Skip to main content

Table 3 Characteristics of the random sample of reviews with at least one included study rated at high risk of outcome non-reporting bias

From: Rethinking the assessment of risk of bias due to selective reporting: a cross-sectional study

Characteristics Number (%), of n = 100
Number of included studies  
 Total number of studies included in review, median (IQR) 13 (7–32)
 Number of studies per review at high risk of outcome non-reporting bias, Median (IQR) 2 (1–5)
 Percentage of studies per review at high risk of outcome non-reporting bias, Median (IQR) 20 (11–39)
High risk outcome(s) stated in the risk of bias table  
 One non-/partially reported outcome clearly specified 27 (27)
 More than one non-/partially reported outcome clearly specified 52 (52)
 No outcome specified (e.g. only stated that “Some outcomes were not reported”) 21 (21)
Reviewer-perceived importance of high-risk outcome(s)  
 At least one was a primary review outcome 52 (66)a
 All were secondary review outcomes 27 (34)a
Synthesis of high-risk outcome(s)  
 At least one outcome was synthesised in a meta-analysis (based on data from studies that completely reported the outcome) 51 (65)a
 All outcome(s) were synthesised/summarised narratively 28 (35)a
Location in the review where the synthesis of at least one high-risk outcome was reported  
 Main text 79 (100)a
 Abstract 63 (80)a
 Summary of Findings table 47 (59)a
  1. aThe denominator is 79 because 21 reviews did not specify in the risk of bias table the outcome that was not/partially reported