BACKGROUND: Of the estimated 10 million people affected by (TB) each year, one-third are never diagnosed. Delayed case detection within the private healthcare sector has been identified as a particular problem in some settings, leading to considerable morbidity, mortality and community transmission. Using unannounced standardised patient (SP) visits to the pharmacies, we aimed to evaluate the performance of private pharmacies in the detection and treatment of TB. METHODS: A cross-sectional study was undertaken at randomly selected private pharmacies within 40 districts of Vietnam. Trained actors implemented two standardised clinical scenarios of presumptive TB and presumptive multidrug-resistant TB (MDR-TB). Outcomes were the proportion of SPs referred for medical assessment and the proportion inappropriately receiving broad-spectrum antibiotics. Logistic regression evaluated predictors of SPs' referral. RESULTS: In total, 638 SP encounters were conducted, of which only 155 (24.3%) were referred for medical assessment; 511 (80·1%) were inappropriately offered antibiotics. A higher proportion of SPs were referred without having been given antibiotics if they had presumptive MDR-TB (68/320, 21.3%) versus presumptive TB (17/318, 5.3%; adjusted OR=4.8, 95% CI 2.9 to 7.8). Pharmacies offered antibiotics without a prescription to 89.9% of SPs with presumptive TB and 70.3% with presumptive MDR-TB, with no clear follow-up plan. CONCLUSIONS: Few SPs with presumptive TB were appropriately referred for medical assessment by private pharmacies. Interventions to improve appropriate TB referral within the private pharmacy sector are urgently required to reduce the number of undiagnosed TB cases in Vietnam and similar high-prevalence settings.