Benefit:Risk Evaluation in Diagnostics: An Example Comparing Two Rapid Molecular Diagnostics Testing Imipemem Susceptibility in Acinetobacter spp.

Background: Methods to assess benefit:risk of diagnostic methods to help clinicians treat multidrug-resistant Acinetobacter spp. are needed. Important considerations include effects on diagnostic yield [(i.e., the distribution of true resistance (TR), false resistance (FR), true susceptibility (TS), and false susceptibility (FS)], which is affected by the susceptibility rate.

Methods: We developed a slide-rule profile plot of the expected diagnostic yield as a function of the susceptibility rate. The number-needed-to-test (NNT) for susceptibility (NNT-S) [i.e., the expected number of susceptible patients that must be tested with one test (vs. another test) to result in one additional TS] was defined, as was the NNT for resistance (NNT-R). We apply the method to the evaluation of two rapid molecular diagnostic platforms [PCR/ESI-MS and Molecular Beacons™ (MB)] utilized for imipenem susceptibility testing based on the presence/absence of blaMBLs and blaOXA genes vs. reference MIC determination in Acinetobacter spp. (Ab) infection. Expected diagnostic yield was evaluated based on a population of 12,000 Ab cases.

Results: When the imipenem susceptibility rate is 40%, the expected yield for MB is (TR=6408; FR=384; TS=4416; FS=792) and for PCR/ESI-MS is (TR=6912; FR=816; TS=3984; FS=288). Thus the between-platform difference in expected yield is 504 more TR for PCR/ESI-MS and 432 more TS for MB. When the susceptibility rate is 60%, the difference in expected yield is 336 more TR for PCR/ESI-MS and 648 more TS for MB (Figure 1). Tradeoffs for any susceptibility rate can be obtained (Figure 2). The NNT-S is 11.1 (MB vs. PCR/ESI-MS), and the NNT-R is 14.3 (PCR/ESI-MS vs. MB).

 

Conclusion: Here we develop profile plots to illustrate how diagnostic yield varies by susceptibility rate. Between-platform differences in TR and TS plotted as a function of the susceptibility rates and NNTs provide a relative comparison of two AST methodologies. Applying these approaches helps us understand the benefits and risks of each new diagnostic platform when applied to treatment decisions.

 

  Figure 1: Slide-Rule Profile Plots of Diagnostic Yield for Imipenem Susceptibility Testing in Ab

 

      

 

 

Figure 2: Between-Platform Difference in TR and TS as a Function of the Susceptibility Rate