To Slice or Perish

Redundant publication is a common but detrimental practice that has pervasive implications for science and society at large. The most well-known type of such malpractice is a duplicate publication, which refers to the inclusion of identical or very similar data in multiple papers.^1,2 A less scrutinized form of redundant publication is salami publication, also known as salami slicing and salami publishing, which refers to the unjustified segmentation of a coherent dataset into its smallest publishable units (i.e., the minimum amount of data that warrants publication).^1,3 The sliced publications tend to have similar authors, research questions, hypotheses, methodologies, and participant cohorts, but may differ by specific outcomes or subgroups.^3,4 For example; a two-arm interventional study may be inappropriately published as two articles, one reporting on efficacy and the other on safety.

While duplicate publication is considered a universal taboo, ethical perspectives on salami-slicing are incongruous.^5,6 This heterogeneity may be partially due to the lack of a definitive threshold that separates appropriate from inappropriate fragmentation of data.^1,2,7 In general, data should not be divided if, in comparison to multiple papers, a single article of reasonable length would be more cohesive and informative.⁸ Instances in which it is appropriate to segment data include 1) large projects such as clinical trials and epidemiological studies in which numerous research questions are investigated, and a copious amount of data is generated, and 2) multifaceted projects in which it is beneficial to convey different messages or perspectives to separate target audiences.^4,9 Salami slicing occurs when the intent to advance scientific knowledge is superseded by motive of external, secondary gain. Among the driving forces underlying redundant publication, the competitive ‘publish or perish’ nature of academia is unanimously considered the strongest impetus.^{3,4,7,10–16} Limitations on human and material resources as well as academic honors and positions perpetuate fierce competition between scientists. Exacerbating this issue is the overemphasis on publication quantity and the consequent depreciation of quality and impact. In response to lofty selection criteria from funding bodies and academic institutions, some researchers are pressured to derail from the noble quest for truth and pursue ethically questionable ventures to maximize publication output and citation count. Driving forces for redundant publication other than career advancement include institutional pressure, monetary profit, desire for peer recognition, and pharmaceutical marketing.^10,17

The reported prevalence of salami publication among medical disciplines ranges from 0.4% to 11.6%.^18–22 These prevalence rates are likely underestimated because numerous journals with lower visibility or impact factors were not investigated, and it is challenging to assess for salami publication without prior knowledge of the relevant literature. A recent content analysis of 209 journals in epidemiology, public health, and general and internal medicine identified a paucity of journal policies on salami-slicing. Whereas 49% of journals had explicit policies on duplicate publication, only 13% had policies on salami publication.³ Furthermore, in contrast to the uniformity of policies on duplicate publication across journals, those on salami slicing were inconsistent and lacked well-specified definitions and consequences for misconduct.³ In ophthalmology, salami-slicing has not yet been systemically studied, but it is not unreasonable to suspect similar deficiencies. One review found an article redundancy rate of 1.39% among 70 ophthalmology journals between 1997 and 2000, but only complete duplicates were analyzed.¹⁶

Salami publication has received harsh criticism for myriad reasons. Inappropriate data fragmentation encourages other forms of malpractice to accommodate multiple papers, including omission of key information (e.g., details of original dataset or similar studies), inappropriate extrapolation of results, data dredging, and data falsification.^7,23 Salami slicing also decreases motivation to pursue large-scale, methodologically rigorous studies that confirm and expand on preliminary findings. Consequently, scientific progress is stalled, and the literature is polluted with the smallest publishable units of scant significance. Although inflation of publication output seems beneficial in the short term, a high publication count is meaningless if the overall value of accumulated work is deficient. Furthermore, research visibility and citation count may be decreased rather than enhanced by salami slicing as fragmented data tend to be published in lower impact factor venues and journals in non-related disciplines.¹⁴ Discovery of salami publication may damage the author’s reputation and negatively impact their future career, especially if duplicitous attempts were made to conceal this malpractice.

In addition to harming authors, inappropriate data fragmentation has numerous detrimental effects on science and society. Salami slicing perpetuates academic cultures and reward systems that prioritize quantity over quality of research output, deepening inequities in funding and career opportunities.^24,25 Salami slicing also squanders precious resources in academic publishing, notably the time and energy of editors, reviewers, and funding bodies, and diverts attention away from higher quality research. Unlike the automation of detecting duplicated text, assessing for salami publication is a manual, time-consuming process that requires in-depth expertise of the specific topic.¹⁵ On the receiving end of research dissemination, salami slicing prevents readers from holistically appraising the original dataset, obscuring valuable insights on the inter-relation of data and diminishing the significance of the work.⁴ Even worse, data fragmentation may cause misinterpretation or distortion of evidence. An example of the latter is when two publications involving the same participant group are believed to represent two distinct samples. As a result, disproportionate attention and importance may be given to a set of findings, and data may be excessively weighed in meta-analyses, skewing effect estimates.^9,26 Distorted evidence builds flawed guidelines for clinical practice and leads to patient care that is at best suboptimal and at worst dangerous.

A multitargeted approach addressing different elements of the academic system must be adopted to curtain and prevent redundant publication. Authors should prioritize quality, comprehensiveness, and impact when developing a publication strategy.^3,27 Aiming for meaningful scientific contribution often translates to minimizing the number of publications generated from the same dataset. In the event that data segmentation is deemed the best course of action, article overlap should be minimized, and several key pieces of information should be provided to ensure academic integrity. Detailed references should be made to similar articles, including the original dataset and all sibling publications derived from it. Strong reasoning should be provided in the cover letter to editors and in the manuscript to justify the fragmentation of data into multiple outlets. Relatedly, authors should include supporting evidence for how their work constitutes a novel, substantial, and stand-alone contribution to the field.^3,13,27

For journal editors, a standardized definition for salami publication should be established to inform the development of relevant policies and punitive sanctions. Policies should be developed with the expertise of researchers and peer reviewers and incorporate strategies for detecting salami publications. Examples of such strategies include random data audits and chronological implementation of checkpoints in the publication process.^3,10 Editors may require authors to provide a comprehensive list of articles that may have relevance to the manuscript as part of the submission package. Since this process relies on the integrity of authors, peer reviewers and editors should routinely assess for salami publication and investigate suspect cases.^3,27 Publication policies should be directly and conspicuously stated on the journal’s website to facilitate ease of access; isolated references to generic guidelines and checklists should be avoided.³ Consequences for non-compliance should be enforced based on severity and may range from cautioning authors and rejecting manuscripts to block listing authors and informing relevant governance and funding bodies.^1,3 The severity of redundant publication may be determined using criteria such as level of transparency, distinctness of research questions, originality of data, and degree of contextualization within related literature.¹

The final, and perhaps most important, defense against salami publication is a revolution of the current climate of academia. Changes must occur at the individual, institutional, and government levels to reduce unhealthy pressure to publish and reward ethical research practices. Evaluation systems that determine funding allocation and career progression should be revamped to prioritize quality over quantity of scholarly output.³ Consensus should be reached regarding how publication merit should be defined and assessed. Potential methods include assigning more weight to the scientific or real-world impact of individual publications, which may be aided by citation metrics such as the h-index, and limiting the number of articles that authors may submit in application processes.^3,10,28 Systemic initiatives must be complemented by direct mentorship from senior researchers to instill academic integrity in trainees and incorporating pointers to redundant publications in manuscript checklists and philosophies of manuscript rejection.^29,30 Young researchers may then propagate ethical values and etiquette to future generations.

Continuously harvesting low-hanging fruit is complacent and harmful, and the repercussions are especially pernicious in academia. Although the goal of scientific inquiry may intermittently conflict with professional and personal interests, the question of, ‘Is this publishable?’ should never supplant, ‘Is this good science?’ All members of the scientific community share the responsibility of preventing this decadence.

Disclosure statement

None