Time trade-off user manual: props and self-completion methods

Abstract

In 1992, the Measurement and Valuation of Health (MVH) Group at the Centre for Health Economics conducted a study with Social and Community Planning Research (SCPR) comparing different methods of valuing health states (Dolan et al, 1993). A random sample of 335 members of the general population were interviewed in their own homes by specially trained interviewers. Each respondent was asked to value a series of health states using two different valuation methods – Standard Gamble (SG) and Time Trade-Off (TTO). Considerable time and energy went into the production of the protocols for the interviews. Standard methodology (derived primarily from research in Canada and the USA) for both the SG and TTO methods involves the use of specially designed boards and cards. He SG procedure typically uses a ‘probability wheel’ which allows different probabilities of health outcomes to be presented to the respondent, while the TTO typically uses a horizontal sliding scale which allows the length of time spent in a health state to be varied. SG and TTO boards based on the standard methodology were piloted as part of the MVH study and it was found that substantial modifications were required to simplify the material for both the interviewer and the respondent. In addition the standard boards were found to be too large and were difficult to operate. The substantive change made to the SG board during the course of the piloting was the use of a sliding scale rather than a wheel, and a new TTO board was designed so that both sides could be used – one for states considered better than death, and the other for states considered worse than death. A fundamental question arising from the pilot work was the advantage of using props such as boards and cards in the interview. To address this issue, an alternative method of administering the SG and TTO tasks was developed representing a significant departure from the standard methodology. In these modified procedures, the respondent was able to take a much more active role and in fact completed much of the valuation task by him/herself without the use of a board. All four methods performed very well in the main study, to the extent that no single method proved decisively superior to all others from an administrative point of view. Ultimately the choice of method was based on empirical grounds, with the result that the TTO ‘Props’ (with board and cards) was selected as the ‘best’ method for valuing health states in population surveys. Although the MVH Group is now concentrating on the TTO ‘Props’ method in further work, there are certain to be other researchers who want to use the SG method or the TTO in its self-completion form. Thud we want to ensure that all our methods are available to other interested parties in the field of health status measurement. The health states used in this study were based on the EuroQol descriptive system (kind et al, 1994), but these valuation procedures have a general application and can be used for any health state descriptive system. Being aware of the considerable work required in designing and piloting any new methods, we felt that it would be useful if other researchers were able to gain access to a detailed account of the procedures that we had developed. In order to maximise the availability of these designs, we have decided to supplement the initial report describing the piloting and interview design (Thomas and Thomson, 1992) with specific User Guides detailing the four valuation methods. Standard Gamble: Props and Self-completion Time Trade-Off: Props and Self-completion Revisions to the TTO Props method as a result of more recent survey work have also been included. We hope others will be able to pick up from where we have left off, either to make use of the methods in their current form or to modify them further as they wish. In either event we look forward to interest to hearing of the results.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 104 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.