S147 Yield and cost of screening for active and latent tuberculosis among high-risk groups attending london emergency departments

Abstract

Background Emergency Department (ED) presentations could present a valuable opportunity to identify high-risk individuals with latent tuberculosis infection (LTBI) and active tuberculosis (TB) disease. We aimed to investigate the prevalence and cost-effectiveness of screening for LTBI and pulmonary TB disease among migrants and under-served groups attending London EDs. Methods We recruited adults attending EDs (18/7/2013–25/5/2017) who were either recent entrants from high TB-incidence countries; or had a history of homelessness, imprisonment, or substance use disorder. We assessed the yield of screening for LTBI (using an interferon-gamma release assay; IGRA) and active pulmonary TB (with a symptom screen, followed by chest radiograph and sputum Xpert MTB/RIF if positive). A cost-effectiveness analysis was performed by calculating cost per true-positive LTBI test. Results A total of 1420 participants were screened for LTBI. Median age was 45 years (IQR 34–60), and the majority were born outside the UK (1,162/1,420; 81.8%). A total of 241/1,420 participants (17.0%) had a history of either substance use disorder, homelessness or imprisonment. Of those with available IGRA results, 214/1,246 (17.2%) were positive. Male sex, age >35 years, non-UK country of birth, and previous TB contact were independent risk factors for being IGRA-positive. Of the 214 with a positive IGRA, 120 (56%) were followed-up by health record review for evidence of progression to TB disease; 5/120 (4.2%) were diagnosed with TB disease (median interval to TB 162 days (IQR 55–108.5)). In the economic analysis, the average cost per patient recruited of correctly diagnosing an LTBI case was £834 (compared to an estimated £1007 per LTBI case diagnosed through primary care). Only 14/513 (2.7%) participants screened for active TB had a positive symptom screen so were eligible for further testing. No cases of active TB were diagnosed. Conclusions Targeted LTBI screening in EDs could be considered to complement the primary care LTBI screening programme. In contrast, active TB screening in EDs using our symptom screen-led algorithm should not be pursued as the yield is likely to be minimal. Other active TB screening models (e.g. routine, automated reviews of ED chest radiographs) could be considered as a focus of future research.