Population-based Time Preferences for Future Health Outcomes

Abstract

CONTEXT: Time preference (how preference for an outcome changes depending on when the outcome occurs) affects clinical decisions, but little is known about determinants of time preferences in clinical settings. OBJECTIVES: To determine whether information about mean population time preferences for specific health states can be easily assessed, whether mean time preferences are constant across different diseases, and whether under certain circumstances substantial fractions of the patient population make choices that are consistent with a negative time preference. DESIGN: Self-administered survey. SETTING: Family physician waiting rooms in four states. PATIENTS: A convenience sample of 169 adults. INTERVENTION: Subjects were presented five clinical vignettes. For each vignette the subject chose between interventions maximizing a present and a future health outcome. The options for individual vignettes varied among the patients so that a distribution of responses was obtained across the population of patients. MAIN OUTCOME MEASURE: Logistic regression was used to estimate the mean preference for each vignette, which was translated into an implicit discount rate for this group of patients. RESULTS: There were marked differences in time preferences for future health outcomes based on the five vignettes, ranging from a negative to a high positive (116%) discount rate. CONCLUSIONS: The study provides empirical evidence that time preferences for future health outcomes may vary substantially among disease conditions. This is likely because the vignettes evoked different rationales for time preferences. Time preference is a critical element in patient decision making and cost-effectiveness research, and more work is necessary to improve our understanding of patient preferences for future health outcomes.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 60 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.