Intermediary-facilitated TB reporting lifted notifications across 15 Vietnamese provinces
A large public-private mix strategy in Viet Nam was associated with a measurable rise in tuberculosis notifications, and the gain was not confined to a narrow pilot zone. The intervention linked national tuberculosis services with 3,154 non-program providers across 15 provinces, covering 40.6 million people, or about 40% of the country’s population. In a disease where missed reporting keeps transmission chains alive and delays treatment, the size of the catchment matters as much as the size of the effect.
The model did not rely on goodwill alone. It used an intermediary agency to connect private and other non-national program providers to the public TB system through screening, referral, reporting, provider incentives, patient support, and integrated data systems. Those moving parts are important because TB care often fragments at the very point where symptoms first reach the health system: pharmacies, private clinics, hospitals, and other settings outside routine national surveillance. When those providers are left outside the reporting structure, patients can still receive care, but the public health system loses sight of them.
Population-level gains after staggered rollout
The evaluation used quarterly provincial notification data from 2016 through 2025 and applied a heterogeneity-robust extended two-way fixed-effects framework to estimate the average treatment effect on the treated. It also adjusted for province and calendar-quarter fixed effects and for COVID-19 disruptions. That analytic approach matters because the intervention was rolled out in stages and with different intensity across provinces, making a simple before-and-after comparison vulnerable to distortion from secular trends and pandemic-era disruption.
The main result was a gain of 316.8 additional notifications per province-quarter, with a 95% confidence interval of 23.4 to 610.2 and a _p_ value of 0.034. The effect was not just a one-time burst. Dose-response models showed that more intensive implementation was associated with more notifications, including a 1.4% increase in notifications per 100 verbal assessments per 100,000 population. In practical terms, the intervention behaved like a functioning surveillance-and-referral engine rather than a static administrative program.
One sentence captures the central finding: the more actively the system searched, the more TB it found.
Why an intermediary likely changed the curve
The biological rationale is straightforward even if the delivery model is operationally complex. TB control depends on early identification, microbiologic confirmation, linkage to treatment, and notification into the national system. Every step weakens when private providers work in isolation, especially in settings where people first seek care outside the public sector and where diagnostic pathways vary widely. An intermediary can reduce friction by standardizing referral behavior, aligning incentives, and making reporting easier than non-reporting.
This study’s architecture reflects that logic. Providers were engaged through systematic screening and reporting mechanisms, and the model emphasized active case finding alongside patient support. Such a design is well suited to closing the gap between clinical encounter and national notification, particularly when a large share of initial care-seeking occurs outside the program network. In Viet Nam, that gap is not theoretical; it is built into the structure of care delivery.
National models showed the same pattern
The provincial analysis was not the only signal. National models suggested 3,206 additional notifications annually, with a 95% confidence interval of 1,658 to 4,763 and a _p_ value below 0.001. The intensity-associated increase in notification was 3.5%, again with a highly significant association. An interrupted time-series analysis of national notifications pointed in the same direction, and repeated analyses using PPM notifications as the outcome were concordant. Multiple methods, multiple timeframes, and multiple outcome definitions all converged on the same basic conclusion.
This matters because TB notification is not just a bookkeeping exercise. Notifications are the gateway to treatment tracking, drug-susceptibility testing, public health follow-up, and eventual interruption of transmission. When notification rises in a sustainable way, the system is identifying people who were previously invisible to national control efforts.
Cost per additional notification stayed within a modest range
The intervention’s funder-perspective costs per additional notification were US$166 to US$458. In an environment of tightening global health budgets, that range is likely to interest policymakers as much as the clinical effect itself. A program that moves thousands of cases into the national system without requiring a wholesale redesign of the healthcare market has a different budgetary profile than one that depends on building parallel infrastructure.
The intervention’s value proposition rests on leverage. Rather than replacing the existing provider network, it recruited the network already being used by patients and gave it a pathway into the national TB response. That is a more realistic model for fragmented health systems, where the public sector often cannot force patients to change where they first seek care, but can change how those encounters are captured and linked to treatment.
Non-randomized design leaves room for caution
The main limitation was the non-randomized design. High-burden provinces were intentionally selected for intervention, which can inflate apparent impact if those provinces were already trending differently from comparison areas. The analysis attempted to blunt that risk through conservative control selection and triangulation across methods, including the national time-series and repeated outcome analyses. Even so, the design cannot fully erase selection bias.
The study also concentrated on notification, not on every downstream TB outcome. Notification is essential, but it is still a surrogate for the full chain of diagnosis, treatment initiation, completion, and cure. The available data support a strong association with reporting gains; they do not by themselves prove every notified patient completed care.
What Viet Nam’s experience suggests for TB control elsewhere
The pattern seen here fits a wider TB control problem: treatment coverage gaps persist because many patients enter care outside the national program, and those encounters are often invisible to surveillance systems. An intermediary-facilitated public-private mix can bridge that divide at scale, with enough structure to influence both behavior and reporting volume. In a setting where 45% to 75% of initial health-seeking occurs at non-program providers, the intervention addresses a point of failure that conventional public-sector strategies often miss.
For countries facing similar fragmentation, the most useful signal from Viet Nam is not merely that public-private collaboration can work. It is that the collaboration can be operationalized through an intermediary, expanded from pilots to a large provincial network, and linked to measurable gains in notifications at both local and national levels. The remaining question is not whether private providers matter; it is how quickly national TB programs can build the mechanisms that make those encounters count.