Latent Infection Freshmen Tracking for Tuberculosis

The Latent Infection Freshmen Tracking for Tuberculosis (LIFT-TB) study is a prospective, multi-center, real-world implementation cohort designed to evaluate the programmatic feasibility, safety, and long-term effectiveness of school-integrated short-course tuberculosis preventive treatment (TPT) among matriculating freshmen. Embedded within routine campus health services across 26 educational institutions (secondary vocational schools, colleges, and universities) in Meishan City, Sichuan Province, China, the protocol operationalizes proactive case-finding and preventive intervention at student entry.

Screening cascade and baseline stratification Under mandatory institutional biosecurity regulations, all newly enrolled freshmen underwent entry health evaluations comprising tuberculin skin testing (TST with 5 TU PPD-RT23), active symptom screening, and digital chest radiography (CXR). To maximize population coverage (>99%) and minimize non-response bias, a 2-week catch-up window was instituted for students initially absent. To account for widespread neonatal BCG vaccination, baseline infection risk was operationalized using TST induration thresholds.

Control cohort: Students with TST indurations <15 mm (subdivided into TST <5 mm, 5-9 mm, and 10-14 mm).

Target Cohort: Students with strong TST reactivity (≥15 mm or local severe reactions such as blisters/necrosis) in whom active TB was clinically excluded.

Intervention workflows and adherence monitoring Within the target cohort, TPT initiation was managed through shared decision-making. Eligible participants voluntarily initiated one of two WHO-recommended 3-month short-course regimens: Daily fixed-dose combination isoniazid/rifampicin (3HR). Twice-weekly isoniazid plus rifapentine (3H2P2). Treatment administration was embedded within campus health clinics using a hybrid directly observed therapy model combining in-person clinic DOT with smartphone-based video DOT (vDOT). Participants receiving medical observation alone without medication served as the No-TPT comparison group. Adverse events were routinely graded per CTCAE v5.0, with regular safety monitoring to capture potential hepatotoxicity.

Causal inference and emulation framework To complement primary multivariate Cox proportional hazards modeling and address observational confounding inherent in non-randomized treatment assignment, the analytic workflow incorporated a Target Trial Emulation (TTE) framework as sensitivity analyses. Time zero (T0) was anchored to the date of LTBI eligibility confirmation, incorporating a 14-day grace period to align baseline entry and eliminate immortal time bias. High-dimensional baseline covariates, including demographic factors, BMI, school type, medical major, BCG status, prior TB history, and close contact history, were balanced using 1:1 Propensity Score Matching (PSM) and Inverse Probability of Treatment Weighting (IPTW with stabilized weights) to estimate the Average Treatment Effect on the Treated (ATT) across both modified intention-to-treat (mITT) and per-protocol (PP) populations.

Inclusion criteria

• Matriculating freshmen aged 14-24 years enrolled across 26 participating educational institutions
• Completion of baseline tuberculosis screening, including systematic symptom screening, digital chest radiography, and tuberculin skin test (TST)
• Target cohort requirement: TST strong reactivity (induration ≥15 mm or severe local reactions such as blisters, necrosis, or lymphangitis) with active TB ruled out at baseline
• Control cohort requirement: TST induration <15 mm (including negative <5 mm, weakly positive 5-9 mm, and moderately positive 10-14 mm)

Exclusion criteria

• Presence of active TB at baseline screening
• Voluntary receipt of Vaccae vaccination (for the target cohort)