Background and justification
The End TB Strategy adopted by the World Health Assembly in 2014 aims to end the global tuberculosis (TB) epidemic by 2035 and to reduce TB deaths by 95%. The operationalisation of this strategy calls for innovative approaches and health technologies for TB prevention and care in the most vulnerable and most affected populations. Furthermore, bringing together researchers, end-users, decision-makers and program implementers is crucial to shorten time between the generation of scientific evidence and widespread adoption of policies and improved practices for childhood TB, that can contribute to reaching the Sustainable Development Goals (SDGs).
Each year an estimated 1.2 million children and young adolescents (<15 years) develop TB [2], of which more than 50% are aged below 5 years [3]. Modelling suggests that nearly 250,000 children die from TB yearly, and that over 95% of those dying of TB are undiagnosed. Globally, only 44% of children with TB are notified to the World Health Organization (WHO) by National TB Programs (NTPs), mainly because they are not diagnosed and consequently not treated. In children below 5 years this figure is as low as 35%. Low case detection is largely due to the paucibacillary nature of the disease and challenges in respiratory sample collection contributing to the low microbiological yield in children. Once diagnosed and treated, outcomes for children with TB are generally excellent with <1% mortality, but the case fatality rate for untreated TB can reach 44% in children below 5 years [6]. Children with inadequate immunity [e.g. children living with HIV (CLHIV), or those with severe acute malnutrition (SAM)] or severe pneumonia are at higher risk of underdiagnosis and of dying from TB.
WHO has identified that improvement in TB diagnostics is the highest research priority in the field of child TB. Currently, in the absence of highly sensitive TB diagnostic tool for children, most children are started on treatment only on the basis of high clinical suspicion. Treatment decision algorithms (TDAs), that assign scores to clinical and radiographic features or microbiological tests and recommend TB treatment initiation above a pre-defined total score, can enable rapid and uniform treatment decision-making. In March 2022, WHO issued an interim and conditional overall recommendation to use TDAs to diagnose pulmonary TB in children below 10 years. Furthermore, in the accompanying operational handbook [12], WHO suggested two specific TDAs for use in settings with and without access to chest X-ray (CXR), with a single diagnostic approach in both the general paediatric population and high-risk groups (age <2, CLHIV, and SAM). WHO has stated that external validation of these algorithms is an urgent priority in view of the conditional recommendation[1].
Decentralizing childhood TB services is essential to increase access to TB diagnosis. Use of TDAs at lower levels of care by overburdened Health Care Workers (HCWs) with limited child TB experience will require strengthening clinical skills and treatment decision-making capacity. Data on the diagnostic accuracy of TDAs, their feasibility, acceptability by end-users, effectiveness, and cost-effectiveness are crucial to update the current WHO policy and operational handbook, national policies, and clinical curricula. A comprehensive TB TDA-based approach could integrate other specific TDAs developed for CLHIV and those with SAM if they outperform the WHO-suggested TDAs and would also provide the opportunity to integrate a disease severity assessment step to assess eligibility for a shorter (4-month) treatment for non-severe TB cases in children as recommended by the WHO. Importantly, tools for integrated TDAs should be tailored to the needs and existing clinical practices of HCWs at primary health centre (PHC) and district hospital (DH) level, to which innovative digital tools (such as clinical decision support systems – CDSS) could contribute, in order to enhance adoption, delivery, and quality of decentralized TB services.
