Estimation of a multiattribute utility function for the Spanish version of the TooL questionnaire.

Objectives: To estimate and assess the psychometric properties of a multiattribute utility function (MAUF) for the Spanish version of the Tolerability and Quality of Life (TooL questionnaire).

Methods: Balanced data on 243 patients diagnosed with schizophrenia or bipolar disorder were gathered. In addition to the demographic and clinical variables and the usual generic health-related quality of life (HRQoL) questionnaires (EuroQol-5D [EQ-5D] and Short Form-6D [SF-6D]), instruments considered included the Spanish versions of the Positive and Negative Symptoms of Schizophrenia Scale (PANSS), Young Mania Rating Scale (YMRS), Montgomery-Asberg Depression Rating Scale (MADRS), Udvalg for Kliniske Undersogelser (UKU), and Clinical Global Impression Severity (CGIS) scale. MAUF parameters estimation involved a number of visual analogue scale (VAS) and time trade-off (TTO) ratings that proved difficult to be performed by the patients. After checking for inconsistencies in patient responses, the original sample was reduced to a still balanced subsample of 70 individuals. A multiplicative-form MAUF was estimated following the standard methodology.

Results: Good convergent validity was demonstrated because utility estimates from the MAUF presented strong correlations with utilities from the generic HRQoL instruments included: SF-6D (0.66, P < 0.01), EQ-5D (0.69, P < 0.01), and moderate correlations with the rest of instruments considered: PANSS (-0.27, P = 0.10), YMRS (-0.30, P = 0.08), MADRS (-0.48, P < 0.01), UKU (-0.35, P < 0.01). Criterion validity was also met because differences in mean utilities by clinical severity were found (P < 0.01). Utilities from the MAUF covered a wider range of health states [0.04,1.00] than those from the SF-6D [0.53,1.00] and EQ-5D [0.23,0.96].

Conclusions: Utilities from the MAUF showed good psychometric properties, serving as a complement to generic health utilities. If misapplied, however, utilities from this instrument might favor the positive evaluation of drugs showing fewer associated side effects.