OBJECTIVES: A 2014 CEA from the Singapore healthcare system perspective was published, claiming that Metformin and Lifestyle modification treatments were cost-effective in Pre-Diabetes. The analysis used ICERs and made a classic error in calculating ICERs vs. a common (Placebo) alternative. Our objective was to show NMB advantages over ICERs.

METHODS: We reproduced the original costs, QALYs and ICERs, then calculated ICERs properly. We then calculated NMB values for each treatment, based on a willingness to pay (WTP) threshold of 50,000/QALY. Lastly, we performed a Probabilistic Sensitivity Analysis (PSA) that relied on NMB values in each iteration.

RESULTS: Successful reproduction of the cost and QALY numbers indicated consistency with the original analysis. The simple NMB calculation showed that Lifestyle had the maximum NMB when the WTP is a conventional range ($50-$150,000). Correct ICER calculations confirmed this result, both analyses indicating the clear error in the original analysis that labelled both Lifestyle and Metformin as cost-effective.

CONCLUSIONS: NMB calculations easily avoid a classic error to which ICER calculation is prone when multiple treatments are involved (failure to account for extendedly dominated alternatives). Moreover, this method allows a simple calculation of decision uncertainty from a PSA that ICERs could achieve only with a great deal of time and attention to detail as interpretation is contingent on the quadrant of the Cost-Effectiveness Plane for each iteration of the PSA. The Singapore DPP CEA conclusion, using ICERs improperly, was that both the Lifestyle and Metformin were cost-effective. This is incorrect and is seen easily via the very straightforward NMB calculations. Only Lifestyle is cost-effective at conventional WTP thresholds.