1.1

Introduction

Scientific congresses provide medical and scientific researchers with an opportunity to present novel work in their particular domains. The oral presentations and posters presented at congresses are selected on basis of the abstract alone, which introduces bias .

Publication in a peer-reviewed international journal is the best way to ensure widespread dissemination of research work and the latter’s incorporation into the scientific debate. The fact that a journal’s selection process is more stringent (i.e. due to peer review) means that only some of the work presented at a congress is subsequently published in a journal. The abstract publication rate is considered to be an indicator of the congress’s scientific quality . The abstract publication rate varies greatly (from 20% to over 60%, depending on the congress and the field of research or medical specialty in question ). In congresses organized by physical and rehabilitation medicine (PRM) societies, the reported publication rates range from 25% to 34% . However, the abstract publication rate per se only partly reflects scientific quality. In fact, some researchers do not subsequently submit their work to a journal (notably when the findings are negative). Furthermore, there are sometimes significant differences between the data presented at a congress and those subsequently published in a journal.

The French Society of Physical and Rehabilitation Medicine ( Société Française de Médecine Physique et de Réadaptation [SOFMER]) organises an annual congress that attracts researchers from France, the rest of Europe and Africa. Although a great amount of work is presented at the SOFMER congress, we are not aware of any data on subsequent publications.

The primary objective of the present study was to determine the publication rate for the work presented at the 2008 SOFMER congress and to study the factors associated with successful publication of this work. The secondary objectives were to study the differences between the data in the abstracts presented at the congress and those in the publication and to establish the reasons for non-publication.

1.2

Materials and methods

1.3

Results

1.3.2

Publication rate and characteristics of the work published

1.3.2.1

Descriptive analysis

Forty-nine communications had been published at the time of our survey, i.e. an overall publication rate of 21.2%. The mean ± S.D. time to publication was 18.4 ± 14 months and the median time was 21 months. Five studies had been published before the congress and those published after the congress were evenly spread out over time ( Fig. 1 ). The articles were published in 24 different journals, including 19 (38.8%) in the Annals of Physical and Rehabilitation Medicine . The mean IF was 1.59 ± 1.76 (range: 0–6.51; median: 1.53) but 20 articles (40.8%) were published in journals with no IF ( Fig. 2 ), including those published in the Annals of Physical and Rehabilitation Medicine . The PRM journals represented the main vector for publication (29 articles), followed by journals in the fields of neurology, rheumatology & orthopaedics, internal medicine & geriatrics and, lastly, urology ( Table 3 ). The PRM journals had a mean IF of 0.71, which was significantly lower than the other specialities (other than internal medicine & geriatrics journals). Only three articles (6.1%) were published in a language other than English.

Time to publication for the work published (as a Kaplan Meier curve). Five studies were published before the congress.

Impact factors for the journals in which work presented at the SOFMER 2008 congress was published ( n = 49).

Table 3

Distribution of the journal articles by speciality, together with the corresponding impact factor (IF).

Specialty	n (%)	Mean IF ± S.D.
PRM	29 (59.2)	0.71 ± 1.1
Neurology	7 (14.3)	3.9 ± 1.38 *
Rheumatology & orthopaedics	7 (14.3)	1.64 ± 0.91 *
Urology	3 (6.1)	4.51 ± 2.7 *
Geriatrics & internal medicine	3 (6.1)	1.67 ± 0.5

 

Statistically significant differences between the mean IF of the physical and rehabilitation medicine (PRM) journals and those in the other specialities are indicated as * p < 0.05 or ** p <0.001; S.D.: standard deviation.

Relative to the congress abstracts, 14 articles (28.6%) presented minor differences and 10 (20.4%) presented major differences.

1.3.2.2

Factors associated with successful publication

The results of our univariate analyses are presented in Table 4 . The following factors were found to be significantly correlated with publication: (i) the fact that the work was submitted by a university group, (ii) authors from France or another European country and (iii) original research. The presentation mode (poster vs. oral), the theme, the number of participants and (for therapeutic studies) the study’s design and outcome were not significant factors.

Table 4

Factors associated with publication of the work presented as an abstract at SOFMER 2008.

	Number of publications	p
Type of presentation
Oral communication	35 (24.3 %)	0.093
Poster	14 (16.9%)
Field of work
Neurology	21 (23.9%)	0.133
Movement disorders	22 (26.2%)
Other	6 (10.2%)
University affiliation
University group	45 (25%)	0.005
Non-university group	4 (7.8%)
Country/geographic zone
France	43 (22.6%)	0.044
North Africa	1 (3.8%)
Other	5 (33.3 %)
Type of work
Original research	38 (29 %)	0.004
Literature review	1 (16.6 %)
Case report	7 (17.9 %)
Presentation of practice	3 (5.5 %)
Characteristics of original research
Single-centre	33 (25.2 %)	0.143
Multicentre	4 (50 %)
Therapeutic study	17 (30.9 %)	0.414
Non-therapeutic study	21 (27.6 %)
Randomized controlled trial	3 (75 %)	0.083
Other design	14 (27.5 %)
Positive outcome	14 (30.4 %)	0.299
Negative outcome	3 (50 %)

 

In a logistic regression, none of the studied factors were significantly associated with publication of the work described in the abstract.

1.4

Discussion

The objective of the present work was to assess (i) the publication rate for abstracts presented at the 2008 SOFMER congress, (ii) factors associated with successful publication and (iii) reasons for non-publication.

The overall publication rate over the three years following the congress was 21.2%. This rate is slightly lower to those published for other PRM congresses (which range from 25 to 34% ) but it is similar to that of the 2008 European Congress on PRM held in Bruges (21.3%, according to a study that we performed in parallel and whose results have yet to be published). Outside the field of PRM, the publication rate was generally lower than that of congresses in other medical specialities (such as orthopaedics , urology and cardiology ). Moreover, the median time to publication was higher than those reported in the literature . Articles were published evenly over the years following SOFMER 2008, whereas the shape of the publication curve is usually logarithmic .

The mean IFs for the journals in which the work had been published were similar to the values reported by Smith et al. following the American Physical Therapy Association’s congresses but were lower than for other specialities . The PRM journals’ mean IF was lower than in the other specialties; however, this reflects disparities between specialties because the highest IF for PRM journals was 4.5 (versus over 20 for neurology, over 10 for internal medicine and 8 for urology and the locomotor system). The IF is considered to indicate a journal’s quality; it depends on the number of citations of the journal’s articles and the number of articles published by the journal. This explains why articles in a less prominent specialty like PRM are published in journals with a lower IF–especially since the SOFMER’s house journal ( Annals of Physical and Rehabilitation Medicine ) lacks an IF and is the main conduit for publication of work presented at the SOFMER’s annual congress. The highest IFs were those for journals in other specialities (notably neurology).

The observation of minor differences between the abstracts and the corresponding publications is in line with the literature . This is notably due to the presentation of preliminary results at the congress. These minor differences do not appear to be a problem, although major differences are of greater concern.

The affiliation and origin of the research group were linked to publication of the work. University groups were more likely to have published their work; this is unsurprising, since publication is a key part of a university researcher’s activities. Groups from France contributed the majority of the abstracts (as is usually the case for congress of national learned societies ) and had a much higher publication rate than non-European groups; this may be due to a higher proportion of oral communications, better-quality work and easier access to the publication process. Publications aside, it is important that the SOFMER congress continues to act as a showcase for French-speaking researchers from outside France . As with other congresses , original research was more likely to be published than other types of work. In fact, many of the SOFMER abstracts concerned case reports and presentations of practice; this may have lowered the overall publication rate. The associations between the nature of the abstract and other analyzed factors were not statistically significant, although there were trends (as often observed in the literature) towards the predominant publication of oral communications , multicentre studies and randomized controlled trials . A positive or negative outcome did not appear to be a discriminant factor in our study–a finding which contrasts with the publication bias described elsewhere . However, the number of therapeutic studies with negative outcomes was low; this may have translated into a lack of statistical power and might suggest that there was upstream submission bias.

Lack of publication of work presented at a congress can be due to two reasons: the work is either not submitted or does not clear the barrier of peer review. Most of the work communicated at the 2008 SOFMER congress had not been submitted to a journal, and about half of these authors stated that were not intending to submit the work for publication in the future. As is well established in the literature, lack of time appears to be a major factor. In contrast, lack of interest in the subject is only rarely mentioned . The significant number of preliminary results presented at the congress also constitutes an obstacle to rapid publication. The work might not have been continued or may not have been ready for publication within the time interval studied here.

Although peer review does not prevent the publication of poor-quality work , it is still the gold standard for assessing the quality of scientific work. Importantly, the reviewers’ judgement is based on definitive datasets. In this respect, the partial data communicated at a congress must considered with a degree of caution; this is why Hoag et al. suggested that a “congress IF” should be used to judge the scientific quality of a congress on the basis of the publication rate over the following 4 years . With respect to our present study, improvement of the publication rate for work presented at the SOFMER congress will necessarily involve an increase in the proportion of original research and a review process that is more in line with the principles of evidence-based medicine – even though the specific features of PRM often make it difficult to perform studies with high levels of evidence . Furthermore, acquisition of an IF by the SOFMER’s house journal ( Annals of Physical and Rehabilitation Medicine ) would appear to be important for better exploitation of the society’s congress. Lastly, the number of articles submitted to journals should be increased by freeing up more time for the authors and by overcoming the fear of failure to publish (given that once written and perhaps after several revisions, a manuscript has little chance of never being published ). However, a certain number of communications presented at the SOFMER congress have training and teaching components and thus constitute continuing medical education for practitioners. This aspect is substantial for the SOFMER congress and must not be neglected – even though these communications are not intended for scientific publication.

The present study had several limitations. Firstly, the overall publication rate is probably an under-estimate: in fact, 45% of the questionnaire’s respondees stated that the submission/publication process was underway, which could increase the publication rate in the years to come. It might therefore be of value to repeat this analysis in a few years’ time. Secondly, one can question whether the 2008 congress is representative of the SOFMER annual congresses as a whole; this would limit generalization of the present results but suggests that it would be useful to perform comparative analyses over several years. Furthermore, the PubMed database does not list all scientific publications (as was the case for nine abstracts in our study). Furthermore, the reply rate for the questionnaire sent to authors of non-published abstracts was relatively low, which is often the case for this type of survey . Although one can suppose that a certain number of authors did not receive the questionnaire (due to a change in workplace or e-mail address, for example), we are not aware of the characteristics of the non-respondees; this lack of response might have limited the impact of our results by introducing bias.

1.5

Conclusion

The publication rate for abstracts presented at the SOFMER’s 2008 annual congress was 21.2% and is at the lower limit of the range of values reported for other national and international congresses (including PRM congresses). The status of the group submitting the work and the type of work appear to be determinants of publication. Much work had not been submitted to a journal at all; the main obstacles to publication were lack of time and the preliminary nature of the results (from aborted or on-going studies). These results underline the progress that PRM has still to make in terms of improving its level of scientific efficiency while taking account, of course, of the specialty’s particular characteristics.

Disclosure of interest

A. Thevenon is a member of SOFMER and is Editor-in-Chief of the Annals of Physical and Rehabilitation Medicine , the SOFMER’s house journal.

2.1

Introduction

Les congrès scientifiques sont l’occasion pour les équipes médicales et scientifiques de présenter des travaux qui apportent une nouveauté dans la discipline dans laquelle elles exercent. Les présentations orales et affichées qui y sont présentées sont sélectionnées sur les seules bases des résumés, avec tous les risques de biais que cela comporte .

La publication d’une étude dans une revue internationale avec comité de lecture est le meilleur gage d’une large diffusion des résultats et de leur intégration dans le débat scientifique. Du fait d’un processus de sélection plus poussé, notamment de validation par les pairs, seuls une partie des travaux présentés dans un congrès sont publiés dans une revue ; c’est ainsi que le taux de publication des résumés est considéré comme un indice du niveau de qualité scientifique d’un congrès . Ces taux de publications sont très variables toutes disciplines confondues, allant de 20 à plus de 60 % , tandis que pour les congrès organisés par les sociétés scientifiques de médecine physique et de réadaptation (MPR), les études montrent qu’ils varient de 25 et 34 % . Mais le taux brut de publication n’est qu’un reflet partiel. En effet, certaines équipes ne soumettent leurs travaux à aucune revue, notamment quand les résultats sont négatifs, et il existe parfois des différences importantes entre les données présentées en congrès et celles figurant dans l’article final.

La Société française de médecine physique et de réadaptation (SOFMER) organise chaque année son congrès, qui réunit des équipes principalement françaises, européennes et africaines. De nombreux travaux y sont présentés et, à notre connaissance, aucune donnée concernant leur publication ultérieure n’est disponible.

L’objectif principal de cette étude était d’étudier le taux de publication des travaux présentés lors du congrès de la SOFMER 2008 et les facteurs de succès qui y étaient associés. Les objectifs secondaires étaient d’une part d’étudier les différences entre les données des résumés présentés au congrès et celles de la version finale de l’article, et d’autre part de s’enquérir des raisons de la non-publication des travaux.

2.2

Matériel et méthode

2.3

Résultats

2.3.1

Caractéristiques des travaux présentés à la SOFMER 2008

Deux cent trente et une communications ont été présentées lors du congrès de la SOFMER 2008, dont 144 communications orales (62,3 %) et 87 affichées (37,7 %). Les pathologies neurologiques et de l’appareil locomoteur représentaient plus des deux tiers des travaux présentés ( Tableau 1 ). Le Tableau 2 présente les principales caractéristiques de ces travaux. Ils étaient issus en grande majorité d’équipes françaises (82,3 %), loin devant celles du Maghreb (11,3 %) puis des autres pays européens et africains. Ces équipes étaient universitaires pour 77,9 % d’entre elles. Les travaux présentés étaient en majeure partie des études originales (56,7 %), puis des présentations de pratique (23,8 %), des report de cas (16,9 %) et des revues de la littérature (2,6 %). Parmi les études originales, toutes étaient cliniques, le plus souvent monocentriques, la médiane du nombre de sujets inclus s’élevait à 35. Parmi les études thérapeutiques, seules quatre (7,3 %) étaient contrôlées randomisées et leurs résultats étaient le plus souvent positifs.

Tableau 1

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

Should abstracts for rehabilitation congresses be subject to more rigorous selection? Interest of rehabilitation in healing and preventing recurrence of ankle sprains Psychosocioprofessional and familial becoming of formers Buruli ulcer patients in Benin Analyse de livre Walking tests during the exercise training: Specific use for the cardiac rehabilitation Gait characteristics of post-poliomyelitis patients: Standardization of quantitative data reporting

Stay updated, free articles. Join our Telegram channel

Join