Project 453056

Canada's federal government and Inuit Tapiriit Kanatami (ITK) are seeking to eliminate tuberculosis (TB) across Inuit Nunangat by 2030. The ITK TB Elimination Framework calls for ensuring "equitable access to high-quality and new/emerging diagnostic tools" for TB prevention and care. Chest X-ray (CXR) plays a central role in TB care pathways. Yet most Inuit Nunangat clinics and hospitals do not have on-site radiologists for rapid, accurate CXR reading. This introduces delays-and at times, errors-in TB care. In 2021, the World Health Organization (WHO) approved computer-aided detection (CAD)-artificial intelligence-based radiology analysis-to detect TB on CXR, as an alternative to human readers. CAD provides an immediate CXR reading at the point of care. However, WHO recommends that sites validate CAD using their own data prior to implementation because of variability in the technology's accuracy-i.e. its capacity to correctly identify the presence or absence of TB on CXR. We will evaluate CAD in Nunavik, Quebec, where TB occurs at a rate much higher than the provincial average. Our study uses existing CXR and clinical data from Nunavik, to estimate the accuracy of CAD, and the economic implications for the healthcare system if CAD is used instead of the status quo. A Steering Committee will provide community oversight and guidance for the research. Our research team includes representatives of the Nunavik Regional Board of Health & Social Services (NRBHSS), and frontline clinicians. At the end of the project, we will hold a stakeholder meeting to determine if CAD should be used, or further studied, or neither. The NRBHSS has a number of initiatives to address social determinants of TB and to remove barriers to accessing TB services, which will be essential to achieving TB elimination in Nunavik. Our research is essential for determining if CAD could also contribute to TB elimination by improving the use of CXR for TB care in Nunavik.