AUGS CONFERENCE SUBMISSION

The Fate of Abstracts Presented at Annual Meetings of the American Urogynecologic Society from 2007 to 2008 Tyler M. Muffly, MD,* Camille S. Calderwood, MD,Þ Karlotta M. Davis, MD, MPH,* and Kathleen A. Connell, MD*

Objectives: The purpose of this study was to determine the publication rate of full-text articles after the presentation of abstracts at consecutive annual meetings of the American Urogynecologic Society (AUGS) over a 2-year period. Methods: Using abstract books published by the Journal of Pelvic Medicine and Surgery, we obtained the text of all abstracts presented at the 2007 and 2008 AUGS annual meetings. A literature search was performed in the US National Library of Medicine to identify the peerreviewed publications arising from each of those abstracts and to calculate the full-text publication rate and the mean duration from presentation to publication. Results: The overall full-text publication rate was 56% (239/438 publications). The mean time from presentation to publication was 15.7 months. The published articles appeared in 49 peer-reviewed journals, with notable distribution in the International Urogynecology Journal (27%), American Journal of Obstetrics and Gynecology (26%), and Obstetrics and Gynecology (10%). Oral presentations (85%) were more likely than posters (71%) to have first authors with university affiliations (P = 0.002). The greater the number of abstract authors, the more likely that the article was to be published (P = 0.0059). The odds of an oral presentation being published were 7 times the odds of a poster being published (odds ratio, 6.99; P G 0.001). Conclusions: Because not all presentations are published, it is questionable whether it is acceptable to cite AUGS abstracts that have not passed a journal’s peer review process and to implement their results in clinical practice. Key Words: abstracts, female pelvic medicine and reconstructive surgery, publication rate (Female Pelvic Med Reconstr Surg 2014;20: 137Y140)

E

ach year, scientific meetings provide excellent venues for presentation of exciting data. For the presenters, scientific meetings provide a forum to disseminate information. For the participants, these gatherings provide educational opportunities and updates on current topics of research. Presenters want to know how the scientific community will interpret data and which presentations will be accepted in peer-reviewed journals. Publication in a peer-reviewed journal validates the strength of the presented data and positively affects the research providing the data. The quality of a scientific meeting can be rated based

From the *Division of Female Pelvic Medicine and Reconstructive Surgery, University of Colorado Anschutz Medical Campus, Aurora; and †Department of Obstetrics & Gynecology, Kaiser Permanente Medical Center, Denver, CO. Reprints: Tyler M. Muffly, MD, Division of Female Pelvic Medicine and Reconstructive Surgery, University of Colorado Anschutz Medical Campus, Aurora, CO. E-mail: [email protected]. Presented as a poster at the 34th Annual Scientific Meeting of the American Urogynecologic Society, October 17, 2013. No reprints available. The authors have declared they have no conflicts of interest. Copyright * 2014 by Lippincott Williams & Wilkins DOI: 10.1097/SPV.0000000000000081

Female Pelvic Medicine & Reconstructive Surgery

&

on the percentage of presented abstracts that achieve publication in peer-reviewed journals. How many of the abstracts presented at scientific meetings make it to publication in peer-reviewed journals? A review of the literature shows that the range of abstract publication for different medical specialties is between 21% and 69%.1,2 Publication rates for dentistry meeting abstract publications are 21%, whereas otorhinolaryngology is on the higher range at 69%. In the middle lies anesthesiology (43%Y50%),3,4 general surgery (57%),5 and ophthalmology (60%Y66%).6,7 These ranges of publication rates give a measure of the quality of each scientific meeting. Female pelvic medicine and reconstructive surgery has seen an increase in research publications since the first Gynecologic Urology Society meeting in 1979. The fate of abstracts presented in the American Urogynecologic Society (AUGS) has not been evaluated. The purpose of this study was to determine the publication rate of full-text articles after the presentation of abstracts at consecutive annual meetings of the AUGS over a 2-year period.

MATERIALS AND METHODS The AUGS meeting abstract compendiums published by the Journal of Pelvic Medicine and Surgery were retrospectively reviewed for the 2007 through 2008 meeting years. Abstract regarding videos, tips, and tricks were excluded. These years were selected to allow enough time for publication. The US National Library of Medicine (pubmed.gov) database was searched to investigate which abstracts resulted in publication. The lead author and a key word from the title were cross-indexed in PubMed. If this initial search did not find a publication, then the last author and key subject word were cross-indexed to find a published article. If an article was still not found, then it was recorded as an abstract that had not been published in a peerreviewed journal retrievable by PubMed. Abstract information was collected and managed using Research Electronic Data Capture tools hosted at the University of Colorado.8 Research Electronic Data Capture is a secure, web-based application designed to support data capture for research studies, providing (1) an intuitive interface for validated data entry, (2) audit trails for tracking data manipulation and export procedures, (3) automated export procedures for seamless data downloads to common statistical packages, and (4) procedures for importing data from external sources. Responses were analyzed using statistical software (SPSS Statistics 20.0). An exemption from the Colorado Multiple Institutional Review Board approval was obtained for this study.

RESULTS There were 438 total abstracts presented at the annual scientific meetings of AUGS in 2007 and 2008 (Table 1). Of the 438 abstracts, 239 (56%) were subsequently published in a peerreviewed journal. Seven abstracts of the 438 abstracts were unable to be confirmed as published or unpublished. The rate of publication for oral presentations (82%) was much higher than

Volume 20, Number 3, May/June 2014

www.fpmrs.net

137

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Muffly et al

Female Pelvic Medicine & Reconstructive Surgery

&

Volume 20, Number 3, May/June 2014

TABLE 1. Characteristics of Oral and Poster Presentations from 2007 to 2008 AUGS Overall Published in a peer-reviewed journal (n = 431) Yes No Time to publication, mo First author, sex (n = 390) Male Female First author institution affiliation (N = 438) University hospital Community hospital Other affiliation First author location (N = 438) United States International First author ACOG district (n = 437) I (CT, ME, MA, NH, RI, VT) II (NY) III (DE, NJ, PA) IV (DC, FL, GA, MD, NC, SC, VA, WV) V (IN, KY, OH, MI) VI (IL, IA, MN, NE, ND, SD, WI) VII (AL, AR, KS, LA, MS, MO, OK, TN) VIII (AK, AZ, CO, HI, ID, MT, NV, NM, OR, UT, WA, WY) IX (CA) XI (TX) International Study design of presented abstracts (n = 420) Cohort Other/basic science Cross-sectional Randomized controlled trial Case-control Case study Meta-analysis Study design of presented abstracts (N = 438) Retrospective Prospective Study subject of presented abstracts (N = 438) Clinical Basic science No. study sites (n = 430) Single site Multisite No. study participants (n = 395) G100 participants Q100 participants National Institutes of Health network participant (N = 438) Yes No

the rate for posters (42%), which represents a significant difference in publication rates for the 2 types of abstract presentations (P G 0.001). Oral presentations (85%) were more likely than posters (71%) to have first authors with university affiliations (P = 0.002),

138

www.fpmrs.net

Poster

Oral P

n

%

n

%

n

%

239 192 239

56 44 15.7 (12.1)

118 166 118

42 58 17.3 (13.0)

121 26 121

82 18 14.2 (10.9)

G0.001

181 209

46 54

124 127

49 51

57 82

41 59

0.11

331 87 20

75 20 5

205 71 14

71 24 5

126 16 6

85 11 4

402 36

92 8

260 30

90 10

142 6

96 4

0.02

37 25 69 49 47 50 25 33 38 28 36

9 6 16 11 11 11 6 8 9 6 7

29 21 45 30 29 31 18 19 24 14 30

10 7 16 10 10 11 6 7 8 5 10

8 4 24 19 18 19 7 14 14 14 6

5 3 16 13 12 13 5 10 10 10 3

0.07

251 73 43 28 13 10 2

60 17 10 7 3 2 1

182 39 28 15 9 7 0

65 14 10 5 3 3 0

69 34 15 13 4 3 2

49 24 11 9 3 2 2

168 270

38 62

123 167

42 58

45 103

30 70

378 60

86 14

256 34

88 12

122 26

82 18

382 48

89 11

262 26

91 9

120 22

85 15

0.045

197 198

50 50

140 127

52 48

57 71

45 55

0.14

12 426

3 97

1 289

1 99

11 137

7 93

0.048

0.002

0.01

0.01

0.09

G0.001

whereas posters (24%) were more likely than oral presentations (11%) to have first authors with community affiliations. Oral presentations were more likely to have authors from the United States than posters (96% vs 90%; P = 0.02). Posters (42%) were more * 2014 Lippincott Williams & Wilkins

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Female Pelvic Medicine & Reconstructive Surgery

&

Volume 20, Number 3, May/June 2014

Abstracts at the AUGS Annual Meetings

TABLE 2. Journals of Publication for Oral and Poster Presentations from 2007 and 2008 AUGS Overall Publication Journal (n = 239) International Urogynecology Journal American Journal of Obstetrics and Gynecology Obstetrics and Gynecology Other Journals of publication

Poster

Oral

n

%

n

%

n

%

P

64 61 23 91

27 26 10 37

39 12 5 62

33 10 4 53

25 49 18 29

21 40 15 24

0.03 G0.001 0.005 G0.001

the model involved the key study variable and those study design features whose relationships with publication or nonpublication approached significance by having a bivariate P value less than 0.15. The initial full model accounted for approximately 24% of the variability in whether or not the abstract was ultimately published (R2 = 0.241). Among the factors included in the model, only presentation type was a significant predictor of whether or not the abstract found its way to publication. The odds of an oral presentation being published were almost 7 times the odds of a poster presentation being published (odds ratio, 6.99; P G 0.001).

likely than oral presentations (30%) to be a retrospective study design (P = 0.01). The mean (SD) time to publication was 15.7 (12.1) months. The mean (SD) time between presentation and publication for posters was 17 (13) months, whereas it was only 14 (11) months for oral presentations (P = 0.048). Among the 239 published, survival estimates showed that 49% of abstracts were published within 1 year, an additional 28% were published by the end of the second year, an additional 17% during the third year, with 5% in the fourth year, and 2% in the fifth year. Abstracts accepted for oral presentations (58%) were more likely than posters (39%) to be finally published within 1 year (P = 0.004). Oral presentations (84%) were more likely than posters (70%) to be published within 2 years (P = 0.011). There were no differences in publication rate between posters and oral presentations after the second year (P = 0.011). Publication dates for the full-text articles ranged from the year of presentation to December 31, 2012. Publications were identified in 51 different peer-reviewed journals (Table 2), including 3 of the following top journals: International Urogynecology Journal (64 publications, 27% of all publications); American Journal of Obstetrics and Gynecology (61 publications, 26% of all publications), and Obstetrics and Gynecology (23 publications, 10% of all publications). In 127 of the 239 publications (53%), the number of authors was identical with that of the corresponding abstract (Table 3). In 83 publications (35%), at least 1 coauthor had been added. The number of additions ranged from 1 to 7 authors with a median of 1 additional author. For 29 publications (12%), there were fewer authors on the publication than on the abstract. The number of subtractions ranged from 1 to 3 authors with a median of 1. Rarely (3 publications, 1%), the original abstract first author was dropped completely from the publication authorship. Only 10 abstracts had a single author. A multiple logistic regression was performed using a set of key factors to determine the factors that best predict the publication or nonpublication of an abstracts (Table 4). The factors included in

DISCUSSION This study is the first longitudinal review regarding the fate of AUGS abstracts (MEDLINE; January 1966 to November 2013; English language; search terms: ‘‘abstracts,’’ ‘‘female pelvic medicine and reconstructive surgery,’’ and ‘‘publication rate’’). The overall publication rate of abstracts presented at AUGS meetings in 2007 and 2008 was 56%. Oral presentations had a higher publication rate than poster presentations. Not surprisingly, randomized controlled trials and prospective studies had higher rates of publication, as they usually denote higher quality studies. Evaluation of publication rates also provides feedback to researchers regarding the mean time to publication of the presented work. A critical examination of the data provides feedback to the program selection committee regarding the stringency of their selection criteria. The abstracts selected for presentation at the AUGS meetings are subjected to peer review by the program committee, who has to rely on minimal data submitted in abstracts to make their decision. The rate of publication is indicative of the quality of abstracts submitted to the meeting. Large numbers of studies presented in scientific meetings never achieve publication (44% for AUGS). One reason is that authors choose not to proceed with publication. Dickersin et al9 found that ‘‘lack of time’’ is the main reason investigators give for failure to publish. Other reasons researchers give are ‘‘low priority and anticipated rejection.’’10 The dissemination of unpublished

TABLE 3. Author Characteristics of Published Oral and Poster Presentations from 2007 and 2008 AUGS Overall Author count (abstract to publication) (n = 239) Fewer authors on publication Same number of authors More authors on publication

No. abstract authors No. publication authors * 2014 Lippincott Williams & Wilkins

Poster

Oral P

n

%

n

%

n

%

29 127 83

12 53 35

18 63 37

15 53 32

11 64 46

9 53 38

n

Mean (SD)

n

Mean (SD)

n

Mean (SD)

P

438 239

5.0 (2.4) 5.7 (2.7)

290 118

4.8 (1.9) 5.4 (1.9)

148 121

5.3 (3.1) 6.1 (3.3)

0.08 0.05

0.27

www.fpmrs.net

139

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Muffly et al

Female Pelvic Medicine & Reconstructive Surgery

&

Volume 20, Number 3, May/June 2014

TABLE 4. Logistic Regression for Predicting Publication of Abstracts Presented in 2007 and 2008 at the AUGS Meetings Factor Oral versus poster presentation First author university affiliation Retrospective/prospective Multiple/single center Randomized controlled trial (Intercept)

A 1.94 j0.47 0.17 j0.36 j0.72 2.65

df 1 1 1 1 1 1

P G0.001 0.06 0.45 0.36 0.19 G0.001

research outcomes raises ethical concerns, particularly if the studies used prospective patient or animal participation. In addition, nonpublication may lead to the unnecessary repetition of studies by others with additional time, cost, and resource implications. There is no set parameter for what is an appropriate rate of subsequent abstract publication for a scientific meeting. In comparison with other major national meetings, subsequent abstract publication rates were as low as 21% for the dental conferences, International Association for Dental Research/ American Association for Dental Research, and as high as 69% for the Otorhinolaryngological Research Society.2 Unlike these other associations, AUGS has an affiliated peer-reviewed journal and has required submission of a manuscript at the time of oral presentation since 2004. Strict rules on nonsubmission mean that effectively all accepted oral abstracts are converted into articles and submitted for review, thereby increasing the likelihood of publication. Reviews of other medical specialty abstract to publication data reveal that most publishable data achieve publication within 6 years after the abstract presentation at a scientific meeting.11 Ninety-four percent of AUGSpresented abstracts that were eventually published achieved publication within 3 years. The number of single-author publications has decreased, whereas more articles have a large numbers of authors. There were more likely to be additions in the number of authors (35%) than deletions of authors (12%). These author additions may indicate that new researchers are added to complete a project. Furthermore, only a 12% deletion of authors may indicate that once a project is underway, the researchers see it to completion. Although we fully agree that the increase in the number of authors in the recent decades raises several questions about gift authorship, teamwork is an important aspect of science that should not be discouraged. However, research teams should be careful not to gift authorship when not earned. A limitation of this study is that journals not represented on PubMed were excluded. Abstracts may have undergone such major changes to both study design and authorship that they may be unrecognizable in a PubMed search. Furthermore, there was no analysis of rejected abstracts or citation index. This study does not evaluate the citation index of articles published with full text. Our regression analysis was not able to evaluate for factors such as ‘‘timeliness’’ or ‘‘novelty’’ of a study that may have pushed it toward publication as well. We also did not contact authors of unpublished works to find out if they were published elsewhere or never submitted at all. This limitation was because of a lack of contact information for abstract authors. Presentations at female pelvic medicine and reconstructive surgery meetings tend to focus on results and conclusions.

140

www.fpmrs.net

Odds Ratio 6.99 0.62 1.19 0.70 0.49 14.09

95% Confidence Intervals Lower 4.16 0.38 0.76 0.32 0.17

Upper 11.75 1.03 1.85 1.51 1.42

During the presentation, it is difficult to evaluate the quality of study design and data. Despite these shortcomings, many attending the meeting may change their practice strategies based on information that has not undergone the rigorous scrutiny of a journal’s peer review process.12 The benefits of going to the meeting are not solely from the presentations that are likely to be published, and we do not suggest that attendees stay home and just read the published articles. REFERENCES 1. Corry AM. A survey of the publication history of randomly selected IADR/AADR abstracts presented in 1983 and 1984. J Dent Res 1990;69:1453Y1455. 2. Roy D, Sankar V, Hughes JP, et al. Publication rates of scientific papers presented at the Otorhinolarygological Research Society meetings. Clin Otolaryngol Allied Sci 2001;26:253Y256. 3. Yentis SM, Campbell FA, Lerman J. Publication of abstracts presented at anaesthesia meetings. Can J Anaesth 1993;40:632Y634. 4. Meranze J, Ellison N, Greenhow DE. Publications resulting from anesthesia meeting abstracts. Anesth Analg 1982;61:445Y448. 5. Easterbrook PJ, Berlin JA, Gopalan R, et al. Publication bias in clinical research. Lancet 1991;337:867Y872. 6. Juzych MS, Shin DH, Coffey J, et al. Whatever happened to abstracts from different sections of the association for research in vision and ophthalmology? Invest Ophthalmol Vis Sci 1993;34:1879Y1882. 7. Scherer RW, Dickersin K, Langenberg P. Full publication of results initially presented in abstracts. A meta-analysis. JAMA 1994;272:158Y162. 8. Harris PA, Taylor R, Thielke R, et al. Research electronic data capture (REDCap)Va metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform 2009;42:377Y381. 9. Dickersin K, Min YI, Meinert CL. Factors influencing publication of research results. Follow-up of applications submitted to two institutional review boards. JAMA 1992;267:374Y378. 10. Weber EJ, Callaham ML, Wears RL, et al. Unpublished research from a medical specialty meeting: why investigators fail to publish. JAMA 1998;280:257Y259. 11. von Elm E, Costanza MC, Walder B, et al. More insight into the fate of biomedical meeting abstracts: a systematic review. BMC Med Res Methodol 2003;3:12. 12. International Committee of Medical Journal Editors. Uniform requirements for manuscripts submitted to biomedical journals. International Committee of Medical Journal Editors. Ginecol Obstet Mex 2002;70:361Y370.

* 2014 Lippincott Williams & Wilkins

Copyright © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

The fate of abstracts presented at annual meetings of the american urogynecologic society from 2007 to 2008.

The purpose of this study was to determine the publication rate of full-text articles after the presentation of abstracts at consecutive annual meetin...
100KB Sizes 0 Downloads 4 Views