Conventional and chained standard gambles in the assessment of coronary heart disease prevention and treatment

The authors compared the abilities of descriptive and valuational health-re lated quality-of-life measures to discriminate healthy participants (n = 39 ) from those on diets for dyslipidemia (n = 35) and angina patients (n = 30 ). On the rating scale, the time tradeoff, and the General Health Perceptio n subscale of the SF-36 Health Survey, the participants with dyslipidemia o r angina reported lower mean scores than the healthy participants. No diffe rences were detected between these groups on conventional or chained standa rd gamble (SG) scales. The distribution of the conventional and the chained SG scores was very skewed, with the vast majority of scores being equal or very close to the maximum score. It is concluded that in this study the di scriminant ability of the chained SG was comparable to that of the conventi onal SG and inferior to descriptive and non-risky valuational scaling techn iques. This may be explained by the distortion of probabilities, by a misun derstanding of the SG chained assessment, and by a strong certainty effect.