. 2014 Sep 16;9(9):e107177. doi: 10.1371/journal.pone.0107177

Table 1. Bias assessment.

Bias domains	Number of trials with low risk of bias	Number of trials with uncertain risk of bias	Number of trials with high risk of bias
Were patients assembled at a common point in the course of their disease?¹	9	17	42
Were the follow up data complete?²	13	55	0
Were outcome criteria either objective or applied in a blind fashion?³	68	0	0
How likely are the outcomes over time?⁴	68	0	0
How precise are the prognostic estimates?⁵	15	30	23

Low risk of bias if all patients are assembled at a common time point in the course of their disease. Unknown risk of bias if relevant information for assessment of bias can not be assembled. High risk of bias if patients are assembled at different time points in the course of their disease.

Low risk of bias if all patients are acounted for and losses to follow up not likely to affect the outcome estimate. Uncertain risk of bias if data on losses to follow up are missing/not accounted for. High risk of bias if losses to follow up are likely to affect outcome estimate.

All trials low risk of bias as HCC and/or mortality are objective outcome measures.

⁴

All trials low risk of bias as the review only included studies with adequate follow up period (>1 year).

⁵

Low risk of bias if standard deviation of follow up <25% of the mean follow up. Uncertain risk of bias if standard deviation of follow up is 25–50% of the mean follow up. High risk of bias is standard deviation of follow up is >50% of the mean follow up.