Start4Kids: Evaluating the performance, feasibility, acceptability and cost effectiveness of computer-aided detection (CAD) of chest X-ray (CXR) and concurrent testing using novel diagnostics and non-sputum specimens for childhood tuberculosis (TB) in Bangladesh, Cameroon, Kenya and Viet Nam.

Each year an estimated 1.25 million children and adolescents under 15 years old become sick due to tuberculosis (TB). Despite TB being preventable and curable, in 2023, approximately 191 000 children died from the disease (Global Tuberculosis Report 2024). While treatment is both effective and tolerable, case detection continues to challenge many national TB programmes (NTPs), particularly in children under 5 years old. Less than half of children and young adolescents with TB are diagnosed and notified each year. Diagnostic gaps The reasons for this large gap in case detection are multiple. In children, obtaining sputum is particularly challenging as they rarely expectorate, and invasive collection procedures such as gastric aspiration or induced sputum are often required. This limits the feasibility of routine bacteriological testing, especially within lower levels of care. When sputum can be obtained, specimen collection, particularly in younger children, requires trained and experienced staff. Bacteriological confirmation is less frequent due to the paucibacillary nature of disease in children and the resultant reduction in sensitivity when using available diagnostic tests. To address these limitations, research studies and programmes often apply the NIH clinical case definition as a reference standard, recognising the challenges of achieving microbiological confirmation in this age group. However, this approach also has limitations, as chest X-ray, a key component of the definition, has inherent limitations in diagnostic accuracy, even when interpreted under standardised conditions by multiple experts and particularly so in small children. Further, clinical syndromes are often non-specific and can be rapidly progressive, resulting in frequent misdiagnosis and associated early mortality.

To overcome some of these barriers, health systems often require referral of children with signs, symptoms, or screening investigations consistent with TB ("presumptive TB") to secondary or tertiary health centres so that expert staff can be consulted, and relevant specimens, or further specimens, taken. This may result in additional cost for both the child and their family and the health system, cause further delays in diagnosis and treatment initiation, and may ultimately result in greater morbidity and mortality.

To reduce the case detection gap and to avoid diagnostic delays, policymakers and programme implementers have promoted child and adolescent TB assessment at primary healthcare (PHC) or child health centres, often through integration into existing child services diagnosing similar conditions. While child TB case notifications prior to the COVID-19 pandemic had been increasing, they declined sharply between 2019 and 2020 due to COVID. Childhood notifications recovered more slowly compared to adults in 2021, and efforts continue to be hampered by limited PHC staff expertise, poor access to diagnostic tests including chest X-ray, and a lack of dedicated time for over-burdened health care staff.

Inclusion Criteria:

1. Age 14 and below on the day of assessment
2. Fulfils the criteria of presumptive TB according to the definition 10.1 of the protocol, or TB associated risk factors according to the definition 10.2. of the protocol
3. Diagnosed with TB disease

Exclusion Criteria:

1. On treatment for TB disease at the time of evaluation
2. Parent or guardian does not allow the child to participate in the study or declines to sign the ICF.
3. The child does not wish to participate in the study or declines to sign the Assent Form (where applicable)

Anonymised IPD will be sared with partner country researchers, WHO, UNITAID. Dataset will be available from LSTM according to their data storage and sharing conditions.