Treatment coverage gaps remain a major barrier to ending tuberculosis. Many people with TB seek care from private and other non-program providers who are not routinely linked to national reporting systems. Public–private mix (PPM) strategies aim to engage these providers to increase case finding and notifications. Intermediary agencies have been proposed as operational mechanisms to bridge national TB programs (NTPs) and out-of-network providers, but robust population-level evidence has been scarce.
This study evaluated an intermediary-facilitated PPM implemented at scale in Viet Nam to measure its association with TB notifications and to assess dose–response effects and cost per additional notification.
Between 2020 and 2023 an intermediary agency implemented a PPM model across 15 provinces in Viet Nam, covering approximately 40.6 million people, equivalent to about 40% of the national population. The intervention engaged 3,154 non-NTP providers including private clinics, pharmacies, and hospitals.
Core components included systematic screening and referral pathways, mechanisms for routine reporting into NTP systems, active case finding, provider incentives, patient support, and integration of data systems to facilitate notification. The program emphasized generating PPM notifications as an intermediate outcome and raising overall provincial and national TB notifications as the primary public health outcome.
The evaluation used a non-randomized, population-level quasi-experimental design. Quarter-level province notification data from 2016 through 2025 were analyzed to estimate the association between the intervention and TB notifications. The primary analytic framework was an extended two-way fixed-effects model robust to heterogeneity from staggered intervention deployment and variable intervention intensity. Models included province and calendar-quarter fixed effects and adjustments for COVID-19 disruptions.
Poisson pseudo-maximum-likelihood fixed-effects models were used to estimate average treatment effects on the treated and dose–response relationships. The authors triangulated findings using interrupted time-series analysis of national notifications and repeated analyses using PPM notifications as an intermediate outcome.
The primary outcome was quarterly province-level TB notifications reported to the NTP. The intermediate outcome was PPM notifications originating from the engaged non-NTP providers. The analysis quantified both the average change associated with the intervention and incremental effects linked to intervention intensity (e.g., numbers of verbal assessments).
Using the quasi-experimental fixed-effects approach, the intervention was associated with an average increase of 316.8 additional notifications per province-quarter (95% confidence interval [CI] 23.4 to 610.2; p = 0.034).
Dose–response analysis indicated that for every 100 verbal assessments per 100,000 population there was a 1.4% increase in notifications (incidence rate ratio [IRR] = 1.00014; 95% CI 1.00002 to 1.00025; p = 0.018).
At the national level, interrupted time-series models suggested approximately 3,206 additional notifications annually associated with the intervention (95% CI 1,658 to 4,763; p < 0.001), and a 3.5% intensity-associated notification increase (IRR = 1.00035; 95% CI 1.00018 to 1.00052; p < 0.001).
Analyses that used PPM notifications as an intermediate outcome produced concordant results with the primary models, supporting a consistent association between intermediary-facilitated PPM activities and raised notifications.
From a funder perspective the authors estimated marginal costs per additional notification in the range of US$166 to US$458. The study presents these estimates to inform affordability and budgetary planning for scaling intermediary-facilitated PPM approaches under constrained global health financing.
The principal limitation reported was the non-randomized design: intervention provinces were purposively selected and tended to be high-burden areas. This selection could overstate the intervention’s effect if treated provinces were on trajectories different from controls prior to the program. The authors mitigated this concern through conservative selection of control areas, use of analytic methods robust to heterogeneity and staggered rollouts, adjustment for COVID-19 effects, and triangulation across methods and outcomes. Details about other potential unmeasured confounders or implementation fidelity over time were not reported in the source beyond the described mitigation strategies.
The study provides population-level, methodologically rigorous evidence that intermediary-facilitated public–private mix interventions can be associated with significant, dose-responsive increases in both PPM notifications and overall TB notifications at provincial and national scales.
Given the observed effects and the reported cost per additional notification, intermediary agency models are presented as an affordable and scalable mechanism to operationalize PPM and to close TB treatment coverage gaps, particularly in settings with fragmented care and constrained funding. The authors recommend considering intermediary-facilitated PPM as a practical strategy to sustain progress toward ending TB, noting the need to interpret effect sizes in light of the study’s non-randomized design.