Indirect comparisons in cost-effectiveness analysis: are we being naïve?