OBJECTIVES

Derive social values of the EuroQol EQ-5D-5L in Peru using a “Lite” valuation protocol.

METHODS

Random sample of the Peruvian population in Lima, Arequipa and Iquitos. We followed the Standard EQ Valuation Technology (EQVT) protocol design (in which subjects undertook both a Time Trade Off -TTO- and a Discrete Choice Experiment -DCE- component) in 300 subjects (30 total states in the TTO task, split in three blocks of ten plus the worst possible state; 8 blocks of 10 pairs in the DCE task); and a “DCE only” survey for the remaining 700 subjects (196 EQVT pairs divided over 28 blocks of 10 pairs). Trained interviewers conducted home face to face interviews and a strict quality control protocol applied. In our main analysis, a hybrid modelling approach is used in order to combine TTO (i.e continuous) and DCE (i.e dichotomous) preference values. Models that accounted for censoring and heteroskedasticity were assessed. We additionally tested TTO only (GLS and Tobit GLS) and DCE only models (Logit and Probit). All models included 20 incremental parameters.

RESULTS

We included 993 subjects (55% females; mean age 45), reported problems ranged from 8% in “self-care” to 55% in “pain/discomfort”. All models tested (TTO only, DCE, and hybrids) showed logically consistent parameters (97.5% TTO only models; 100.0% DCE only models; 98.3% hybrid models). Model coefficients were also mostly statistically significant (82.5% TTO only models; 75% DCE only models; 83.3% hybrid models). Hybrid model incremental coefficients ranged from Usual Activities level 3 (-0.00005; NS) to Mobility level 4 (0.25; p<0.001).

CONCLUSIONS

We derived the EQ-5D-5L value set in Peru, the first to date to be based in a Lite protocol less reliant on TTO only interviews and relying also on DCE, using a hybrid modeling approach. These results will help inform decision making for resource allocation decisions.