Cost–Utility Analysis

The means and SDs were calculated for the various resources used and other variables related to the 2 breast reconstruction approaches. The calculated costs and effectiveness were tabulated for each procedure. Effectiveness was measured as QALYs; therefore, if one reconstructive approach had a lower mean cost and provided more QALYs, it was deemed the “dominant approach” as it fell in the “win–win” quadrant of the cost-effectiveness plane (Fig. ​(Fig.2).2). Such approach would then be labeled cost-effective. If both the costs and the effectiveness were higher, an ICER was calculated. This ratio is the difference of costs (y axis) divided by the difference of the effectiveness (x axis) of the 2 reconstructive approaches (Fig. ​(Fig.2).2). If this ICER fell within the acceptability threshold of $50,000/QALY, then the procedure would still be considered cost-effective. The use of $50,000/QALY is a commonly cited threshold within the literature.¹⁶ The ICER represents the additional costs required to gain one additional unit of benefit (ie, cost per QALY gained). To explore and quantify the uncertainty in this economic evaluation, nonparametric bootstrapping to quantify the joint effect of uncertainty around the costs and QALY variables was undertaken.¹⁶ This technique randomly draws with replacement samples of the original cost and QALY data over 1000 replications. These bootstrapped cost–effect pairs are graphically represented on an incremental cost-effectiveness plane. The bootstrapped estimates can be used to construct cost-effectiveness acceptability curves (Fig. ​(Fig.3).3). These show the probability that AAT is cost-effective compared with TE/I from the perspectives of the MOH and of society.

Illustrative cost-effectiveness plane.

Cost-effectiveness acceptability curves for societal and Ministry of Health (MOH) perspectives. Cost-effectiveness acceptability curves for TE/I breast reconstruction across 2 perspectives. Cost-effectiveness acceptability curves were based on 1000 bootstrap cost–effect pairs. Yellow line represents the MOH perspective; blue line represents the societal perspective. At the willingness-to-pay threshold of $50,000/QALY, there is >90% probability that the TE/I is cost-effective.

When analyzing data from the HUI-3, 1 in TE/I and 2 in AAT responses were missing at each time point. Group mean was used to impute the missing values. Together with probabilistic sensitivity analysis, the impact of the missing data and the sampling uncertainty for health utilities was sufficiently dealt with through analyses. Cost-effectiveness analysis data was performed in Microsoft Excel; demographic and HRQoL differences between groups were analyzed in SPSS version 25.²⁷

As the time horizon for the study was 12 months, no discounting of costs and QALYs was necessary.¹⁵ The methodology used for the economic evaluation was obtained from Drummond and colleagues¹⁵ and the Methodological guide in performing cost–utility analysis comparing surgical techniques.²⁸