Purpose: When predicting valuations for pain profiles one may use a QALY like, Area Under the Curve (AUC) approach based on time and intensity. Another strategy would be to apply a more holistic, Multi-Attribute Utility Theory (MAUT)-based approach, focusing on distinct pain characteristics. It is hypothesized that both approaches will lead to similar results.

Methods: Annual pain profiles, which comprehended several attributes describing the course of the pain (in writing as well as graphically), were constructed based on real pain data. Valuations for these profiles were holistically obtained by means of a Time Trade Off (TTO) technique. Each respondent valuated 12 unique profiles. The valuation data were modeled by means of Tobit models using 1) the AUC, 2) the distinct profile-attributes (MAUT-approach), and 3) both as explanatory variable(s). Estimates were described (level of significance: p < 0,05) and performance of the models was assessed by means of Akaike's Information Criterion (AIC). For the nested models, the differences in fit were tested for significance by means of the Likelihood Ratio test.

Results: In total 711 valuations of unique profiles were collected from 68 respondents. In the AUC-model, AUC was strongly significant (p = 0,000). In the MAUT-model, p-values of the attributes varied from 0,049 to 0,000. AIC's were 7092 (AUC) and 6992 (MAUT). In the combined model, AUC did not attribute to the explanation of the valuations (p = 0,11). Difference in fit between the AUC and the combined model was strongly significant (p = 0,0000), while the difference between the MAUT-model and the combined model was not (p = 0,11).

Conclusions: The valuation associated with pain profiles is much better captured by a description of the distinct pain characteristics than by a measure of the cumulative intensity of pain. Similar findings might result when considering QALYs instead of pain, which raises doubt about the validity of some of the assumptions underlying the QALY-model, most notably in case of health profiles in which quality of life fluctuates over time.