Walker, D; (2001) Economic analysis of tuberculosis diagnostic tests in disease control: how can it be modelled and what additional information is needed? The international journal of tuberculosis and lung disease, 5 (12). pp. 1099-1108. ISSN 1027-3719 https://researchonline.lshtm.ac.uk/id/eprint/17301
Permanent Identifier
Use this permanent URL when citing or linking to this resource.
https://researchonline.lshtm.ac.uk/id/eprint/17301
Abstract
As the goal of tuberculosis (TB) control programmes is to get as many cases as possible onto correct treatment, the critical role of diagnosis is obvious. Under-diagnosis contributes to further spread of the disease, while over-diagnosis wastes scarce resources on inappropriate treatment. Various techniques are available with which to diagnose TB, including sputum smear microscopy, chest X-ray, antibiotics and culture. While it may be possible to improve the application of these diagnostic tools, there is universal recognition of their limitations. Although new technologies have been developed, their appropriateness for use in resource-poor settings has been questioned due to concerns associated with their complexity and high cost. Given the constrained resources of most countries heavily affected by TB, economic evaluation provides a powerful tool to facilitate the prioritisation of resources. However, economic appraisals of diagnostic procedures provide several unique methodological challenges, including specification of appropriate alternatives, measuring costs, and measuring outcomes. Nevertheless, the use of decision analytic models can provide an opportunity to explore many of these issues. This paper presents key problems in the diagnostic process, reviews existing diagnostic techniques, and discusses possible improvements of these techniques and the introduction of new techniques. Next it examines the usefulness of modelling the economics of TB diagnosis, highlighting gaps in information requirements. Finally, the article identifies economic and epidemiological factors that are likely to change results in different settings.