Hoeg 2009.

Methods	Identification of subsequent full‐length publications Searched electronic database Embase, PubMed, Google/Google Scholar to April 2007 Search completed by trained librarian Searched by all authors and keywords Matching criteria not reported Contacted abstract authors directly
Data	Included 559 abstracts presented at the 1997, 1999, and 2001 American Society of Clinical Oncology meetings Included all abstracts of phase II trials, excluded those with preliminary results, ongoing studies, or those with a 'phase III component'
Comparisons	Proportion of abstracts published, overall and by meeting Median time to publication by meeting and overall Cumulative proportion of abstracts published Sample size > 35 versus sample size < 35 Oral or poster presentation versus print only Industry sponsor versus none North American or European versus rest of the world origin Subspecialty Type of treatment
Outcomes	361 of 559 abstracts published 90/126 abstracts presented at the 1997 meeting, 124/177 at the 1999 meeting, and 147/256 at the 2001 meeting published Proportion of abstracts published by time Median time to publication = 41 months (95% CI = 37 to 46 months) overall Median time to publication = 33 months for 1997 American Society of Clinical Oncology meeting Median time to publication = 41 months for 1999 American Society of Clinical Oncology meeting Median time to publication = 48 months for 2001 American Society of Clinical Oncology meeting Cumulative proportion of abstracts published at 84 months showed proportion published = 66.8% (373/559 abstracts) Factors related to proportion of abstracts published included 76/259 abstracts with sample size > 35 versus 119/296 abstracts with sample size < 35 published 32/39 abstracts presented orally versus 163/233 abstracts presented as posters versus 166/287 abstracts appearing in print only published 117/174 abstracts with an industry sponsor versus 207/251 abstracts with a non‐industry sponsor versus 37/134 abstracts without funding or not reporting funding published 192/294 abstracts originating from North America versus 116/177 abstracts originating from Europe versus 53/88 abstracts with rest of world origin published
Notes	Oncology Funding by the Gundersen Lutheran Center for Cancer and Blood Disorders and the Gundersen Lutheran Medical Foundation
Risk of bias
Item	Authors' judgement	Description
Sampling method?	Yes	Included all abstracts that described specific subgroup with reasonable exceptions, so low risk of bias.
Search for publications?	Yes	Searched 3 databases and contacted abstract authors.
Follow‐up time?	Yes	All meetings had at least 6 years follow‐up.
Matching?	Unclear	Matching criteria not reported.
Adjustment for confounding?	Unclear	Examined association of meeting year, sample size, type presentation, industry sponsor, country of origin, type treatment, and subspecialty with publication using stratified analysis and Chi2 tests and Cox regression models for time to publication.