Endogenous Patient Responses and the Consistence Principle in CEA

Presenter: Liqun Liu, Texas A&M University

Abstract

Rationale: Medical interventions often have further or “unrelated” financial and health impacts, in addition to incurring direct treatment costs and generating direct health benefits that improve longevity and/or health-related quality of life. For example, any life-extending intervention would cause increased future consumption costs and/or increased future labor earnings, referred to in the literature on medical CEA as survival consumption costs and survival earnings, respectively. These “unrelated” financial and health impacts derived from a medical intervention, as opposed to the direct treatment costs and the direct health effects of the intervention, have been considered as imposing a challenge to the medical CEA. Specifically, an issue frequently raised in the medical CEA literature is when calculating the C/E ratio for an intervention, which of its unrelated financial and health impacts should be explicitly accounted for?

Objective: This paper reconciles two alternative approaches to this issue: the approach based on lifetime utility maximization as exemplified by Meltzer (JHE 1997) and the approach based on logical consistence (the “consistence principle”) as exemplified by Nyman (HE 2004). The “consistence principle” states that, in the evaluation of a medical intervention, one should either include both the cost and the utility benefit of a change (directly or indirectly) caused by the intervention or include neither of them. We focus on two questions: (1) whether survival consumption costs and earnings due to an intervention should be included in the numerator of the C/E ratio when only the QALY gain from the intervention is included in the denominator and (2) how to choose among alternative internally consistent exclusion/inclusion rules regarding the costs and utility benefits of an intervention.

Methodology: We generalize the QALY-consistent lifetime utility maximization model of Bleichrodt and Quiggin (JHE 1999) to also include leisure, in addition to longevity, health-related quality of life and consumption. A QALY-consistent model has the advantage that the health improvements from an intervention are measured by a gain in the QALY measure. We also explicitly distinguish between two types of changes caused by an intervention: those exogenous changes directly brought about by an intervention and those endogenous changes resulting from the patient’s optimal responses to the exogenous changes.

Results: Regarding whether survival consumption costs and earnings should be included in the C/E ratio, we find that both welfare maximization and the consistence principle require that survival consumption costs and earnings directly associated with an intervention should be included in the numerator of the C/E ratio even though the denominator of the C/E ratio only includes the direct QALY gain from the intervention. Regarding how to choose among alternative internally consistent exclusion/inclusion rules about the costs and utility benefits of an intervention, we find that for an endogenous change resulting from the patient’s optimal response to an intervention, excluding or including both the cost and the utility benefit of the change makes no difference to the evaluation of the intervention but the exclusion option certainly has an informational advantage.

Conclusion: There exists significant ambiguity about the appropriate way to conduct medical CEA due to seemingly conflicting approaches. This paper reconciles the two main existing approaches and sheds light on questions that have been puzzling health economists working on cost-effectiveness analysis. Hopefully our research helps reduce the ambiguity in medical CEA.