Approaches to Consider Spillover Effects Among Significant Others in Decisions About Health Care Interventions

OBJECTIVES: Recently, there has been an increased interest in how to consider spillover effects among significant others in healthcare decision-making. The purpose of this study was to analyze how different approaches to consider these effects may influence the results of health economic evaluations and decision-making.

METHODS: A Markov-model was developed to evaluate the cost-effectiveness of two hypothetical treatments. The first treatment (A) had effect on quality of life (QoL) and life expectancy of the patients, while the second treatment (B) only influenced the patients’ QoL. For the assessment of each treatment, five approaches were applied and compared to a base case without spillover effects.

RESULTS: In the approach where spillover effects were presented beside the cost-effectiveness analyses, the judgment on how much these effects should influence the decision was left to decision-makers. When the spillover effects were considered by applying a QALY multiplicator, a linear relationship between the effects on patients and significant others was assumed, and the cost per QALY gained decreased proportionally with the size of the multiplicator. When QoL of the significant others and costs of informal care were included in the model, the cost per QALY gained decreased when the treatment influenced the patients’ QoL only, but increased when it also influenced the survival of the patients. The results were influenced by the assumptions regarding the significant others after the death of patients. In the approach where both QoL of significant others and costs for informal care were included in the cost-effectiveness analysis, these influenced the results in the same direction.

CONCLUSIONS: The approach to consider spillover effects among significant others and associated methodological choices may influence the results of cost-effectiveness analyses and may consequently impact decisions regarding allocation of resources. When including both QoL of significant others and costs of informal care, there is a risk of double counting.