Table 1 Criteria for assessing the risk of bias in included studies ^a

Eligibility

· Recruited from general HIV-infected population, eg. HIV clinics

· Recruited from moderately selected population, eg. STI clinics

· Highly selected population (eg. FSW, active GUD)

· Unclear selection criteria applied

HSV-2 ascertainment

· High quality type-specific serology assay

· Culture-based diagnosis

· Methods unclear

· PCR assay

· Clinical diagnosis

Endpoint ascertainment

· Similar between groups

· Similar between groups

· Methods unclear

· Regular timing

· Irregular timing

Confounding

· Accounted for ART, and

· Some of these confounders accounted for

· None of these confounders accounted for or unclear

· Accounted for acyclovir, and

· Accounted for CD4/stage of HIV disease

Analysis

· Sample size or power calculation done, and

· No concerns with analysis

· Problems identified with analysis

· No concerns with analysis

Attrition

· Minimal attrition (<10%) and

· Moderate attrition (10-20%)

· High attrition (>20%)

· Attrition explained

· Attrition not explained

OVERALL

· Most items at low risk of bias, including both HSV-2 ascertainment and confounding

· Most items at low to moderate risk of bias

· Most items at moderate risk of bias

· Not more than two items at moderate risk of bias

· No item at high risk of bias

· At least one item at high risk of bias