SEARCH

SEARCH BY CITATION

Keywords:

  • Abstract;
  • publication rate;
  • systematic review;
  • transplantation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Full publication of abstracts presented at scientific meetings ranges from 25–74%. To determine the rate and factors associated with publication in organ transplantation, we examined abstracts presented at the American Transplant Congress in May 2000. Of 1147 abstracts, 607 (53%) achieved full publication at 4.5 years (mean 1.32 ± 0.88 years). Fifty-nine percent (357/607) were published in three transplantation journals. For randomized trials, the proportion published was 61%. On multivariate analysis, industry sponsorship (OR 1.78; 95% CI 1.04–3.06), basic science research (OR 1.68; 95% CI 1.32–2.14), non-American center (OR 1.67; 95% CI 1.28–2.20) and oral presentation (OR 1.36; 95% CI 1.07–1.73) were independent predictors of full publication. Nearly half of all abstracts presented at a transplantation meeting remain unpublished. This finding needs to be considered when interpreting systematic reviews in the field of transplantation.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

International scientific meetings provide a unique forum for the dissemination of novel research. However, full publication of abstracts presented at a variety of meetings is often low with estimated rates between 25 and 74% (1). Factors that influence full publication are of debate within the scientific literature, however, they do include author confidence in study quality and their motivation to publish (2–5). Acceptance of abstracts for conference presentation, and acceptance for oral presentation in particular, has been associated with successful full-text publication in other specialties (1–4,6,7). Similarly, positive results favoring experimental interventions, randomized controlled trial design, large sample size and industry involvement are frequently associated with increased likelihood for publication (1,3,6,8–10). Abstract publication and analysis of the factors associated with full publication have not been assessed within the field of solid organ transplantation. Accordingly, the objective of this study was to review all abstracts presented at a major transplantation meeting (American Transplant Congress, May 2000) to determine the proportion of presented abstracts that reach full publication and the factors associated with full publication.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

We reviewed all abstracts (n = 1178) included in the Scientific Program of Abstracts (11) for Transplant 2000, the First Joint Annual Meeting of American Society of Transplant Surgeons and the American Society of Transplantation held May 13–17, 2000, in Chicago, Illinois.

For each abstract, data from the following predetermined domains were identified: name of first and senior author, department and country of origin at the time of abstract submission. Abstract qualification included presentation type (oral vs. poster), research type [clinical (involving intact human subjects) vs. basic science], study methodology as defined by traditional epidemiological definitions (12) (randomized controlled trial, cohort study, case-control study, case series or other), organ type, human or non-human study subjects, sample size, reported statistical significance of primary outcome, collaborative group studies and single center versus multi-centered. The later included research studies with two or more independent research sites. When more than one country was listed, the country of the first author was considered the country of origin for the abstract. For the purposes of analysis, organ type was categorized as liver, heart, kidney and other (lung, pancreas, islet, multi-organ, small bowel, cells or skin, hematopoetic, retina). Industry sponsorship was considered present if a pharmaceutical company was listed within the abstract as a sponsor. We did not use the disclosure index included in the program of abstracts to determine industry sponsorship since the specific industry relation was not identified (i.e. disclosure could have been for several reasons such as study sponsorship, speakers fees, advisory board etc.). Primary data abstraction was carried out by three investigators. Validation of abstracted data was carried out by another investigator who randomly selected abstracts for review. Differences were resolved by consensus.

To determine full publication status, an English database search of Medline, PubMed and EMBASE was conducted. The first, second and final author, as well as abstract keywords were searched individually and as combined (AND) search terms. The identified publications were reviewed as per their objectives, methodology and results, to confirm the publication resulted directly from the research described in the corresponding abstract. Time to publication was the difference in days, between the date of publication (rounded to the first day of the month) and date of abstract presentation. The final date for the publication search was November 1, 2004.

Predictors of full-text publication were determined using logistic regression analysis. The odds ratio (OR) and corresponding 95% confidence intervals (CI) were presented for each variable. Predictor variables that showed a trend toward statistical significance (p < 0.2) on univariate testing were included in the multivariate model. Only variables that reached statistical significance (p < 0.05) were retained in the final model. No data imputation was done for missing data. Variables that had more than 5% of values missing were excluded from the logistic regression analysis (13).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

There were 1178 abstracts accepted for presentation. Three abstracts were withdrawn prior to the meeting and 28 abstracts were published as full-text articles prior to presentation at the meeting (range 833 days to 1 day prior to meeting date). Thus, 1147 abstracts were included in this analysis. Abstract characteristics are outlined in Table 1. Of the abstracts accepted for the conference, 547 (48%) were oral presentations, and 600 (52%) were poster presentations. Of these, 653 (57%) were clinical science studies, and 494 (43%) were basic science studies. With regard to study methodology, the majority were observational studies with only 64 (6%) randomized trials (Table 1). Industry involvement was declared in 66 (6%) abstracts. A statistically significant primary outcome was cited in 397 (35%) of the abstracts.

Table 1.  Characteristics of the 1147 abstracts
Characteristicn (%)Published, n (%)
  1. *Other includes lung, pancreas, islet, multi-organ, small bowel, cells or skin, hematopoetic, retina.

Presentation type
 Oral presentation547 (48)308 (56)
 Poster presentation600 (52)299 (50)
Collaborative study56 (5)28 (50)
Research type
 Clinical science653 (57)306 (47)
 Basic science494 (43)301 (61)
Study methodology
 Randomized trial64 (6)39 (61)
 Cohort study563 (49)288 (51)
 Case-control study84 (7)45 (54)
 Case series284 (25)140 (49)
 Other152 (13)95 (63)
Industry sponsor66 (6)45 (68)
Country of origin
 United States804 (70)398 (50)
 Other314 (27)197 (63)
 Not specified29 (3)12 (41)
Organ type
 Kidney382 (33)199 (52)
 Liver250 (22)121 (48)
 Heart180 (16)98 (54)
 Other*335 (29)189 (56)
Primary outcome statistically
 significant (p < 0.05)
 Yes397 (35)208 (52)
 No234 (20)95 (41)
 Not specified516 (45)304 (59)
Multi-center study371 (32)220 (59)
Human subjects710 (62)340 (48)

Four-and-a-half years after initial abstract presentation, 607 of 1147 (53%) abstracts were published as full-text articles. For studies published after presentation at the annual meeting (n = 607), the median time to publication was 1.0 year (mean time 1.32 ± 0.88 years). Of the 64 randomized trial abstracts, 61% were published within 4.5 years of presentation. Eighty percent of abstracts were published within 2 years of presentation (Figure 1). Only 7% of abstracts were published more than 3 years after presentation (Figure 1). Nearly 60% of the published abstracts appeared in three transplant journals (Transplantation Proceedings, Transplantation and the American Journal of Transplantation) (Table 2). Very few abstracts were published in general medical or general surgical journals (Table 2).

image

Figure 1. Time to publication for 1147 abstracts presented at Transplant 2000 Meeting.

Download figure to PowerPoint

Table 2.  Journal of full text publication for 607 abstracts
Journaln (%)
  1. *Each of these journals published fewer than three abstracts.

Transplantation Proceedings185 (30.5)
Transplantation139 (22.9)
American Journal of Transplantation33 (5.4)
Journal of Immunology30 (4.9)
Liver Transplantation17 (2.8)
Journal of Heart & Lung Transplantation13 (2.1)
Clinical Transplantation12 (2.0)
Journal of the American Society of Nephrology12 (2.0)
Annals of Surgery10 (1.6)
Kidney International8 (1.3)
Human Immunology7 (1.2)
American Journal of Kidney Diseases6 (0.9)
Hepatology6 (0.9)
Journal of Clinical Investigation6 (0.9)
New England Journal of Medicine6 (0.9)
Transplant International5 (0.8)
Circulation4 (0.7)
Lancet4 (0.7)
Transplant Immunology4 (0.7)
Clinical Pharmacology & Therapeutics3 (0.5)
Journal of Thoracic & Cardiovascular Surgery3 (0.5)
Microsurgery3 (0.5)
Nephrology Dialysis Transplantation3 (0.5)
Pediatric Transplantation3 (0.5)
Other*85 (14.0)

Variables associated with full-text publication are shown in Table 3. On univariate analysis, the following variables were significant: industry sponsor (OR 1.98; 95% CI 1.16–3.37); basic science research type (OR 1.77; 95% CI 1.40–2.24); non-American center (OR 1.72; 95% CI 1.32–2.24); animal research subject (OR 1.67; 95% CI 1.31–2.13); multi-center study (OR 1.44; 95% CI 1.12–1.86); and, oral presentation (OR 1.30; 95% CI 1.03–1.64). On multivariate analysis, oral presentation (OR 1.36; 95% CI 1.07–1.73), basic science research (OR 1.68; 95% CI 1.32–2.14), industry sponsorship (OR 1.78; 95% CI 1.04–3.06) and non-American center (OR 1.67; 95% CI 1.28–2.20) were statistically significant (Table 3). The variables sample size and statistical significance were excluded from the logistic regression analyses because greater than 5% of data points were missing.

Table 3.  Factors associated with full text publication*
VariableUnadjusted OR (95% CI)p-ValueAdjusted OR (95% CI)p-Value
  1. *Final multivariate model contains only variables with p-value < 0.05.

Industry sponsorship (vs. none)1.98 (1.16–3.37)0.011.78 (1.04–3.06)0.04
Basic science research (vs. clinical)1.77 (1.40–2.24)0.00011.68 (1.32–2.14)0.0001
Non-American center (vs. American center)1.72 (1.32–2.24)0.00011.67 (1.28–2.20)0.0001
Oral presentation (vs. poster)1.30 (1.03–1.64)0.031.36 (1.07–1.73)0.01
Animal research (vs. human)1.67 (1.31–2.13)0.0001
Multi-center study (vs. single center)1.44 (1.12–1.86)0.004
Randomized trial (vs. other study design)1.41 (0.84–2.37)0.19
Collaborative study group0.88 (0.51–1.50)0.65
Organ type (vs. other)
 Liver0.73 (0.52–1.01)0.06
 Kidney0.84 (0.63–1.13)0.25
 Heart0.92 (0.64–1.33)0.67

The univariate and multivariate analyses were repeated after excluding the 185 publications that appeared in Transplantation Proceedings. The univariate analysis yielded similar results with the same six variables being significant: industry sponsor (OR 1.67; 95% CI 1.01–2.75); basic science research type (OR 1.36; 95% CI 1.07–1.73); non-American center (OR 1.39; 95% CI 1.07–1.82); animal research subject (OR 1.30; 95% CI 1.02–1.67); multi-center study (OR 1.31; 95% CI 1.01–1.67); and, oral presentation (OR 1.33; 95% CI 1.04–1.69). In addition, randomized trial design was also significantly associated with publication (OR 1.66; 95% CI 1.00–2.76). On multivariate analyses, oral presentation (OR 1.34; 95% CI 1.05–1.72), basic science research (OR 1.37; 95% CI 1.07–1.77) and non-American center (OR 1.36; 95% CI 1.03–1.78) remained independent predictors of full-text publication. In addition, randomized trial design remained statistically significant (OR 1.82; 95% CI 1.07–3.10) while industry sponsorship was no longer an independent predictor of publication.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Publication of full-text research is crucial not only for the dissemination of knowledge, but also in creating critical dialogue around novel research findings. We found that 53% of abstracts presented at a major transplantation meeting attained full publication by four-and-a-half years, with the majority being published within 2 years. This finding is consistent with studies in other medical specialties. A recent meta-analysis of 79 studies evaluating publication rates following abstract presentation (29 729 abstracts) revealed an abstract publication rate of 44.5% at 2 years, and an estimated publication rate of 52.6% at 9 years (1). The median time to publication was 17 months (1).

We found that oral presentation was associated with full-text publication. The influence of podium presentation on publication has been well documented in other specialties (1,2,6,9,14–16). Selection for an oral presentation may be seen as a marker for study quality. However, there may also be bias in selecting positive outcome studies for presentation. Koren et al. showed that abstracts rejected for presentation at a scientific meeting were more likely to report negative findings but that the quality was higher than the abstracts with positive findings (17).

The impact of industry sponsorship on publication is controversial. While we and others (9,16) have shown an association with industry sponsorship, other reports have shown that industry funded research is less likely to be published (18,19). Earlier studies have shown that the most common reasons for non-publication were lack of time, funds or other resources by the investigator (2–5). One could speculate that the industry partner may be able to provide needed resources to assist with eventual full-text publication. Scherer et al. support this concept as they found that some form of external financial support was associated with subsequent publication (1). We also found that basic science abstracts were more likely to be published compared to clinical research abstracts. This result was previously documented in the systematic review by Scherer et al. (1). They found that clinical science abstracts had a 21% reduction in publication compared to basic science abstracts (1). Our data do not explain why basic science abstracts are more likely to be published but one could speculate that this is just a marker for external funding (1). However, confirmation of this would require a more detailed study of abstract funding, which has been recommended by Scherer et al. (1).

Publication of abstracts originating from non-American centers was significantly higher compared to American centers. This finding has been previously documented in the gastroenterology literature. Eloubeidi et al. found that abstract acceptance to the 1994 American Society of Gastrointestinal Endoscopy conference did not favor studies from outside the United States (20). However, full publication of abstracts was significantly higher for the non-American centers compared to the American centers (20). In contrast, Timmer et al. found that country of origin was not associated with abstract publication following pancreas meetings (21). They did, however, show that abstracts originating from North America or Western Europe were more likely to be published in a high impact journal (21).

Certain limitations of our study methodology should be noted. This analysis was highly dependent on the quality of the abstracts. Incomplete reporting of clear study objectives and methodology hindered data abstraction. We attempted to minimize inconsistent data abstraction by establishing a priori definitions as related to study methodology (12) and limiting the number of investigators abstracting data. Our study likely underestimated the true publication rate, as identification of articles by an electronic database search is inherently limited with retrieval rates estimated at 80% (3,22). Finally, despite a detailed analysis of key abstract characteristics, we did not assess abstracts for quality since rating schemas are specific for certain types of trials and cannot be broadly and reliably applied across basic science and clinical studies.

In conclusion, this analysis has shown that nearly half of all abstracts presented at a transplantation meeting remain unpublished. Further methodological research needs to be conducted to determine the effect of including or excluding data that are published only in abstract form in transplantation systematic reviews.

Acknowledgment

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Dr. Gourishankar is a clinical investigator funded by the Alberta Heritage Foundation for Medical Research.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References
  • 1
    Scherer RW, Dickersin K, Langenberg P. Full publication of results initially presented in abstracts. A meta-analysis. JAMA 1994; 272: 158162.
  • 2
    Timmer A, Hilsden RJ, Cole J, Hailey D, Sutherland LR. Publication bias in gastroenterological research – a retrospective cohort study based on abstracts submitted to a scientific meeting. BMC Med Res Methodol 2002; 2: 110.
  • 3
    De Bellefeuille C, Morrison CA, Tannock IF. The fate of abstracts submitted to a cancer meeting: factors which influence presentation and subsequent publication. Ann Oncol 1992; 3: 187191.
  • 4
    Weber EJ, Callaham ML, Wears RL, Barton C, Young G. Unpublished research from a medical specialty meeting. JAMA 1998; 280: 257259.
  • 5
    Sprague S, Bhandari M, Devereaux PJ et al. Barriers to full-text publication following presentation of abstracts at annual orthopedic meetings. J Bone Joint Surg 2003; 85: 158163.
  • 6
    Callaham ML, Wears RL, Weber EJ, Barton C, Young G. Positive-outcome bias and other limitations in the outcome of research abstracts submitted to a scientific meeting. JAMA 1998; 280: 254257.
  • 7
    Carrol AE, Sox CM, Tarini BA, Ringold S, Chirstakas DA. Does presentation format at the Pediatric Academic Societies’ annual meeting predict subsequent publication. Pediatrics 2003; 112: 12381241.
  • 8
    Dickersin K, Min YI, Meinert CL. Factoring influencing publication of research results. Follow-up of applications submitted to two institution review boards. JAMA 1992; 267: 374378.
  • 9
    Krzyzanowska MK, Pintilie M, Tannock IF. Factors associated with failure to publish large randomized trials presented at an oncology meeting. JAMA 2003; 290: 495501.
  • 10
    Ioannidis JP. Effect of the statistical significant of results on the time to completion and publication of randomized efficacy trials. JAMA 1998; 279: 281286.
  • 11
    Scientific Program of Abstracts, Supplement to Transplantation (Volume 69 (8) April 27, 2000). Original presentation of abstracts at ‘Transplant 2000’, the 1st Joint Annual Meeting of ASTS and AST held May, 2000, in Chicago, sponsored by AST (American Society of Transplantation).
  • 12
    Hennekens CH, Burnings JE. Epidemiology in Medicine. 1st Edn. USA : Little Brown & Company. 1997.
  • 13
    Allsion PD. Missing Data. Thousand Oaks , California : Sage Publications. 2001: 1104.
  • 14
    Gilbert WM, Pitkin RM. Society for maternal-fetal medicine meeting presentations: what gets published and why? Am J Obstet Gynecol 2004; 191: 3235.
  • 15
    Juzych MS, Shin DH, Coffey J, Juzych L, Shin D. Whatever happened to abstracts from different sections of the association of research in vision and ophthalmology? Invest Ophthalmol Vis Sci 1993; 34: 18791882.
  • 16
    Barnes DE, Bero LA. Why review articles on the health effects of passive smoking reach different conclusions. JAMA 1998; 279: 15661570.
  • 17
    Koren G, Graham K, Shear H, Einarson T. Bias against the null hypothesis: the reproductive hazards of cocaine. Lancet 1989; 2: 14401442.
  • 18
    Lexchin J, Bero LA, Djulbegovic B, Clark O. Pharmaceutical industry sponsorship and research outcome and quality: systematic review. BMJ 2003; 326: 11671170.
  • 19
    Bekelman JE, Li Y, Gross CP. Scope and impact of financial conflict of interesting biomedical research: a systematic review. JAMA 2003; 289: 454465.
  • 20
    Eloubeidi MA, Wade BS, Provenzal B. Factors associated with acceptance and full publication of GI endoscopic research originally published in abstract form. Gastrointest Endosc 2001; 53: 275282.
  • 21
    Timmer A, Blum T, Lankisch PG. Publication rates following pancreas meetings. Pancreas 2001; 23: 212215.
  • 22
    Hopewell S, Clarke M, Lusher A, Lefebvre C, Westby M. A comparison of handsearching versus MEDLINE searching to identify reports of randomized controlled trials, (abstract). Stat Med 2002; 2: 16251634.