burnsg

For those two reversal thresholds, would you vary observed-period and extrapolated-tail utilities independently, or link them when the same utility estimate is carried forward beyond follow-up?

That comparison needs to separate utility accrued during observed follow-up from utility projected in the extrapolated tail. Report the incremental quality-adjusted life-years from each period under measured and assumed utilities, then vary the utility nearest the decision threshold while holding survival and costs fixed. A reversal confined to the tail would indicate dependence on extrapolation rather than on measured patient experience. Without those period-specific results, persistence of the cost-effectiveness conclusion cannot be determined from the publication title alone.