Figure 4.

(a) and (b). Incremental cost-effectiveness scatterplots, showing pairwise comparison of strategy 2 (methotrexate + TT (a)) and strategy 3 (methotrexate + b/tsDMARD1 (b)), both compared to strategy 1 (methotrexate monotherapy) as the base case. The expected values for cost and effectiveness for strategy 1 are located at the intercept of the y and x axis, 0 cost, and 0 effectiveness. All iterations of strategies 2 and 3, shown as dots, indicate the incremental difference in cost and effectiveness to the average values for strategy 1. In most iterations, both strategies are more effective and more costly than the base case, indicated by the majority of iterations location in the North-East quadrant of the graphs. The dotted line delineates the willingness to pay threshold for Zanzibar (WTP) set at USD 282. With the estimated WTP, the more effective treatments are, with high probability, not cost-effective treatment options for Zanzibar, shown as red dots. While more costly and more effective, the incremental gain in effectiveness can hardly be justified with the corresponding incremental cost, when assuming the ICER threshold for Zanzibar. Only 4.7% of iterations suggested strategy 2 as cost-effective treatment in this comparison. Comparing strategy 3 to strategy 2, 23.5%, a higher share of iterations favor strategy 3. This suggests less certainty for a potential decision in favor of strategy 1 in this comparison with the uncertainty we introduced in the model.