|
In a preference-based
approach, respondents assess the desirability of life in certain
health states. The health state rated may be:
- The respondent's
own current health, as when ratings are done as an outcome measure
in a clinical trial, or
- A hypothetical state
other than their own current health presented to respondents
for rating.
The ratings are performed
on a scale that is conventionally anchored at best possible health
(equal to 1) and worst possible health or death (equal to 0).
Ratings are then comparable across a wide variety of diseases,
treatments, and outcomes.
The preference-based
approach to health-related quality of life measurement is well-suited
to decision analysis and cost-effectiveness
analysis (CEA), where the value that the decision-maker
places on outcomes is critical to the decision. This framework
for decision-making is grounded in
von Neumann Morgenstern VNM utility theory (von
Neumann and Morgenstern, 1944). Utility theory for individual
decision making has been expanded to application for health policy
decisions through cost-effectiveness analysis (Garber,
Weinstein et al., 1996).
Well-known techniques
for preference-based rating of health states include: (Froberg
and Kane, 1989).
- The standard gamble
(SG), and
- The time trade-off
(TTO).
These techniques are
discussed in more detail later in this chapter (Section
5: Direct Utility Elicitation).
Examples of Preference-Based
Utility Measures
There are two general
approaches to ascribing utilities, or quality-of-life weights,
to health states.
- In direct utility
assessment, utilities are elicited from respondents using
methods such as TTO or SG.
- In contrast, in
health state classification systems the respondent's
health status is ascertained, and then utilities assessed or
derived mathematically from a reference group are mapped onto
that health state.
The latter method is
sometimes called indirect utility assessment, because the
quality-of-life weights are not re-derived from each new subject,
or multiattribute utility assessment, because it incorporates
multiple domains of health.
Section 5 describes
methods of direct utility assessment and health classification
systems.
|