Utility of pain states: what influences the value people give to being in pain

Abstract

Pain has a significant burden on individuals and society; however challenges remain in the measurement of pain-related utility. Research was conducted to design a measurement framework for valuing pain states, for use in policy making. Within the theoretical framework of utility theory, the bio-psychosocial model of pain was used to select co-variates impacting the pain-utility relationship. The applicability of two utility theories (prospect theory and hedonist theory) to pain was evaluated, with EQ-5D-5L and utility values for scenarios of different pain intensities elicited using time-trade-off (TTO) and willingness-to-pay (WTP). Prospect theory was tested using the pain-utility relationship in 600 members of the general population (‘genpop’) by examining the presence of a reference point and the relationship between pain decreases/increases (gains/losses) and utility. Hedonist theory was tested through examination of predicted, experienced and remembered utilities, and ‘peak-end’ effects, principally using data from 56 university athletes experiencing training pain. Pain intensity had the greatest influence on EQ-5D values, with present and worst pain showing additive effects. Duration of the pain episode, general health, mood, age and gender also had significant impacts. In addition to pain intensity, TTO ‘genpop’ models included age, income, experienced pain and general health; WTP ‘genpop’ models included residual pain, age and income. The TTO and WTP models did not fit the athletes’ data well and alternative models were developed. The data did not confirm prospect theory: a reference point could not be identified (‘genpop’ and athletes) and the shape of the utility curve contradicted theory (athletes). Results for hedonist theory were inconclusive: predicted utility was consistently higher than experienced utility in athletes but the relationship depended on pain levels in ‘genpop’; remembered and experienced utility differed despite being consistent for pain levels; peak-end effects were not found. Further research in controlled environments is recommended for further theory testing.