Table 5: GRADE (Grading of Recommendations Assessment, Development and Evaluation)^32,37-39 assessment of the quality of the body of evidence regarding the accuracy of antibody tests compared to RNA detection in screening for chronic hepatitis C virus infection in 1 cross-sectional study (cohort-type accuracy study)

Outcome	Author, yr; no. of participants	Quality assessment domain*					Pretest probability, %†; expected no. per 1000 patients tested (95% CI)			Overall quality‡
Outcome	Author, yr; no. of participants	Risk of bias	Indirectness	Inconsistency	Imprecision	Publication bias	0.96	0.61	1.34	Overall quality‡
True positive	Kosan et al.,²⁶ 2010 11	++++§	++¶	+++**	++††	None	8 (6-10)	5 (4-6)	11 (8 -13)	Very low
False negative	Kosan et al.,²⁶ 2010 11	++++§	++¶	+++**	++††	None	2 (0-4)	1 (0-2)	2 (0-5)	-
True negative	Kosan et al.,²⁶ 2010 17 840	++++§	++¶	+++**	++++‡‡	None	987 (986-988)	991 (990-992)	984 (983-985)	Very low
False positive	Kosan et al.,²⁶ 2010 17 840	++++§	++¶	+++**	++++‡‡	None	3 (2-4)	3 (2-4)	3 (2-4)	-

Note: CI = confidence interval.

*The overall quality is determined by downgrading, from a rating of "high" (4 plus signs), for any concerns as follows: 0 (no concerns), -1 (serious concerns) or -2 (very serious concerns).

†The values correspond to the HCV seroprevalence estimate in the general Canadian population of 0.96% (95% CI 0.61%-1.34%).²

‡The 4 possible ratings for overall quality are high, moderate, low and very low. We had serious or very serious concerns regarding several of the domains, resulting in an overall rating of "very low."

§Based on the Quality Assessment of Diagnostic Accuracy Studies version 2 tool,³⁶ the risk of bias was assessed to be low (Appendix 4).

¶This domain was downgraded by 2 points because 1) the study population was different from that in our research question (Turkish volunteers who underwent a mandatory physical examination before donating blood versus the general Canadian population) and 2) the study answered an indirect question about the accuracy of diagnostic testing rather than directly assessing the effectiveness of screening the general population for HCV infection.

**This domain was downgraded by 1 point because we expect that other studies may have different estimates.

††This domain was downgraded by 2 points because the very low number of true-positive and false-negative results led to a very wide 95% CI for the sensitivity estimate.

‡‡This domain was not downgraded because the large number of true-negative results led to a very narrow 95% CI for the specificity estimate.