Peng 2006.

Methods	Identification of subsequent full‐length publications Searched electronic database MEDLINE to February 2005 Person completing the search not reported Searched by all authors and keywords Matched abstract to full‐length publication by All authors Title Methodology Results
Data	Included 473 abstracts presented at the 1999 American Academy of Otolaryngology Head & Neck Surgery meeting Included all abstracts
Comparisons	Proportion of abstracts published Median time to publication 'Positive' versus not 'positive' Oral versus poster presentation RCT design versus observational study design versus case series versus opinion 'Prospective' versus 'retrospective' versus cross‐sectional design Case reports versus other Grant support versus other 'Research award' versus other
Outcomes	237 of 473 abstracts published Proportion of abstracts published by time Median time to publication = 16 months (IQR = 12 to 29 months) Factors related to proportion of abstracts published included 189/337 'positive' (defined as experimental better than control) versus 13/26 not 'positive' abstract results published 135/213 abstracts presented orally versus 102/260 abstracts presented as posters published 55/84 abstracts with RCT design versus 96/173 abstracts with observational study design versus 82/209 abstracts with case series design versus 2/3 opinion abstracts published 90/215 abstracts with 'prospective' study design versus 92/222 abstracts with 'retrospective' study design versus 20/36 with cross‐sectional study design published 18/70 abstracts of case reports versus 219/403 abstracts not of case reports published 25/37 abstracts reporting grant support versus 211/435 abstracts without or not reporting grant support published 9/10 abstracts with 'research award' versus 227/462 abstracts without 'research award' published
Notes	Otolaryngology head and neck surgery Funding not reported
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts.
Search for publications?	No	Searched only 1 database.
Follow‐up time?	Yes	The meeting had 6 years follow‐up.
Matching?	Yes	Matched by 4 different criteria.
Adjustment for confounding?	Yes	Examined association of positive results, type of presentation, study design, prospective status, presence of grant support, and award winning with publication using multivariable logistic regression analyses.