OBJECTIVES: Discrete Choice Experiments (DCEs) have received considerable interest in EQ-5D valuations because these methods are simple and choice based. Some investigations using DCEs to develop EQ-5D Value Sets include implausible states. The DCE literature discourages the use of implausible options, citing the links between cognitive burden (facing respondents) and random variability in responses.  It has been suggested that giving DCE respondents choice sets with more than 2 alternatives may improve respondent effort/concentration. METHODS: A D-Efficient DCE design for EQ-5D-3L comprising 15 choice sets of 3 alternatives each, that excluded states which combine a ‘3’ on Mobility with a ‘1’ on either Usual Activities, or Self Care or both was developed. As a pilot test, and as a first step towards developing an on-line elicitation exercise, an elicitation exercise was put into a  Microsoft PowerPoint file which was sent out to 189 respondents at various work locations in Trinidad and Tobago with a submission deadline of 2 weeks.  The elicitation process was based on a protocol using sequential comparisons that break the choice process into s series of simple ranking steps. RESULTS: Sixty-five Respondents returned DCE results within the 2 week period of which the results from 43 respondents were useable (66%). A utility function was estimated based on data from the 43 useable files. This produced an internally valid model with all coefficients bearing the correct signs, and relative magnitudes, and all significant to the 5% level.  CONCLUSIONS: This DCE design allows for EQ-5D-3L valuation studies that can be conducted with small samples (no blocking) and it avoids the use of implausible states. It introduces a DCE protocol that allows respondents to easily make choices out of choice sets comprising 3 EQ-5D-3L States.  Increasing the number of alternatives in each choice set allows for designs with fewer choice sets.