Tuberculosis attributed to transmission within healthcare facilities, Botswana-The Kopanyo Study

Abstract

Objective: Healthcare facilities are a well-known high-risk environment for transmission of M. tuberculosis, the etiologic agent of tuberculosis (TB) disease. However, the link between M. tuberculosis transmission in healthcare facilities and its role in the general TB epidemic is unknown. We estimated the proportion of overall TB transmission in the general population attributable to healthcare facilities.

Methods: We combined data from a prospective, population-based molecular epidemiologic study with a universal electronic medical record (EMR) covering all healthcare facilities in Botswana to identify biologically plausible transmission events occurring at the healthcare facility. Patients with M. tuberculosis isolates of the same genotype visiting the same facility concurrently were considered an overlapping event. We then used TB diagnosis and treatment data to categorize overlapping events into biologically plausible definitions. We calculated the proportion of overall TB cases in the cohort that could be attributable to healthcare facilities.

Results: In total, 1,881 participants had TB genotypic and EMR data suitable for analysis, resulting in 46,853 clinical encounters at 338 healthcare facilities. We identified 326 unique overlapping events involving 370 individual patients; 91 (5%) had biologic plausibility for transmission occurring at a healthcare facility. A sensitivity analysis estimated that 3%-8% of transmission may be attributable to healthcare facilities.

Conclusions: Although effective interventions are critical in reducing individual risk for healthcare workers and patients at healthcare facilities, our findings suggest that development of targeted interventions aimed at community transmission may have a larger impact in reducing TB.

Keywords: Tuberculosis; healthcare facilities; nosocomial transmission; transmission.

Fig. 1.

Visualization of source-secondary event. Black indicates either inpatient stays (rectangles) or outpatient visits (circles) at the same healthcare facility. The solid red indicates the infectious period of a potential source case (patients A, H, and Q), as defined in the Methods section. Shaded red areas highlight infectious overlap with potential secondary transmission events (red outline). Participants who potentially became infected at the healthcare facility during the overlapping events are shown in bold (patients B, C, D, G, and I). All patients were diagnosed with TB either during (source cases) or after the overlapping event. Note that this illustration highlights a uniquely complex event for illustrative purposes; the median number of patients in a source-secondary event was 2 (interquartile range, 2–3).

Fig. 2.

Percentage of transmission potentially occurring at a healthcare facility. The empirical and resampling estimates of the proportion of transmission potentially occurring at the healthcare facility across all possible year thresholds. Our primary analysis considered any overlapping event where a case was later diagnosed within 2 years (grey dotted box). Grey and black lines represent the interdecile and interquartile range, respectively, of 15,000 pseudopopulations resampled according to the methods; red diamonds indicate the median. Purple triangles indicate the empirical estimates using only data with original genetic profiles.