Diagnostic tests play a crucial role in the control and surveillance of infectious diseases, and to ensure effective clinical management. Novel diagnostic tests are traditionally evaluated in terms of their accuracy (sensitivity and specificity). Using mathematical models, I examine the impact of different novel diagnostic tests on the tuberculosis (“TB”) and severe acute respiratory syndrome coronavirus 2 (“SARS-CoV-2”) epidemics, and investigate how the context in which these tests are used may affect this impact. In chapter 3, I evaluate the use of a hypothetical biomarker test that can detect individuals at imminent
risk of progressing to active TB disease (incipient TB) and subsequent TB preventive treatment
(“TPT”) initiation in a high TB burden setting. I demonstrate that biomarker-led TPT can have a
significant impact on TB incidence in a high TB burden setting; however, the cost of implementing
such a strategy is likely to be prohibitive given the testing effort needed to identify those with
incipient TB, even if testing is targeted to populations at high risk of TB.
Next, in chapter 4, I evaluate the use of urine-based tests for active TB disease, in a high TB and HIV
burden setting. Although current urine-based tests for TB suffer from poor sensitivity, these tests are
continuously improving, and are essential for TB diagnosis amongst patient subgroups who struggle
to produce quality sputum and who are therefore missed by traditional methods of TB diagnosis. I
demonstrate that although urine-based diagnostic tests reduce mortality amongst people living with
HIV, population-level epidemiological impact is not seen unless the tests are deployed outside of HIV
care and into routine TB care.
In chapter 5, I investigate the cost and epidemiological impact of expanding different TB publicprivate sector engagement services. My results reveal that services involving the use of Xpert, a
highly accurate but costly test for diagnosing active TB disease are epidemiologically impactful, but
costly, and thus have the highest cost per TB case or TB death averted than other services.
Finally, in chapter 6, I explore the context under when cheap but less accurate rapid antigen
diagnostic tests (“Ag-RDTs”) offer greater public health value than more accurate but costly nucleic
acid amplification tests (“NAATs”) that often have high turnaround times. My results highlight that
Ag-RDTs-led strategies, despite their imperfect sensitivity and specificity, are more impactful at a
lower cost than NAATs under different use-cases.
Overall, the context in which diagnostic tests are used is crucial in anticipating their impact. Factors,
including but not limited to testing eligibility, levels of current testing and turnaround time, can affect
6
the potential epidemiological impact of a diagnostic test. Thus, future evaluation of diagnostic tests
should move away from focussing exclusively on accuracy and move towards clearly defining
different use-cases and investigating which factors other than accuracy, may affect the
epidemiological impact of a diagnostic test.