Objectives: To assess the extent of nondisclosure of known HIV status among sexual health clinic attendees and to quantify the impact of nondisclosure on estimates of undiagnosed HIV prevalence and of the proportion of patients remaining undiagnosed on leaving the clinic.
Methods: Serum samples from the unlinked anonymous survey of clinic attendees' archive were tested for antiretrovirals. Estimates of undiagnosed HIV were adjusted using the findings.
Results: Antiretrovirals were detected in 27% of samples taken from 'previously undiagnosed' attendees, who did not have an HIV test but were HIV positive as detected by unlinked anonymous testing, indicating nondisclosure; 24% of such samples from MSM had antiretrovirals present compared with 32% of heterosexual men and women. Antiretrovirals were detected in 33% of samples from London clinics and in 21% from non-London clinics. Following adjustment, the estimated prevalence of undiagnosed HIV decreased nonsignificantly from 3.04% (95% confidence interval 2.71-3.41) to 2.66% (2.35-3.01) among men who have sex with men (MSM), 0.31% (0.26-0.37) to 0.30% (0.25-0.36) in heterosexual men and 0.40% (0.35-0.46) to 0.37% (0.32-0.43) in women; 7% of MSM who do not have an HIV test at a clinic visit will be infected with HIV and remain unaware of their infection.
Conclusion: Nondisclosure of HIV status to healthcare professionals occurs among clinic attendees. Adjustment for nondisclosure results in a small, nonsignificant decrease in the prevalence of undiagnosed HIV estimated from the unlinked anonymous survey in sexual health clinics. Testing the population of MSM not having an HIV test remains a priority as levels of undiagnosed HIV are high.