Standard Gamble Questions Research Articles

10.1162/153244304773633843

We present an approach to elicitation of user preference models in which assumptions can be used to guide but not constrain the elicitation process. We demonstrate that when domain knowledge is available, even in the form of weak and somewhat inaccurate assumptions, significantly less data is required to build an accurate model of user preferences than when no domain knowledge is provided. This approach is based on the KBANN (Knowledge-Based Artificial Neural Network) algorithm pioneered by Shavlik and Towell (1989). We demonstrate this approach through two examples, one involves preferences under certainty, and the other involves preferences under uncertainty. In the case of certainty, we show how to encode assumptions concerning preferential independence and monotonicity in a KBANN network, which can be trained using a variety of preferential information including simple binary classification. In the case of uncertainty, we show how to construct a KBANN network that encodes certain types of dominance relations and attitude toward risk. The resulting network can be trained using answers to standard gamble questions and can be used as an approximate representation of a person's preferences. We empirically evaluate our claims by comparing the KBANN networks with simple backpropagation artificial neural networks in terms of learning rate and accuracy. For the case of uncertainty, the answers to standard gamble questions used in the experiment are taken from an actual medical data set first used by Miyamoto and Eraker (1988). In the case of certainty, we define a measure to which a set of preferences violate a domain theory, and examine the robustness of the KBANN network as this measure of domain theory violation varies.

Quality of life, health-state utilities and willingness to pay in patients with psoriasis and atopic eczema.

Skin diseases have been shown to have a significant adverse impact on the health-related quality of life of patients that may be underestimated by objective assessments of clinical severity. The main aim of this study was to measure the health-state utilities on a scale between 0 (dead) and 1 (full health) of patients with psoriasis and atopic eczema, and to measure the willingness to pay for a cure for psoriasis and atopic eczema. A second aim was to analyse how these measures are related to different dimensions of health-related quality of life, as measured by general and disease-specific quality of life instruments and a subjective measure of disability activity. This study was based on data from a questionnaire administered to, and interviews conducted with, 366 patients with psoriasis and atopic eczema aged 17-73 years, attending the dermatology outpatient clinic in Uppsala, Sweden from November 1996 to December 1997. The survey included: a rating scale question, a time trade-off question, a standard gamble question, a dichotomous choice willingness to pay question, a bidding-game willingness to pay question, a generic quality of life instrument (SF-36), a disease-specific quality of life instrument (the Dermatology Life Quality Index) and a subjective measure of disease activity (on a visual analogue scale). The mean health-state utility was 0.69 (rating scale), 0.88 (time trade-off) and 0.97 (standard gamble) for patients with psoriasis. The corresponding health-state utilities for patients with atopic eczema were 0.73, 0.93 and 0.98. On average, patients were willing to pay between 1253 and 1956 Swedish crowns (SEK) per month for a psoriasis cure and between SEK 960 and 1083 per month for an atopic eczema cure ($1 = SEK 8.25 and pound1 = SEK 13.23). The health-state utilities were related to SF-36, the Dermatology Life Quality Index and disease activity in the expected direction and the correlations were strongest for rating scale and weakest for standard gamble. The willingness to pay was correlated with the Dermatology Life Quality Index and disease activity, but not with SF-36. The study indicates that it is feasible to measure health-state utilities and willingness to pay in this patient population, and the sizeable willingness to pay suggests that skin diseases are associated with substantial reductions in quality of life.

What price information? Modelling threshold probabilities of fetal loss

This paper is an extension of previous work in which an alternative method of measuring the benefits of antenatal screening was proposed. The method is based on the elicitation of threshold probabilities of fetal loss at which women would be indifferent between having and not having an amniocentesis for the prenatal diagnosis of fetal abnormalities. The aim of this paper is to extend the previous work by modelling the preferences of a larger sample of women and investigating the consistency and validity of their responses. The threshold probabilities are elicited using standard gambles and modelled using Tobit estimation. The results indicate that it is possible to model these probabilities in this way and that it is possible to obtain a high degree of consistency in response to standard gamble questions. While establishing the validity of the responses is more difficult there is some evidence that the technique can provide valid responses.

Deriving a Preference-Based Single Index from the UK SF-36 Health Survey

This article presents the results of a study to derive a preference-based single index from the SF-36. The study was an attempt to reconcile a profile health status measure, the SF-36, with the “quality adjusted life years” approach. The study undertook a parsimonious restructuring of the SF-36 using explicit criteria to form the SF-6D health state classification. A sample of multidimensional health states defined by this classification were valued by a convenience sample of health professionals, managers, and patients, who responded to a set of visual analogue scale ratings and standard gamble questions, with highly complete and consistent answers. Statistical models were estimated to predict single index scores for all 9000 health states defined by the new classification. The resultant algorithms can be applied to existing SF-36 data sets and used in the assessment of the cost-effectiveness of health technologies. This preliminary work forms the basis of a larger study currently being undertaken in the UK.

On the Contingent Valuation of Safety and the Safety of Contingent Valuation: Part 2 - The CV/SG "Chained" Approach

This article reports the results of a study aimed at estimating a willingness-to-pay based value of statistical life for road risks using a multi-stage approach which involves “chaining together” responses to contingent valuation and standard gamble questions. The rationale for employing a multi-stage approach is to break the wealth/risk of death trade-off down into a number of conceptually manageable steps, thereby trying to attenuate the various biases that appear to be pervasive in responses to more direct contingent valuation questions in the health and safety field.

Evaluating the benefits of antenatal screening: an alternative approach

This paper presents an alternative method of valuing the benefits of antenatal screening programmes based on individual preferences. The methods of benefit measurement used to date, namely averted costs, QALYs, and willingness to pay, are critically reviewed. The alternative method assumes that screening can be regarded as an investment in improved information, and that the benefits of screening can be measured by the value placed upon this information. The information is valued by presenting individuals with standard gamble questions framed in terms of a prenatal diagnostic choice. The approach is developed in the context of antenatal screening for Down's syndrome and for cystic fibrosis carrier status. Some empirical testing of the approach is reported for the latter.