The CEA study will identify the most cost-effective NMEP activity. The units of this CEA are likely to be NMEP specific and not comparable to external benchmarks. Cost-consequence measures, on the other hand, are less useful internally ( since each activity has distinct mechanisms) but may be more useful externally. Given a particular cost-effective policy, cost-consequence data represent a useful metric for comparing to external benchmarks, such as similar activities (e.g., Web sites) used by other programs.
If the MCBS does not contain sufficient data on NMEP activities to generate meaningful comparisons of the relative cost-effectiveness of different activities, a pure costing study and a cost-consequence study appear feasible and would still permit external comparisons of the NMEP to other programs and policies.
Even with good data, CEA results must be used cautiously when recommending major policy changes, such as phasing out one NMEP activity and redirecting those resources to another. First, redistributing resources may have unintended consequences. For instance, the 1-800-MEDICARE helpline may serve many beneficiaries without Internet access. Even if the helpline costs twice as much for the same change in outcome, eliminating the helpline and emphasizing the Web site may not serve all beneficiaries equally well, regardless of the cost savings. Second, the marginal cost measurement in CEA is not appropriate for “large” policy changes. Costs and outcomes may vary considerably from the environment in which they were measured. For example, capacity constraints on infrastructure may limit the ability for large change or affect costs in unplanned ways.