Reducing spatial clustering to prevent tuberculosis transmission in a busy Zambian hospital: a modelling study based on person movements, environmental and clinical data

BackgroundHospitals in high-burden tuberculosis (TB) settings are important sites of Mycobacterium tuberculosis (Mtb) transmission, yet the impact of infection prevention and control (IPC) measures targeting crowding is poorly understood. We assessed the effects of simple interventions to reduce spatial clustering and airborne transmission in a Zambian hospital. Methods and findingsFrom June to August 2024, we prospectively collected clinical data on presumptive (symptom-based screening) and confirmed (Xpert-positive, chest X-ray) TB patients, indoor CO2 levels, bioaerosol samples, and continuous person movements (optical sensors) in the hospitals main waiting hall. Airborne Mtb DNA was detected using hourly cyclonic bioaerosol sampling. Using a spatiotemporal Wells-Riley model integrating ventilation, proximity between visitors, and movement patterns, we estimated Mtb transmission risk under routine conditions and during two IPC interventions: (1) an optimised waiting-area layout with physical distancing measures; and (2) an added one-way patient flow system. During 52 days, 668 presumptive and 45 confirmed TB patients visited the hospital, and 671 840 person movements were recorded. Despite excellent natural ventilation (median CO2 455 ppm; 8.0 air changes per hour), airborne Mtb was detected on six days. The first intervention reduced spatial clustering by an estimated 24% (95% credible interval [CrI] 13-32) and the second by 13% (95% CrI 1-23). The interventions lowered Mtb transmission risk by an estimated 39% (95% CrI 29-48) and 21% (95% CrI 9-32), respectively. Over the four weeks of implementation, they collectively averted an estimated 16 (95% CrI 8-26) infections. ConclusionsIn this high-burden, well-ventilated hospital, short-range exposure remained a driver of airborne Mtb transmission. Simple, low-cost operational changes to reduce clustering substantially decreased proximity-driven transmission risk. Integrating proximity-focused strategies can meaningfully strengthen IPC in resource-limited healthcare settings.