BACKGROUND Research presented to a scientific meeting is inaccessible to clinicians, unless it is also published in a cited journal.
AIMS To assess the publication rate of studies presented to two UK national paediatric meetings: the Paediatric Research Society (PRS) and the British Paediatric Association (BPA).
METHODS A Medline search in December 1999 for the first authors of all plenary abstracts presented in 1996. If not found, authors contacted by postal questionnaire.
RESULTS Information was obtained on 88/89 presentations. Twenty five of 48 PRS and 31 of 40 BPA studies were published in Medline listed journals. The major reason for non-publication was that they had not been submitted (PRS 15/48, BPA 6/40). Some authors were still hoping to do so (PRS 7, BPA 2). Other reasons were: publication in other forms (theses, book chapters, non-Medline journals) (PRS 5, BPA 2), or still being reviewed (PRS 3, BPA 1). Ten of 11 randomised, controlled trials were published, but only 20 of 37 observational studies were submitted and published.
CONCLUSION Presenters to paediatric meetings need help in submitting and publishing their work.
Statistics from Altmetric.com
Research presented to a scientific meeting is inaccessible to clinicians, unless it is also published in full, in a cited journal. For paediatricians in training, especially those wishing to pursue an academic career, publication is an important step in furthering their career. I aimed to assess the publication rate of studies presented to two national paediatric meetings: the Paediatric Research Society (PRS) and the British Paediatric Association (BPA) plenary sessions. I also tried to determine the reasons why presentations were not published.
I carried out a computerised Medline search in December 1999 for the first authors of all plenary abstracts presented to the PRS and the BPA in 1996. Studies presented to these sessions represent all areas of paediatrics, rather than individual subspecialties. If a presented study was not found, I contacted the first author by postal questionnaire. A reminder was sent out one month later, with a duplicate copy to a PRS/BPA member who was a co-author.
All presenters to the PRS were contacted and asked about their job when the study was performed and if they had submitted other papers for publication.
PAEDIATRIC RESEARCH SOCIETY
Forty nine presentations were given to the PRS spring and autumn meetings in 1996. I found 25 (51%) on Medline or “in press” by December 1999, 21 as papers and four as letters. These were published in 17 different journals, six of which were paediatric journals (table1). Of the studies published as papers, 48% were in paediatric journals. Letters were published in 1996–97, but 13 papers were published in 1998, two years after the work was presented, four in 1999, and at least three more had been submitted but not yet published.
Twenty three presentations were not identifiable on Medline search (table 2); five were published in non-Medline journals or as theses, four were still being reviewed for publication or had been rejected. The data from 15 presentations had not been submitted for publication; seven of these authors claimed that they hoped to submit for publication, but had not yet done so (table 2). Information on one presentation was not available.
Only two presentations described randomised controlled studies, both being published in Archives of Disease in Childhood. Twenty three presentations described observational studies; five of these were not submitted for publication, while 12 (52%) were published in full.
Of the 46 presenters, 41 were paediatricians, of whom 36 (88%) were trainees at the time of the presentation. Twenty eight presentations were by those in research or academic posts. Presenters in academic posts were no more likely than others to submit their work for publication (18/28 versus 14/20) or have it published (15/28 versus 13/20). Thirty eight presenters had submitted other papers for publication. This meant only three presenters had never submitted for publication, two in academic posts.
BRITISH PAEDIATRIC ASSOCIATION
Forty plenary presentations were given to the 68th BPA Annual Scientific meeting in 1996. I found 31 studies (78%) on Medline or “in press” by December 1999. These were published in 19 different journals, five of which were paediatric journals (table 1). Of the studies published as papers, 35% were in paediatric journals. The median date of publication was 1997 (n = 12); two studies had been published before presentation at the meeting.
I could not identify nine on Medline search: one was being submitted, two were still to be submitted, one was published as a book chapter, four had not been submitted, and one was published in a nursing journal (table 2).
Only seven presentations described randomised controlled studies, six of which had been published in Medline listed journals (two inArchives of Disease in Childhood, two inBritish Medical Journal, one inLancet, one inHepatology); one is still to be submitted. Eight of the 14 (54%) presentations describing observational studies were published.
After three years the publication rates of presentations to the two main general paediatric meetings in the UK were: PRS 51% and BPA 78%. The main reason why presentations were not published was that investigators did not submit them for publication, although 43% were still planning to do so three years after presentation. Most randomised, controlled trials were published, but observational studies were submitted and published less often. Studies were published in 32 journals, most commonly in Archives of Disease in Childhood, usually two years after presentation.
I did not include studies presented to the BPA subspecialty group sessions (225 other abstracts). The overall publication rate of the BPA meeting, therefore, is likely to be lower, since the plenary session includes those presentations thought by the academic board to be of the greatest merit or interest.
Previous studies have found similar publication rates; 90% of these papers are published within four years of presentation at a meeting.1
Many reasons have been suggested for non-publication of studies presented to scientific meetings, including rejection after peer review because of unsound methods or negative results2 or small sample size.3 I found only one study rejected for publication and not finally published. Others suggest that studies are not published because they report only “work in progress” or are not completed by the researchers.2
The major factor in the present study was whether presenters prepared and submitted a manuscript. It has been found that if supervisors help prepare manuscripts these are likely to be published.4Failure to submit or publish by those in academic posts suggests variations in the quality of supervision given to young researchers.5
Only 40% of papers were published in paediatric journals, compared with 61% in a previous study.2 “Best paediatric evidence” is found in a number of diverse journals.6Those wishing to access data presented at the PRS and BPA plenary sessions would have to search 32 different journals. Randomised controlled trials were more likely to be published than observational studies.
I conclude that doctors who present to scientific meetings need time and encouragement to prepare and submit manuscripts. Once submitted, publication is likely, unsurprising as the plenary session presentations are selected from the best abstracts submitted to the subspecialty groups. Thus, a plenary presentation should be considered when short listing or assessing paediatricians in training, even though authors are discouraged from using abstracts as references in peer reviewed journals.7
I would like to thank Dr Michael Quinn and the executive committee of the Paediatric Research Society for their encouragement, and Dr Alistair Thomson and the referee for helpful comments. This work was presented in part to the Paediatric Research Society, Stoke-on-Trent, September 1999.