Comparison of online and face-to-face valuation of the EQ-5D-5L using composite time trade-off

Abstract

Objective The aim of this study was to compare online, unsupervised and face-to-face (F2F), supervised valuation of EQ-5D-5L health states using composite time trade-off (cTTO) tasks. Methods The official EuroQol experimental design and valuation protocol for the EQ-5D-5L of 86 health states were implemented in interviewer-assisted, F2F and unsupervised, online studies. Validity of preferences was assessed using prevalence of inconsistent valuations and expected patterns of TTO values. Respondent task engagement was measured using number of trade-offs and time per task. Trading patterns such as better-than-dead only was compared between modes. Value sets were generated using linear regression with a random intercept (RILR). Value set characteristics such as range of scale and dimension ranking were evaluated between modes. Results Five hundred one online and 1,134 F2F respondents completed the surveys. Mean elicited TTO values were higher online than F2F when compared by health state severity. Compared to F2F, a larger proportion of online respondents did not assign the poorest EQ-5D-5L health state (i.e., 55555) the lowest TTO value ([Online] 41.3% [F2F] 12.2%) (p < 0.001). A higher percentage of online cTTO tasks were completed in 3 trade-offs or fewer ([Online] 15.8% [F2F] 3.7%), (p < 0.001). When modeled using the RILR, the F2F range of scale was larger than online ([Online] 0.600 [F2F] 1.307) and the respective dimension rankings differed. Conclusions Compared to F2F data, TTO tasks conducted online had more inconsistencies and decreased engagement, which contributed to compromised data quality. This study illustrates the challenges of conducting online valuation studies using the TTO approach.

United States Utility Algorithm for the EORTC QLU-C10D, a Multiattribute Utility Instrument Based on a Cancer-Specific Quality-of-Life Instrument

“…Although we used quota sampling for sex, age, race, and ethnicity to try to achieve representation across these characteristics, there may be differences between participants in online panels and the general population in unmeasured variables due to selection bias. 61 We observed some nonrepresentativeness, with our sample being somewhat more educated than the general population. Potential nonrepresentativeness of the sample is a valid concern, and it remains unknown whether health preferences are typically different in subpopulations defined by education, marital status, or other demographic characteristics.…”

“…Regarding the use of online panels per se (i.e., independent of differences in respondent characteristics), research indicates that online completion of DCE tasks provides similar findings to those derived from in-person administration, 62 although some studies demonstrate differences between face-to-face and online DCEs results. 61 There are also challenges with achieving representativeness with face-to-face administration and with other methods of recruiting general population samples. Second, the inclusion of only English-literate participants may have introduced some bias, if cultural differences translate into difference in how health states are valued.…”

EQ-5D-Y Value Set for Slovenia

“…For this reason, it could not reach part of the population with limited access to the internet or those without sufficient technical skills, such as certain groups of the elderly population or those living in remote areas. Online survey methods have been linked to concerns about whether the participants properly engage with and understand the task [35]. As the researchers had no face-to-face contact with the participant, it was not possible to directly estimate any lack of understanding, engagement or interest of the participant in the task.…”

Estimation of an EORTC QLU-C10 Value Set for Spain Using a Discrete Choice Experiment

“…First, it used an online administration procedure. While this has the advantage of being cheaper and less time consuming than face-to-face administration, it may be associated with poorer quality of data due to responders’ reduced engagement with the task and strategic behaviours [ 54 – 56 ]. To assess whether this occurred, time stamps were investigated, without finding any relevant problem.…”