Conversion of Plastic Surgery meeting abstract presentations to full manuscripts: a brazilian perspective

Objective: to assess the conversion rate of Plastic Surgery meeting abstract presentations to full manuscript publications and examine factors associated with this conversion. Methods: we assessed the abstracts presented at the 47th and 48th Brazilian Congresses of Plastic Surgery by cross-referencing with multiple databases. We analyzed the Abstracts’ characteristics associated with full manuscript publications. Results: of the 200 abstracts presented, 50 abstracts were subsequently published in full, giving the conference a conversion rate of 25%. The mean time to publish was 15.00±13.75 months. In total, there were 4.93±1.63 authors per abstract and 67.8±163 subjects per abstract; 43.5% of the abstracts were of retrospective studies; 69% comprised the plastic surgery topics head and neck, and chest and trunk, and 88.5% had no statistical analysis. Overall, 80% of the manuscripts were published in plastic surgery journals, 76% had no impact factor and 52% had no citations. Bivariate and multivariate analyses revealed the presence of statistical analysis to be the most significant (p<0.05) predictive factor of conversion of abstracts into full manuscripts. Conclusion: the conversion rate found from this bibliometric research appeared a bit lower than the conversion trend of international plastic surgery meetings, and statistical analysis was a determinant of conversion success.


INTRODUCTION
B razil's scientific production has increased substan- tially, reaching an average growth rate of 10.7% per year 1 .However, as presentations of abstracts at different Brazilian medical meetings have been accompanied by a relatively low conversion rate into full manuscript publications in peer-reviewed, indexed journals [2][3][4][5][6][7][8] , this scientific growth could potentially be higher.Therefore, a continuous analysis of the conversion rate should be performed to encourage members of each academic society to publish full manuscripts [2][3][4][5][6][7][8] .
In this context, there has been no data about the conversion rate from Brazilian Plastic Surgery meetings, although international Plastic Surgery meetings have been analyzed [9][10][11][12][13] .In addition, we are not aware of any investigation of possible predicting factors for the publication of plastic surgery meeting abstracts as full manuscripts [9][10][11][12][13] .Thus, the purposes of this quantitative, descriptive, bibliometric study were to assess the conversion rate of Brazilian plastic surgery meeting abstracts into full peer-reviewed, indexed manuscripts and to examine possible predicting factors of this conversion.We hypothesized that despite increased diffusion of the necessity of full manuscript publications within the Brazilian plastic surgery community 14,15 , the conversion rate after Brazilian plastic surgery meeting presentations would be inferior than the international plastic surgery trends [9][10][11][12][13] .

Identification of abstracts
Two independent authors identified and analyzed the abstracts through examination of the online supplements published by the Brazilian Society of Plastic Surgery (Sociedade Brasileira de Cirurgia Plástica -SBCP) from the 47 th and 48 th Brazilian Congresses of 1 -SOBRAPAR Hospital, Institute of Plastic and Craniofacial Surgery, Campinas, Sao Paulo State, Brazil.

Original Article A B S T R A C T
Objective: to assess the conversion rate of Plastic Surgery meeting abstract presentations to full manuscript publications and examine factors associated with this conversion.Methods: we assessed the abstracts presented at the 47th and 48th Brazilian Congresses of Plastic Surgery by cross-referencing with multiple databases.We analyzed the Abstracts' characteristics associated with full manuscript publications.Results: of the 200 abstracts presented, 50 abstracts were subsequently published in full, giving the conference a conversion rate of 25%.The mean time to publish was 15.00±13.75months.In total, there were 4.93±1.63authors per abstract and 67.8±163 subjects per abstract; 43.5% of the abstracts were of retrospective studies; 69% comprised the plastic surgery topics head and neck, and chest and trunk, and 88.5% had no statistical analysis.Overall, 80% of the manuscripts were published in plastic surgery journals, 76% had no impact factor and 52% had no citations.Bivariate and multivariate analyses revealed the presence of statistical analysis to be the most significant (p<0.05)predictive factor of conversion of abstracts into full manuscripts.5][6][7][8][16][17][18] .

Full manuscript publication search
We identified the publications in peerreviewed journals by a standardized searching of the Medline (PubMed), ISI Web of Knowledge, SciELO, Lilacs, and Google Scholar databases through March 2015.We used the last name and the first letter of the first name of the first author of the abstracts.If we dis not find an exact match or if there were no results to a search, we repeated the process using the second and the last authors of the abstracts.If the result included no publication or several publications with the same author, we applied an additional criterion, such as another author or keywords from the title or text of the abstract to expand or simplify the search.
Whenever we retrieveda peer-reviewed manuscript, we compared the abstract information with those of the full publication to establish whether they matched in accordance with the stringent criteria previously used [16][17][18] .
For each matching publication, we recorded the following data: period between the abstract presentation and the publication (<12 months; 12-36 months; >36 months); journal name; database indexing journal (ISI Web of Knowledge and Medline; Medline; SciELO; or Lilacs); impact factor of the journal according to the Thompson Reuters Journal Citation Report ® on the date of publication; language of the publication (English, Portuguese, or English/Portuguese); and the number of citations of the manuscript according to Google Scholar and ISI Web of Knowledge (if indexed journal) 6,9,12,[16][17][18] .

Statistical analyses
For the descriptive analysis, we used the mean and standard deviations for numerical variables and percentages for categorical variables.We defined the conversion rate as the ratio between the number of subsequently published articles in peer-reviewed journals and the total number of abstracts presented at the scientific meetings 18 .We performed intra-and interperiod (period 1 versus period 2) comparisons.We used Analysis of Variance, Equality of Two Proportions, Paired Student's T, Chi-Square, and Confidence Interval for the Mean tests for statistical comparisons.We performed bivariate and multivariate logistic regression analyses to determine which independent abstract variables (number of authors, number of subjects, plastic surgery topic, presence of statistical analysis and study designs) were significant predictors for the conversion of abstracts into full manuscripts (dependent variable).We also calculated theinterobserver reliability between the two authors who collected all of the data.We performed all analyses using the software program Statistical Package for Social Science (SPSS version 20.0 for Windows, Chicago, IL, USA).We considered as significant values with a confidence interval of 95% (p < 0.05).

RESULTS
We included two hundred meeting abstracts in this bibliometric study, with an excellent interobserver agreement (all values between 0.86 and 0.99) for all evaluated abstracts and manuscript variables.

Meeting abstract characteristics
Overall, there was a significant (all p < 0.05) predominance in the proportion of abstracts including: more than six authors; 11-50 subjects/abstract; head and neck, and chest and trunk plastic surgery topics; absence of statistical analysis; and retrospective study design.The inter-period analyses revealed a significant  1).

Full manuscript characteristics
Overall, there were 50 (25%) matching full manuscripts, and a significant (all p < 0.05) predominance in the proportion of abstracts published as full manuscripts within 12 months after meeting presentations, manuscripts in Portuguese, manuscripts published in plastic surgery journals, journals indexed in the SciELO database, journals without an impact factor, and manuscripts without citations.The interperiod analyses revealed a significant (all p < 0.05) reduction (period 1 > period 2) in the proportion of abstracts converted into full manuscripts and manuscripts in English/Portuguese languages (Table 2).

Bivariate and multivariate analyses
Bivariate analysis revealed that abstracts with 1-3 and 4-5 authors/abstract, one subject/abstract, general plastic surgery topic, and case reports were significantly (all p < 0.05) less converted into full manuscripts.Bivariate and multivariate analyses demonstrated that the presence of statistical analysis was the most significant (all p > 0.05) predictive factor of conversion success (Table 3).

DISCUSSION
Abstract presentations at scientific meetings are an integral component of medical research, in which investigators can share their results with others and meeting attendees may obtain new information 18,19 .
Although acceptance of an abstract by a scientific meeting is prestigious, ideally abstracts should be followed by a full manuscript publication in peerreviewed journals for several reasons: key novel findings and useful information should also be available to the general scientific community who did not participate in the meetings; abstracts alone have many defects, inaccuracies, and only preliminary data; abstracts have been devoid of information that is needed for evaluating validity and reliability; and abstracts are accepted for presentations without a peer-review process or have been reviewed less thoroughly than what is typical of journal manuscripts 18,19 .In addition, adverse consequences of research inaccessibility include: unnecessary duplication, delays in the dissemination of advances in patient-care strategies, harm to patients, waste of limited resources, and loss of (trust in) scientific integrity 20 .In fact, research ethical obligations require the appropriate dissemination and publication of all research outcomes 21 , and empirical evidence indicates the existence of research dissemination bias, as published studies tend to be systematically different from unpublished studies 22 .
It is also noteworthy that the conversion rate may be regarded as an indicator of the scientific level of a certain society's meeting 12 .Interestingly, however, a considerable proportion of medical meeting abstracts have never been published as full manuscripts 18 .In Brazil, the publication rates from different medical fields were previously established [2][3][4][5][6][7][8] .To the best of our knowledge, there has been no similar research from a Brazilian plastic surgery perspective, although there are recent bibliometric plastic surgery data 14,15 .
Therefore, in response to an identified gap in the national plastic surgery literature, we quantified conversion rates from the major Brazilian plastic surgery meeting.Our rationale for evaluating this particular meeting was threefold: first, it is the most important scientific plastic surgery event in Brazil; second, it is applicable to a wide plastic surgery audience, as it is general in nature; and third, Brazilian Portuguese was the primary language of presentation.Plastic surgeons are invited to submit scientific abstracts that are reviewed for relevance and scientific value by a scientific committee called that Department of Scientific Events of SBCP.Our data is mainly from a plastic surgery perspective, but our bibliometric research is also important for the overall Brazilian scientific  proposed in other medical fields 2,7,16,17 .
The abstract characteristics (number of subjects, number of authors, and plastic surgery topics) are in conformity with the previous bibliometric plastic surgery meeting trends [9][10][11][12][13] .Moreover, some of the characteristics of the manuscripts are different from previous published trends [9][10][11][12][13] , with most abstracts being published in journals without an impact factor and only indexed in SciELO.More than 43% of the abstracts were published in the Brazilian Journal of Plastic Surgery (Revista Brasileira de Cirurgia Plástica), the official journal of the SBCP.Previous reports [9][10][11][12][13] also highlight that abstracts presented at meetings organized by scientific societies are preferentially submitted for publication to the official journals.
We also investigated independent abstract factors that may predict full manuscript publication, as they have not been previously investigated in the plastic surgery meeting literature [9][10][11][12][13] .Bivariate analysis revealed that abstracts with 1-3 and 4-5 authors/abstract and general plastic surgery topics were significantly less converted in full manuscripts.The increase in the number of authors per article has been evident in scientific literature, and may reflect a progressive complexity in academic work 23 , which are potentially more likely to be published, and the increasing focus on subspecialtization of plastic surgery 24 may partially explain these findings.However, as we have not ascertained the connection, future research should test these hypotheses.
We also showed that the presence of statistical analysis was a significant determinant for conversion success, as exhibited in different societies' meetings 16,17 .
Although the provision of appropriate statistical tests is an important quality criteria of abstract reporting 16 , details about statistical testing were provided in only 11.5% of the evaluated abstracts; and interestingly, many retrospective studies (the major study design in our report) had a potential sample of patients, but no statistical analysis was applied.Therefore, identifying the right statistical test as an essential component of research design and ensuring that data are correctly obtained with sufficient power to address the original research hypothesis 25 should be the target of plastic surgeons in the future.
In addition, our bivariate analysis showed that abstracts with only one subject and case reports were significantly less converted into full manuscripts, whereas the study design was not significant in predicting the conversion of abstracts into full manuscripts in the multivariate analysis.Although one can assume that studies with increased methodological rigor, such as randomized clinical trials and prospective studies are likely to have higher conversion rates than case series and case reports, there are mixed results regarding the study design of meeting abstracts as a determinant factor of conversion success [16][17][18] .As we also showed an increase of case reports and a decrease of retrospective studies, efforts to increase publishing research with more thorough research methods in Brazilian plastic surgery must be promoted as evidencebased medicine achieves greater acceptance within the plastic surgery community 14,15 .
We found that 75% of abstracts presented at the Brazilian Congress of Plastic Surgery remained unpublished after four years.The reasons why some scientific meeting abstracts remain unpublished are not entirely clear and probably act in a complex multifactorial format 19,26 .Rejection by journals may be a cause of nonpublication; however, as most of the unpublished studies have not been submitted to journals, non-publication of many studies was directly caused by failure of authors to write and submit it to journals 26 .A recent systematic review 19 revealed that among different factors (e.g., "lack of resources," "publication not an aim," "low priority," "incomplete study," and "trouble with co-authors"), "lack of time" was the most frequently reported reason and the most important reason for not subsequently publishing abstracts as full-length manuscripts.The plastic surgery literature 27 has demonstrated that a median of 177 hours is needed to take a retrospective study from idea genesis to publication, and the number of authors or subjects is not correlated with the hours spent per publication, whereas different factors (e.g., medical students, residents, and data collection) were associated with more hours spent per plastic surgery publication 27 .
As we are unable to directly address the real reasons that may have influenced non-publication of abstracts, further studies should be delineated to investigate this issue.
In this context, our data reinforced previous anecdotal perceptions that regular scientific productivity has not been the rule among Brazilian plastic surgeons 14,15 .In fact, Brazil was poorly ranked (number of articles and impact factor) in recent worldwide bibliometric plastic surgery analyses 28,29 , although the Brazilian plastic surgery communitywith more than 5,500 plastic surgeons and 83 plastic surgery residency programs accredited by SBCP -is among one of the biggest plastic surgery communities worldwide.
Therefore, to improve conversion rates, we need major changes in the Brazilian plastic surgery research culture.Given the necessity of disseminating all research findings, mandatory full manuscript submission for publication before meeting presentation may be considered.Although this might be a radical notion and reduce the number of abstracts submitted, the result would be an increased publication yield 10,18 .
Senior authors and experienced research teams should be encouraged to assist junior authors and incipient scientific groups with study completion and manuscript preparation 10,18 .In addition, the next generation of plastic surgeons should receive the knowledge that "an abstract is only a work in progress 30 ," as emphasized by Dr.Joseph Murray, a Nobel laureate plastic surgeon, during the residency training process.For these culture changes, governmental, departmental and SBCP support is urgently required,for both faculties and trainees, including dedicated research time and a research infrastructure.It will be a long-term journey, but with the support of the entire Brazilian plastic surgery community, it is possible to get there.Some inherent limitations to this study design may have influenced our results and must be considered when interpreting our findings.The variability of conversion rates may be related to the quality of meeting presentations, alterations in the criteria for abstract selection, the stringency of the publication criteria for the specialty's peer-review journals, differences of scientific profile of each specialty and medical society, and/or the inherent methodology of determining conversion rates [2][3][4][5][6][7][8][9][10][11][12][13][16][17][18] . Thus, he trend comparisons performed in our study should be interpreted with caution.Our results likely underestimate the overall Brazilian conversion rate after meeting presentations.Given the bibliometric nature of this study, there is the potential for selection bias; it is possible that the search strategy did not accurately identify all publications, although our methods of selection, inclusion, and analysis have been found in previous similar investigations [2][3][4][5][6][7][8][9][10][11][12][13][16][17][18] .We only looked at abstracts presented at one particular plastic surgery meeting, mirroring previous studies [2][3][4][5][6][7][8][9][11][12][13]16,17 .Thus, it is possible that the analysis of other meetings would show different rates of Brazilian plastic surgeons' contributions.Our study is also limited by the data provided by the investigators in the abstracts, as there was no information about the type of institution Finally, a minimum 3-year period was fixed to allow time for publication of each meeting abstract, as the vast majority of publications appeared in the first three to four years after meeting presentations [2][3][4][5][6][7][8][9][10][11][12][13][16][17][18] .
community, as it allows a quantitative assessment of the relative contribution from Brazilian plastic surgery investigators to the scientific scenario.It also permits the initiation of a critical reflection on actual scientific production, with potential modifications of planning by the specialty society, academic community, and governmental actions for research induction, financing, human resources training, among others, as has been involved, academic status, gender of the investigators, among others.A further caveat stems from the fact that we assessed the proportional Brazilian contribution in the form of abstracts.It is conceivable that goodquality Brazilian research throughout the years analyzed was submitted for full publication without presentation at the evaluated plastic surgery meeting.Such possibility would lead to an underestimation of Rev. Col. Bras.Cir.2017; 44(1): 017-026Brazilian productivity.An additional limitation is that we only assessed quantitative aspects of the abstracts.
Plastic Surgery, held in 2010 and 2011, respectively.

Table 3 .
Bivariate and multivariate analyses for conversion rate.