ACT4KIDS-TB

The challenge

Underdiagnosis of tuberculosis in children is a major global challenge. An estimated 1.25 million children develop TB each year, but less than half get diagnosed and treated. Globally, 96% of children who die from TB did not receive treatment; 80% of them are less than 5 years old.

TB diagnosis in children is highly challenging. Symptoms are shared with multiple other conditions. In addition, sputum samples can be difficult to obtain from young children, and TB disease typically involves a relatively small number of bacteria, making them difficult to detect. Chest X-rays can provide useful diagnostic information, but accurate interpretation and experienced readers are limited in many sub-Saharan African countries.

To guide decision-making, WHO has developed treatment decision algorithms for TB diagnosis, with separate versions depending on the availability of chest X-ray results. Computer-assisted diagnosis (CAD) tools have also been developed to automate analysis of chest X-ray results, but are not yet WHO-approved for children. Consequently, there remains considerable scope to improve TB diagnosis in children, particularly at lower levels of care, which may have fewer specialists but see the bulk of patients.

The project

The ACT4KIDS-TB project is developing an app based on WHO treatment decision algorithms but incorporating additional sources of information to provide a new tool to support the diagnosis of tuberculosis in children.

One advance is to integrate the results of Actiphage analyses of patients’ blood samples into the decision-making tool. Actiphage is a relatively simple laboratory technique in which phages (viruses that infect bacteria) are used to break open Mycobacterium tuberculosis (Mtb, the bacterium that causes tuberculosis), increasing the amount of Mtb DNA that can be detected and, consequently, diagnostic sensitivity, particularly in paucibacillary disease, where conventional methods frequently fail to detect infection. This approach provides results within hours, requires only simple laboratory equipment, and is specific for viable Mtb cells, so it detects only active infections. 

In addition, ways to improve Actiphage analysis are being explored, such as using finger-prick blood samples and collection devices that begin processing blood samples even before they arrive at a laboratory, to speed up the delivery of results. The Actiphage approach has shown promising results in TB detection studies involving adults. ACT4KIDS-TB will be the first to explore its use in children, for whom there is a dire need for sensitive, non-sputum-based point-of-care TB diagnostic tests.

A second potential source of information to be incorporated is CAD interpretation of chest X-ray images. The project will validate the application of this technology for children at the bedside.

A co-creation approach will be used to develop the app, which is designed to accept multiple inputs, calculate the probability of a TB diagnosis, and generate a ‘treat/do not treat’ recommendation. The app will initially be trained on a wide range of data sources from patients with well-characterised TB, with Actiphage and CAD data added when they become available. A large-scale consensus-building exercise will be used to establish a probability threshold above which TB treatment would be recommended.

To evaluate the new app, the project will organise a clinical study in Benin and South Africa, two high-burden countries with differing socioeconomic statuses. Recruiting around 200 children with presumed TB, the study will determine whether use of the app increases the number of children who are correctly diagnosed compared with standard methods of assessment. 

The app will also provide guidance on household contact tracing when TB is likely, and also provide health workers with information on nutritional status, based on patient information provided, including recommendations for action for children with malnutrition. The project will additionally explore issues that might affect the implementation and use of the new app within health systems in sub-Saharan Africa.

Impact

The ACT4KIDS-TB project could help to address a major challenge in tuberculosis control and diagnosis of TB disease in children. It will: 

• Create a simple-to-use tool to advise on clinical decision-making, suitable for use at lower levels of care in resource-limited settings.
• Demonstrate whether Actiphage technology is feasible to introduce and increase the likelihood of TB diagnosis.
• Show whether CAD systems for chest X-rays can improve childhood TB diagnosis at the bedside.

Improved diagnosis of TB in children would lead to more children receiving lifesaving treatment, reducing the unacceptably high mortality rate seen in this age group, and reducing the opportunity for TB to be spread to others due to untreated disease.