Vlassov 2008.

Methods	Investigated discrepancies between systematic review abstracts and the main text.
Data	N = 100 systematic reviews Inclusion criteria: New, updated, and existing systematic reviews published in Issue 1, 2008 of the Cochrane Database of Systematic Reviews with included studies. Exclusion criteria: Cochrane Methodology Reviews; Cochrane Reviews without included studies. Year(s) of publication of the systematic reviews: 1996 to 2007 Areas of health care addressed by the systematic reviews: Various (not specifically assessed) Methodological quality of systematic reviews: Not assessed Number (proportion) of systematic reviews that only included RCTs: Not assessed Number (proportion) of systematic reviews that are Cochrane Reviews: 100 (100%) Extent of overlap of RCTs included in the systematic reviews in the empirical study: Not assessed, though unlikely to have impacted on the results
Comparisons	Abstract versus main text
Outcomes	Prevalence of systematic review abstracts which: report the results of other outcomes before the results of the primary outcome; report the results of other outcomes before the results of the primary outcome (where the other outcome was statistically significant); report a primary outcome differently to the primary outcome reported in the main text (defined as different phrasing and unclear description).
Notes	Statistical significance defined as P < 0.05. Additional outcomes measured but not included in the review were prevalence of abstracts reporting the magnitude, direction and statistical uncertainty (e.g. P value, 95% CI) of effect estimates (i.e. measures of 'incomplete reporting') Study published as a conference abstract only. Additional unpublished data retrieved from the author.
Risk of bias
Item	Authors' judgement	Description
Is the empirical study at low risk of selection bias?	Yes	Quote: "It was not a random, but really 'regular' sample: having the DB [database] open in alphabet order I selected first 10 of every 100 abstracts. Reviews without trials found were excluded" (personal communication). Comment: This empirical study included a representative systematic sample of systematic reviews.
Is the empirical study at low risk of selective reporting bias?	Unclear	Comment: No protocol for the study is available so it is unclear whether all measured and analysed outcomes were reported in the conference abstract.