PURPOSE: Several EQ-5D-5L valuation studies have been carried out in Asia, and more are on the way. These all included two valuation techniques, composite time trade off (C-TTO) and discrete choice experiments (DCE). Japan and South Korea have reported value sets based on C-TTO valuations only. However, a method for combining C-TTO and DCE valuations, called hybrid regression, has recently become available. The purpose of this workshop is to introduce the hybrid models, demonstrate their use in estimating EQ-5D-5L value sets, and to debate the merits and drawbacks of hybrid modelling vs. using C-TTO alone. DESCRIPTION: We will introduce participants to the valuation techniques currently employed by the EuroQol group: C-TTO and DCE. Afterwards, we will train them in estimating hybrid models using a Stata add-on (the hyreg command) developed by the discussion leaders. We will cover: 1) how to set up the data, 2) how to handle heteroscedastic data, 3) handling censored C-TTO observations, and 4) how to deal with convergence issues in hybrid model estimation. In order to compare the hybrid estimations against models using C-TTO alone, methods for estimating various C-TTO models will also be introduced: 1) Regular linear OLS models and 2) Non-linear multiplicative models. We will use data collected in China, Singapore, and Hong Kong. The last part of workshop will be an open discussion about the modelling results when using only C-TTO data or when merging C-TTO and DCE data to estimate a country specific value set. Participants will be encouraged to participate by asking questions and sharing their views.