Discrete-choice preference comparison between patients and doctors for the surgical management of oesophagogastric cancer

Background: Establishing preferences for surgery is paramount to preoperative patient counselling. This study aimed to prioritize and compare preferences of patients and doctors towards surgery for oesophagogastric cancer, to aid the counselling process.

Methods: A discrete-choice questionnaire containing hypothetical scenarios was designed to test patient preferences for six treatment attributes: mortality, morbidity, quality of life (QoL), cure rate, hospital type and surgeon's reputation. The survey was mailed to all patients who underwent oesophagogastric cancer resection from 2008 to 2009 at two teaching hospital sites. All doctors at these sites with previous experience in counselling patients for cancer surgery were also identified and presented with the survey. Results were analysed using a random-effects probit regression model. Spearman correlation was used to compare participants' implicit choices from the discrete-choice scenarios (their true preferences) with their explicit choices from the direct ranking of preferences (their perceived preferences).

Results: Eighty-one patients and 90 doctors completed the survey. Some 15 per cent of patients and 31 per cent of doctors based their choices solely on QoL. In order of importance, patients based their implicit responses on QoL (β = 1.19), cure rate (β = 0.82), morbidity (β = - 0.70), surgeon's reputation (β = 0.60), mortality (β = - 0.57) and hospital type (β = 0.26). Doctors similarly indicated QoL (β = 1.14) and hospital type (β = 0.31) as highest and lowest preferences respectively, but placed far greater emphasis on mortality (β = - 0.80) than morbidity (β = - 0.35). Implicit and explicit preferences correlated only for morbidity and surgeon's reputation in the patient cohort.

Conclusion: Clinicians may better meet patients' expectations and facilitate informed decision-making if QoL, cure rate and morbidity are emphasized foremost. A similar study employing preoperative patients is warranted for further clarification of preferences.