Validity of conjoint analysis to study clinical decision making in elderly patients with aortic stenosis

Abstract

ObjectiveWritten case simulations are increasingly being used to investigate clinical decision making. Our study was designed to determine the validity of written case simulations within a conjoint analysis approach. Study design and settingWe developed a series of 32 written case simulations that differed with respect to nine clinical characteristics. These case simulations represented elderly patients with aortic stenosis. The clinical characteristics varied according to a fractional factorial design. We analyzed retrospectively all consecutive patients of 70 years of age or older with an aortic stenosis in three university hospitals. Results34 cardiologists from three Dutch hospitals gave their treatment advice to each of these case simulations on a six-point scale (ranging from ‘certainly no’ to ‘certainly yes’ to surgical treatment). We compared the influence that the clinical characteristics had on the responses to these case simulations with their influence on the actual treatment decision for 147 actual patients in the same three hospitals. We found a strong agreement. This agreement was only slightly affected by the cut-off value used to dichotomize the treatment advice into a recommendation in favor of or against surgical treatment. ConclusionWritten case simulations reflect well how clinicians are influenced by specific clinical characteristics of their patients.