The fate of rejected manuscripts in different biomedical disciplines

Abstract

A scientific journal may reject a submitted manuscript for several reasons. For example, desk rejection occurs when a topic is addressed that does not fit a journal's aims. Or the submitted manuscript has insufficient methodological quality. Therefore, authors need to manage manuscript rejection as a normal part of the publication process. Because methodological quality is not the only reason for rejection, authors will typically submit a rejected manuscript to another journal. This allows research authors to maintain efficiency and avoid wasted effort.¹ Actually, not attempting to publish research material due to one journal's rejection can be considered a form of publication bias.²

The fate of rejected manuscripts has already been researched in different biomedical disciplines. However, the quality and characteristics of the studies assessing rejected manuscripts have not been performed. The aim of the present survey was to examine the characteristics of studies assessing the fate of rejected manuscript submissions in various biomedical disciplines and to provide manuscript rejection rates.

Studies assessing the fate of rejected manuscripts in different biomedical disciplines were included. Similar studies focused on other scientific disciplines were excluded. Only articles written in English were included. The PubMed, Scopus, and Web of Science Core Collection databases were searched on 12 May 2024, and articles published from the database's inception to the date of search were considered for inclusion. The following keywords and Boolean operators were applied in a search of the database: “rejected manuscript” OR “rejected manuscripts” OR “rejected articles” OR “rejected papers.” Additionally, the reference lists of the articles included were scrutinized for further potential studies to include (Section S1). We selected articles based on the eligibility criteria. Those not meeting these criteria were excluded, first in a title/abstract assessment and then in a full-text assessment (Sections S2–S4, Figure S1).

The reasons for exclusion were individually determined. When data on rejection rates were not reported, we calculated them by dividing the number of manuscripts submitted by the number of manuscripts rejected. The data were extracted into an Excel datasheet, and the included information is reported in Table 1. Study selection and data extraction were performed in duplicate for 10 studies. After good agreement (more than 80%), the remaining process was performed by one author (M.C.M.).³ Data were presented as frequencies and percentages and, when applicable, as means and medians.

A total of 36 studies meeting the eligibility criteria were found (Section S5). The most prevalent affiliation of the first author of the studies was the USA (n = 17, 47%), followed by Italy and Switzerland (each n = 3, 9%). The discipline of radiology was the most researched (n = 7, 19%). Authors self reported nine different study designs; however, 23 (n = 64%) studies had design reported. More than half of the studies (n = 19, 53%) did not contain any conflict of interest (CoI) statement. Reporting on sponsorship followed the same pattern; about 60% of the studies (n = 22, 61%) did not have any statement on sponsorship. Of the 12 studies that reported some information on approval of the study protocol by an official board, 5 (14%) reported no need for approval, and 7 (19%) reported approval. The median number of citations of the studies was 19.5 (interquartile range (IQR) = 29); the median h-index of the first and last author of the studies was 27 (IQR = 32) and 46 (IQR = 40.75), respectively. The 36 included studies were published between 1991 and 2022 (median = 20105, IQR = 8.5).

Information on rejection rates was reported in 24 (67%) studies. The median rejection rate was 65.8% (IQR = 28.7%). With data available from 22 studies, the median impact factor (IF) of the journals that rejected the submissions was 2.628 (IQR = 1.922). The median and mean IFs of the journals that accepted the submissions rejected by the first journal were 1.57 and 1.83, respectively (data available from nine studies each).

With data available from two studies, the median time in months for publication after rejection was 15.5 (IQR = 2). The median rate of articles published after rejection was 56% (IQR = 26%). A detailed report of rejection in general, and by biomedical disciplines is presented in Tables S1,S2, respectively.

Most of the studies retrieved information on the fate of manuscripts from only one source, and the search strategy was reported in 75% of the studies. For most studies, it was unclear whether data selection and extraction were conducted in duplicate. There was no report of a pilot/training phase of researchers in any study in the sample. Of the included studies, six^4-9 reported data on the reasons for rejection. The most frequently reported reasons were poor design, a lack of originality and novelty, and the article not fitting the scope of the journal.

Our survey identified characteristics of studies on the fate of rejected manuscripts and some methodological limitations of these studies. We also found that within our sample, the greatest rejection rate was in the field of general and internal medicine (journals assessed: The Lancet, The British Medical Journal, Annals of Internal Medicine), and the lowest rejection rate was in the field of clinical neurology. It is important to emphasize that we used specialty journals as a proxy for medical disciplines. The median IF of the second journal accepting a rejected manuscript was lower than the median IF of the journal that initially rejected the manuscript.

In the field of anesthesiology (second highest rejection rate in our data set), the data were retrieved from only one study that assessed rejection rates from manuscripts submitted to a single journal. This journal (Anaesthesia) had an IF of 10.7 (JCR 2022) and had the highest IF in its category. One can hypothesize that this journal likely receives many submissions due to its relatively high IF, and, therefore, the journal can be very selective in what it publishes. Evidence suggests that IF, together with a journal being indexed and peer-reviewed, are the most important criteria for an author in the medical field in selecting a journal for submission.¹⁰ A larger sample would be needed to fully examine the relationship of journal characteristics and reasons for rejection/acceptance in the case of rejected manuscripts that are accepted by another journal.

Studies assessing rejections could profit from a methodological quality perspective by following guidance on items reported in tools such as the AMSTAR-2 (assessment of multiple system reviews).³ All the items on this checklist would not be applicable, but those related to search/selection/extraction and ethical issues could be easily applied. In fact, studies assessing studies on rejection of manuscripts should also observe a systematic approach to identifying, selecting, and extracting information as any systematic review does. This approach would give the reader the confidence that all potential studies were included, reducing chances of potential publication bias, and that the studies would be developed in the highest methodological quality possible. These studies could also observe a more homogeneous definition regarding their design to standardize them. This standardization would facilitate the assessment of these studies in a systematic review frame as well as the right identification of the type of study by the reader. Another measure that could improve the methodological characteristics of these studies would be the implementation of a strategy of piloting the different phases of the research project. For example, in data selection and extraction, this strategy would potentially minimize the risk of errors in the project development.

Studies assessing studies on rejection of manuscripts could also profit by reporting with more detailed information on the ethics perspective. For example, by clearly stating potential financial and non-financial CoI^{11, 12} and details on whether the study received and economic support to be developed, another domain that could be improved in future studies about this topic is the approval of the research by any ethics board. Two thirds of the studies in this sample did not present any information about ethical approval. In fact, this kind of approval is likely more important when there is the involvement of treated patients. However, interestingly, a few studies in this sample reported some kind of approval, and the scientific community should discuss further the need for such approval for developing this kind of study.

Further research is needed in other biomedical disciplines to compare with the characteristics and rejection rates reported here. This research should involve more journals in each discipline to increase the generalizability of the present findings.¹³

In resume, this study reports that some biomedical disciplines have different rejection rates and that rejected manuscripts, when resubmitted, are published in journals with lower IF than those that rejected the manuscript in the first submission. Furthermore, there is room for improving the reporting of the methodologies of studies assessing the fate of rejected manuscripts in biomedical disciplines.

The authors declare no conflicts of interest.