AN EMPIRICAL COMPARISON OF THREE VERSIONS OF THE TIME TRADE-OFF

Purpose: To determine how the framing of the questions in the time trade-off method (TTO) affects quality of life (QoL) estimates, and to analyze the effect of feedback.

Method(s): We invited 3,465 students to answer an on-line survey. Each respondent was randomly assigned to one of four methods: 1. visual analog scales (VAS); 2. TTOe, which asked how much life time with perfect health is equivalent to some time in a certain health condition; 3. TTOt, in which respondents can give up life time in order to gain QoL; 4. TTOq, in which they can give up QoL to gain life time. We offered them a description of the performance of children having one cochlear implant vs. those having two. The feedback for the TTO versions was to show two VAS’s (one for each health condition); the feedback for the VAS was the TTOe. All the respondents could revise their estimates after receiving the feedback.

Result(s): 583 respondents entered the survey (response rate: 17%), but some of them abandoned without answering the QoL questions. Some subjects assigned to TTOt or TTOq refused to enter the trade-off game. TTOe collected significantly more qualitative estimates (86.8%) than TTOt (69.5%) and TTOq (64.1%). The average QoL for unilateral cochlear implantation was higher for TTOt than for TTOe (0.787 vs. 654, p < 0.001) and for TTOe than for TTOq (0.654 vs. 0.618, p < 0.0001); a similar pattern was observed for bilateral implantation. This was expected from previous psychological studies, which showed that people are reluctant to exchange what they have for goods of similar value. The feedback reduced the variability for VAS, TTOt and TTOe. The respondents considered that TTOt posed more difficult questions than the others.

Conclusion(s): Even though TTOt is the standard technique, we recommend the TTOe because it seems to be the most neutral version, while TTOt and TTOq bias the answer in opposite directions. Additionally, in our experiment TTOe collected more qualitative responses than TTOt and was deemed easier to answer.