BACKGROUND: Poor linkage to HIV care is impeding achievement of the Joint United Nations Programme on HIV and AIDS (UNAIDS) 90-90-90 targets. This study aims to identify risk factors for poor linkage-to-care after HIV counseling and testing, thereby informing strategies to achieve 90-90-90. SETTING: The Thol'impilo trial was a large randomized controlled trial performed between 2012 and 2015 in South Africa, comparing different strategies to improve linkage-to-care among adults aged ≥18 years who tested HIV-positive at mobile clinic HIV counseling and testing. METHODS: In this secondary analysis, sociodemographic factors associated with time to linkage-to-care were identified using Cox regression. RESULTS: Of 2398 participants, 61% were female, with median age 33 years (interquartile range: 27-41) and median CD4 count 427 cells/mm (interquartile range: 287-595). One thousand one hundred one participants (46%) had clinic verified linkage-to-care within 365 days of testing HIV-positive. In adjusted analysis, younger age [≤30 vs >40 years: adjusted hazard ratio (aHR): 0.58, 95% CI: 0.50 to 0.68; 31-40 vs >40 years: aHR: 0.81, 95% CI: 0.70 to 0.94, test for trend P < 0.001], being male (aHR: 0.86, 95% CI: 0.76 to 0.98, P = 0.028), not being South African (aHR: 0.79, 95% CI: 0.66 to 0.96, P = 0.014), urban district (aHR: 0.82, 95% CI: 0.73 to 0.93, P = 0.002), being employed (aHR: 0.81, 95% CI: 0.72 to 0.92, P = 0.001), nondisclosure of HIV (aHR: 0.63, 95% CI: 0.56 to 0.72, P < 0.001), and having higher CD4 counts (test for trend P < 0.001) were all associated with decreased hazard of linkage-to-care. CONCLUSION: Linkage-to-care was low in this relatively large cohort. Increasing linkage-to-care requires innovative, evidence-based interventions particularly targeting individuals who are younger, male, immigrant, urban, employed, and reluctant to disclose their HIV status.