IntroductionAwareness of HIV‐infection goes beyond diagnosis, and encompasses understanding, acceptance, disclosure and initiation of the HIV‐care. We aimed to characterize the HIV‐positive population that underwent repeat HIV‐testing without disclosing their serostatus and the impact on estimates of the first UNAIDS 90 target.MethodsThis analysis was nested in a prospective cohort established in southern Mozambique which conducted three HIV‐testing modalities: voluntary counselling and testing (VCT), provider‐initiated counselling and testing (PICT) and home‐based testing (HBT). Participants were given the opportunity to self‐report their status to lay counsellors and HIV‐positive diagnoses were verified for previous enrolment in care. This study included 1955 individuals diagnosed with HIV through VCT/PICT and 11,746 participants of a HBT campaign. Those who did not report their serostatus prior to testing, and were found to have a previous HIV‐diagnosis, were defined as non‐disclosures. Venue‐stratified descriptive analyses were performed and factors associated with non‐disclosure were estimated through log‐binomial regression.ResultsIn the first round of 2500 adults randomized for HBT, 1725 were eligible for testing and 18.7% self‐reported their HIV‐positivity. Of those tested with a positive result, 38.9% were found to be non‐disclosures. Similar prevalence of non‐disclosures was found in clinical‐testing modalities, 29.4% (95% CI 26.7 to 32.3) for PICT strategy and 13.0% (95% CI 10.9 to 15.3) for VCT. Prior history of missed visits (adjusted prevalence ratio (APR) 4.2, 95% CI 2.6 to 6.8), younger age (APR 2.5, 95% CI 1.4 to 4.4) and no prior history of treatment ((APR) 1.4, 95% CI 1.0 to 2.1) were significantly associated with non‐disclosure as compared to patients who self‐reported. When considering non‐disclosures as people living with HIV (PLWHIV) aware of their HIV‐status, the proportion of PLWHIV aware increased from 78.3% (95% CI 74.2 to 81.6) to 86.8% (95% CI 83.4 to 89.6).ConclusionMore than one‐third of individuals testing HIV‐positive did not disclose their previous positive HIV‐diagnosis to counsellors. This proportion varied according to testing modality and age. In the absence of an efficient and non‐anonymous tracking system for HIV‐testers, repeat testing of non‐disclosures leads to wasted resources and may distort programmatic indicators. Developing interventions that ensure appropriate psychosocial support are needed to encourage this population to disclose their status and optimize scarce resources.