SUMMARY Objective To systematically review Indian literature on delays in TB diagnosis and treatment. Methods We searched multiple sources for studies on delays in pulmonary TB and chest symptomatic patients. Studies were included if numeric data on any delay were reported. Patient delay was defined as the time interval between onset of symptoms and the patient’s first contact with a healthcare provider. Diagnostic delay was defined as the time interval between the first consultation with a healthcare provider and diagnosis. Treatment delay was defined as the time interval between diagnosis and initiation of anti-TB treatment. Total delay was defined as time interval from the onset of symptoms until treatment initiation. Results Among 541 potential citations identified, 23 studies met our inclusion criteria. Included studies used a variety of definitions for onset of symptoms and delays. Median (IQR) estimates of patient, diagnostic and treatment delay were 18.4 (14.3-27.0), 31.0 (24.5-35.4) and 2.5 days (1.9-3.6), respectively, for TB and chest symptomatic patients combined. The median total delay was 55.3 days (46.5-61.5). About 48% of all patients first consulted private providers and 2.7 healthcare providers, on average, were consulted before diagnosis. Number and type of provider first consulted were the most important risk factors for delay. Conclusions These findings underscore the need to develop novel strategies for reducing patient and diagnostic delays and engaging first-contact healthcare providers.

Deep learning (DL) neural networks have only recently been employed to interpret chest radiography (CXR) to screen and triage people for pulmonary tuberculosis (TB). No published studies have compared multiple DL systems and populations. We conducted a retrospective evaluation of three DL systems (CAD4TB, Lunit INSIGHT, and qXR) for detecting TB-associated abnormalities in chest radiographs from outpatients in Nepal and Cameroon. All 1196 individuals received a Xpert MTB/RIF assay and a CXR read by two groups of radiologists and the DL systems. Xpert was used as the reference standard. The area under the curve of the three systems was similar: Lunit (0.94, 95% CI: 0.93–0.96), qXR (0.94, 95% CI: 0.92–0.97) and CAD4TB (0.92, 95% CI: 0.90–0.95). When matching the sensitivity of the radiologists, the specificities of the DL systems were significantly higher except for one. Using DL systems to read CXRs could reduce the number of Xpert MTB/RIF tests needed by 66% while maintaining sensitivity at 95% or better. Using a universal cutoff score resulted different performance in each site, highlighting the need to select scores based on the population screened. These DL systems should be considered by TB programs where human resources are constrained, and automated technology is available.

An aggregate data meta‐analysis is a statistical method that pools the summary statistics of several selected studies to estimate the outcome of interest. When considering a continuous outcome, typically each study must report the same measure of the outcome variable and its spread (eg, the sample mean and its standard error). However, some studies may instead report the median along with various measures of spread. Recently, the task of incorporating medians in meta‐analysis has been achieved by estimating the sample mean and its standard error from each study that reports a median in order to meta‐analyze the means. In this paper, we propose two alternative approaches to meta‐analyze data that instead rely on medians. We systematically compare these approaches via simulation study to each other and to methods that transform the study‐specific medians and spread into sample means and their standard errors. We demonstrate that the proposed median‐based approaches perform better than the transformation‐based approaches, especially when applied to skewed data and data with high inter‐study variance. Finally, we illustrate these approaches in a meta‐analysis of patient delay in tuberculosis diagnosis.

Background Artificial intelligence (AI) algorithms can be trained to recognise tuberculosis-related abnormalities on chest radiographs. Various AI algorithms are available commercially, yet there is little impartial evidence on how their performance compares with each other and with radiologists. We aimed to evaluate five commercial AI algorithms for triaging tuberculosis using a large dataset that had not previously been used to train any AI algorithms.Methods Individuals aged 15 years or older presenting or referred to three tuberculosis screening centres in Dhaka, Bangladesh, between May 15, 2014, and Oct 4, 2016, were recruited consecutively. Every participant was verbally screened for symptoms and received a digital posterior-anterior chest x-ray and an Xpert MTB/RIF (Xpert) test. All chest x-rays were read independently by a group of three registered radiologists and five commercial AI algorithms: CAD4TB (version 7), InferRead DR (version 2), Lunit INSIGHT CXR (version 4.9.0), JF CXR-1 (version 2), and qXR (version 3). We compared the performance of the AI algorithms with each other, with the radiologists, and with the WHO's Target Product Profile (TPP) of triage tests (≥90% sensitivity and ≥70% specificity). We used a new evaluation framework that simultaneously evaluates sensitivity, proportion of Xpert tests avoided, and number needed to test to inform implementers' choice of software and selection of threshold abnormality scores. Findings Chest x-rays from 23 954 individuals were included in the analysis. All five AI algorithms significantly outperformed the radiologists. The areas under the receiver operating characteristic curve were 90•81% (95% CI 90•33-91•29) for qXR, 90•34% (89•81-90•87) for CAD4TB, 88•61% (88•03-89•20) for Lunit INSIGHT CXR, 84•90% (84•27-85•54) for InferRead DR, and 84•89% (84•26-85•53) for JF CXR-1. Only qXR (74•3% specificity [95% CI 73•3-74•9]) and CAD4TB (72•9% specificity [72•3-73•5]) met the TPP at 90% sensitivity. All five AI algorithms reduced the number of Xpert tests required by 50% while maintaining a sensitivity above 90%. All AI algorithms performed worse among older age groups (>60 years) and people with a history of tuberculosis.Interpretation AI algorithms can be highly accurate and useful triage tools for tuberculosis detection in high-burden regions, and outperform human readers.Funding Government of Canada.

Computer-aided reading (CAR) of medical images is becoming increasingly common, but few studies exist for CAR in tuberculosis (TB). We designed a prospective study evaluating CAR for chest radiography (CXR) as a triage tool before Xpert MTB/RIF (Xpert).Consecutively enrolled adults in Dhaka, Bangladesh, with TB symptoms received CXR and Xpert. Each image was scored by CAR and graded by a radiologist. We compared CAR with the radiologist for sensitivity and specificity, area under the receiver operating characteristic curve (AUC), and calculated the potential Xpert tests saved.A total of 18 036 individuals were enrolled. TB prevalence by Xpert was 15%. The radiologist graded 49% of CXRs as abnormal, resulting in 91% sensitivity and 58% specificity. At a similar sensitivity, CAR had a lower specificity (41%), saving fewer (36%) Xpert tests. The AUC for CAR was 0.74 (95% CI 0.73–0.75). CAR performance declined with increasing age. The radiologist grading was superior across all sub-analyses.Using CAR can save Xpert tests, but the radiologist's specificity was superior. Differentiated CAR thresholds may be required for different populations. Access to, and costs of, human readers must be considered when deciding to use CAR software. More studies are needed to evaluate CAR using different screening approaches.