An assessment of nucleic acid amplification testing for active mycobacterial infection

Is NAAT cost-effective?

The base-case ICER for NAAT (the TB mixed scenario) is $90,728/QALY. The addition of NAAT leads to more patients initially receiving the correct treatment, due to improved sensitivity of NAAT in conjunction with AFB and the ability to identify MDR-TB. The incremental cost of NAAT is driven predominantly by the cost of testing, offset by reduced TB transmissions and hospitalisation costs. The incremental QALY gain is driven by the shift of TB patients from being initially untreated (or having standard treatment in the case of MDR-TB) to receiving correct treatment.

The cost-effectiveness of NAAT is affected by the extent of use of clinical judgement in initial treatment decisions. In the extreme scenario, in which clinical judgment is not exerted (i.e. treatment initiation decisions are based on the results of testing), NAAT is most cost-effective due to improved sensitivity in conjunction with AFB, thereby reducing the number of patients who would have been untreated on the basis of AFB results alone. However, in the scenarios in which clinical judgement perfectly identifies TB or in which clinical judgment is used as the basis to treat all patients, the benefits of NAAT are restricted to identifying rifampicin resistance, and so are accrued in a very small proportion of the population tested (2% of 22% = 0.44%).

Substantial uncertainty surrounds a number of variables included in the economic modelling, in particular the prevalence of TB in the tested population. The ICER is most sensitive to changes in this variable; for example, decreasing the estimated prevalence in the tested population from 22% to 10% increases the ICER to $967,000.

The ICER is also sensitive to decreases in the specificity of NAAT, particularly in AFB-negative results (e.g. using the lower limit of the 95%CI increases the ICER to $450,000) and for rifampicin resistance (e.g. using the lower limit of the 95%CI increases the ICER to $253,000). Any decrease in these specificities (from 100%) increases the number of false-positive patients that receive poorly tolerated treatment, leading to increases in cost and poorer quality of life. However, as culture is an imperfect reference standard for diagnosis of TB, some proportion of NAAT false-positive patients may truly have clinical disease, and so the uncertainty in the ICER associated with reductions in the specificity in AFB-negative results may be an overestimate.