At the risk of sounding like a malfunctioning obsolete acoustical storage media, I am returning to the topic of last week’s post: the launch of the Xpert® MTB/RIF diagnostic test which the WHO stated could “revolutionize” TB care and control (Reuters article). WHO approved the test and issued a “Roadmap” to guide purchasers to the test’s best (most cost-effective) use (WHO Roadmap). My question was that, given that the multi-year, multi-million dollar (Gates Foundation and NIAID, expertly-managed and -advised (by FIND and multiple committees) effort resulted in a system that was limited in where it can be used (mid-tier labs with reliable electricity) and how (as a check on false negatives and to find cases of multi-drug resistant TB among HIV patients [an important problem but only about 5% of all cases]), was: what analyses were done (before, during, or after development) to figure out if the test can/will revolutionize TB care and control (or said another way, was the effort worth the investment). Although my point is clearly post (and ad) hoc, these types of analyses are important in prioritize and guiding the use of (likely to be always limited) resources in global health, and specifically in TB diagnostics development where there is an immense and pressing need and multiple technologies available (Medscape article) and being developed (e.g., FIND TB Projects and Rapid Biosensor Systems).
So in my Part II look I found two publications on the Xpert system’s cost effectiveness: Vassell et al., “Cost Effectiveness of the Xpert MTB/RIF” (2010, Stop TB Alliance Resources), which was cited in the WHO Roadmap, and Van Rie et al., “Xpert® MTB/RIF for point-of-care diagnosis of TB in high-HIV burden, resource-limited countries: hype or hope?” (2010, Expert Review article). Vassell et al. used data from the FIND demonstration studies (WHO Roadmap p. 2) to compare the costs and outcomes of current practice (sputum smear microscopy for initial dx and culturing assays for drug sensitivity [drug resistance]) to Xpert MTB/RIF using a probability tree analysis. In my reading, they found:
- Smear plus Xpert and Xpert alone improved case finding (all TB and MDR TB) and improved DALYs (disability-adjusted life years, a standard but criticized measure of improved health); and
- The incremental cost-effectiveness ratios (measured in euros per year of life gained [I think a lower ratio means more bang for the buck]) was better in smear plus Xpert than in Xpert alone (which generated more cost per DALY due to the additional treatment costs for the MDR-TB); but
- With the caveat that the study did not include a retreated patient scenario, effect of diagnostic delay on treatment outcomes or on transmission, and patient costs (and I think they did not include Xpert’s capital, setup, maintenance, and training costs).
Overall, as an unqualified reviewer, I agree with the analysis (the Xpert system is incrementally better and cost-effective), and it was useful to WHO as a basis for their recommendation of the use of Xpert to check the smear negatives and for initial screening in HIV patients in high TB prevalence areas, but did not support the “revolutionary” claim.
Van Rie et al. were cautious the impact of the Xpert test, pointed out its limitations, but did not generate specific cost-effectiveness data. They noted positively that the Xpert test met or exceeded 15 of 20 criteria developed by 2009 workshop organized by Medecins Sans Frontieres (MSF Workshop) with the cost being one criterion that was indeterminate. Of concern to them was that a single Xpert assay of HIV-infected smear-negative patients had a 72.5% sensitivity (i.e., missing about a quarter of positives) so a second assay should be used (and will add cost). Their bottom line (which sounds to me like a justification for more academic research) was: “Large-scale implementation of the Xpert MTB/RIF at POC in resource-limited settings will require a body of research to generate the evidence needed to inform policy and guidelines … [specifically] estimation of the impact of clinical decision making, the effect on healthcare workers’ workload, the impact on patient outcomes, health economic evaluations from both the patients’ and health provider’s perspective and, possibly, also longer-term evaluation of indirect effects, such as the impact on transmission of M. tuberculosis in communities and the prevalence of drug-resistant TB.” They recommend five years of research to figure out where/when to use the assay “to avoid an unwarranted explosion of costs.” Or maybe someone responsible should do some upfront modeling and analysis before committing to a large technology development project.
As for Cepheid, the company on whose platform the Xpert MTB/RIF runs and which received at least $3.3 million in grant funding for their role (Bizjournal Article), the level of interest in selling into the rest-of-world (non-US and EU) TB testing market seems low. While clearly pleased with their involvement in the program (Ceheid press release), the company does not mention in its 2009 annual report (2009 Annual Report) the possibility of sales to the many organizations FIND found as potential users (my post of last week). And the company does not appear to be pursing an FDA approval that would be needed for sales to US-funded organizations. I guess the business development folks at Cepheid ran the numbers and concluded that the effort to sell to public sector organizations trying to stop TB would not be cost-effective (or revolutionary for the company’s valuation). Further I guess FIND should have run the numbers themselves to create a business case for Cepheid increasing their sales meaningfully (the company sells about 500 of its several systems per year and another 50 machines would be a 10% increase and about $1 million in revenue). So after two posts on about the launch of the Xpert® MTB/RIF, my point is: organizations wanting improved diagnostic tests need to analyze (or even guesstimate) costs and benefits (do some business planning) before making a major investment in the test’s development, i.e., figure out if the investment will yield the expected return. Do I hear the sound of a malfunctioning acoustical storage media?