of the unproductive authors with those of the more productive authors, including yourself. One strategy is to strike a deal with the unproductive authors. You might offer authorship on a future paper, if they relinquish the current project. Most authors wish to maintain the appearance of collaboration; therefore offer a schedule of work, forcing the author to choose whether or not to remain an author through delivering the scheduled effort. If the effort is not realized, the lead author has legitimate grounds for removal. Such assertive behavior may render respect of your collaborators.

Regardless, the removal of a disgruntled author is risky, and is best avoided. In most cases, it leaves a black on the reputations of all those involved. Some authors have gone to court to squash manuscripts on the principle of the issue. Sadly, these manuscripts rarely merit the exchange of blows. If the situation worsens to this point, it may be best to drop the manuscript. You can likely significantly change the paper and begin with a new team instead of proceeding under malignant circumstances.

Change in Author Order Less extreme than the removals, changes in authorship order are common. After a project begins, authors may ask for a change in the authorship order, because the assigned tasks required more effort than expected or because they were unable to complete the assigned tasks. This process of renegotiation requires the open participation of all authors, because it may change the ranking. A common tragedy is when collaborators are surprised by a re-ordering at time of publication.

Disagreements over authorship order most often occur because of unexpected work needed to complete the project. For example, a referee may request for a substudy requiring further data collection or an additional analysis. To prevent the arguments, care is needed not to unduly assignment of unpredicted work to the third or four authors. At the top, author contributions may appear more similar, because their tasks include dissimilar activities (e.g., data analysis and site administrator). When the extra work is assign, best practice suggests that the authorship order is clarified.

Quid Pro Quo The demands of academic research have rendered greater rigor as well as a Pandora's box of unethical practices. As a graduate student, a well known professor of labor economics once told me that if I did not want my manuscript refereed by particular colleagues that I should discuss the paper with them and include them in the acknowledgements. In the authorship decision, questionable tactics include 'guest' and 'ghost' authorship [7].

'Guest' authorship is the inclusion of an individual in the by-line who does not meet the authorship criteria. In the end, the journal, editorial board and publisher have little choice but to trust the corresponding author, who may be at the mercy of local influences (`pressured' authorship) or perceive improved publication potential with the inclusion of a noteworthy 'guest' who may not even know about their inclusion. The most challenging cases of 'guest' authorship are when members of a thesis committee or course instructors require authorship of all graduate student submission under their direction, which demoralizes the better students and hurts reputations and student recruitment.

Manuscripts may benefit greatly from the swift pen of a professional writer. 'Ghost' authorship is more treacherous, because the manuscript may be written by another and gifted to the corresponding author, which is obviously incongruent with journal guideline. However, it is also well understood that evidence from particular messengers may be more or less persuasive. These cases also occur in reverse where academic researcher consults on a manuscript under the condition of anonymity, because he or she does not wish attribution for the study results.

Reputation and loyalty are also commodities in research, which may be traded for money, authorship and professional advancement. The authorship decision includes a balance of multiple interests across all those involved. Compared to scientific misconduct (e.g., falsifying results), the trade of authorship for money, reputation, loyalty and professional advance, instead of contribution, is a lesser sin, which may explain why quid pro quo seems increasingly prevalent.

A research project is like starting a new business: choose your coauthors wisely. Look through their CV. If they write many papers with persons under their authority (i.e., junior faculty, graduate students), ask why. The researcher is either exceptionally generous with their time or an unscrupulous leech. Contacting people who have worked with the authorship candidates and checking your authors references before hiring them is simply good practice in academic research.

References
1. Savitz DA. Invited Commentary: What can we infer from author order in epidemiology? Am J Epidemiol 1999;149:5.
2. Sutter M, Kocher M. Patterns of co-authorship among economics departments in the USA. App Economics 2004;36:327- 33.
3. Stokes TD, Hartley JA. Coauthorship, social structure, and Influence within specialties. Soc Stud Sci 1989;19:1.
4. Engers M, Gans JS, Grant S, King SP. First author conditions. J Pol Econ 1999;107:4
5. Laband D, Tollison R. Alphabetized coauthorship. App Econom 2006;38:1649-53.
6. Barnes R and Heaton A. Panel 6: Addressing questions of bias, credibility, and quality in health economic evaluations. Value in Health 1999;2:99-102.
7. Bennett DM, Taylor DM. Unethical practices in authorship of scientific papers. Emergency Med 2003;15:263-70.

  Issues Index | 2008 Issues Index  

Contact ISPOR @ info@ispor.org  |  View Legal Disclaimer
©2013 International Society for Pharmacoeconomics and Outcomes Research. All rights reserved under International and Pan-American Copyright Conventions. 
Website design by Eagle Systems USA, Inc.