Disentangling public preferences for health gains at end-of-life: Further evidence of no support of an end-of-life premium