‘Lead Time’ TTO improves upon conventional TTO by providing a uniform method for eliciting positive and negative values. This research investigates (i) the values generated from different combinations of time in poor health and in full health; and the order in which these appear (lead vs. lag); (ii) whether values concur with participants' views about states; (iii) methods for handling extreme preferences. n = 208 participants valued five EQ-5D states, using two of four variants. Combinations of lead time and health state duration were: 10 years and 20 years; 5 years and 1 year; 5 years and 10 years; and a health state duration of 5 years with a lag time of 10 years. Longer lead times capture more preferences, but may involve a framing effect. Lag time results in less non-trading for mild states, and less time being traded for severe states. Negative values broadly agree with participants' stated opinion that the state is worse than dead. The values are sensitive to the ratio of lead time to duration of poor health, and the order in which these appear (lead vs. lag). It is feasible to handle extreme preferences though challenges remain.