A One-stop Shop for the Same Day Diagnosis and Management of TB and HIV

TB is a major public health problem and the second most common cause of adult death due to infection in many low-income countries. Despite major efforts to de-centralise services, accessibility to diagnosis is still limited, with one third of the 9 million cases occurring each year being missed by national control programmes. New TB diagnostics suitable for use at the point-of-care are emerging. Some of these are intended for screening purposes, as an initial step to identify individuals who may have TB and should undergo further tests for confirmation. These tests may have high sensitivity, but also give false-positive results (low specificity). Other tests aim to be the confirmatory tests for TB (high specificity), but these tests are often more expensive and complex and are only available in hospital laboratories. As these tests have different purposes, it is likely they would work better in combination in a step fashion to optimise their impact and to develop an efficient diagnostic process. Furthermore, as none of the tests is versatile enough to be used in all settings, test combinations will need to consider the health system context in which they would be used. Our aim is to develop and evaluate rapid and accurate diagnostic approaches for TB that facilitate the initiation of appropriate treatment on the same day of the initial consultation in Africa. The objectives are to 1. Evaluate new diagnostics for TB (including among HIV co-infected individuals) that are suitable at the point-of-care; 2. Develop diagnostic algorithms that streamline and accelerate the diagnosis of TB, allowing patients to reach clinical management decisions within a single clinic visit; 3. Determine the impact of using novel point-of-care diagnostic combinations on the proportion of patients correctly initiating TB treatment within 24-48 hours of first attendance; their potential cost effectiveness The investigators conducted studies in 2016-2018 to accomplish the first two objectives and have identified diagnostic tests that are suitable for low and middle income countries. This document therefore refers to objective 3, which aims to 1. Assess the performance of two diagnostic schemes for the diagnosis of TB when compared to culture. 2. Assess the yield of two diagnostic schemes for the diagnosis of TB when compared to Xpert and 3. Assess the cost of the two diagnostic schemes compared to Xpert.

The study will enrol patients in TB clinics based in 4 selected district hospitals (two in Nigeria and two in Ethiopia) and samples will be processed in a single reference laboratory. This diagnostic evaluation trial will comprise two experimental diagnostic schemes which will be compared against the standard of care: * One experimental arm (scheme 1) will screen all patients for HIV using two rapid tests routinely used by the clinics and a rapid CRP. Selected patients will be further tested using ULTRA. Individuals with HIV will undergo an HIV VL using Xpert. * A second experimental arm (scheme 2): will screen individuals for HIV and CRP (as in scheme 1) and selected patients will be tested using Molbio Truenat MTB. Individuals with HIV will undergo an HIV VL using Molbio Truenat HIV-VL and Truenat RIF. * In addition, all patients will be tested using the standard of care consistent of confirmatory HIV and CRP tests, Xpert MTB/RIF and culture. * Randomisation All patients will be randomised at a ratio of 1:1 into schemes 1 and 2. Random numbers will be generated in LSTM by a statistician independent to the study. The scheme allocations will be included in study envelopes assigned to individual study numbers. Equal number of participants will be included in Nigeria and Ethiopia. * Proposed methods for protecting against source bias As this is an open trial, the classification of patients will be based on objective quantitative results of laboratory tests. It is expected the test performances will vary according to HIV status. Participants will be classified according to their experimental test results