Comparing Discrete Choice Experiment with Swing Weighting to Estimate Attribute Relative Importance: A Case Study in...

Area of Science:

• Health Services Research
• Patient Preference Research
• Health Economics

Background:

• Discrete Choice Experiments (DCE) are widely used for patient preference elicitation but are complex and costly.
• Simpler methods like Swing Weighting with Direct Rating (SW-DR) exist, but empirical comparisons are limited.
• Understanding the relative importance of treatment attributes is crucial for shared decision-making.

Purpose of the Study:

• To directly compare attribute importance rankings and weights derived from DCE and SW-DR.
• To assess the usability and respondent experience of both methods.
• To provide guidance on selecting appropriate preference elicitation methods.

Main Methods:

• A total of 307 non-small-cell lung cancer patients in Italy and Belgium participated in an online survey.
• Preferences were assessed using both DCE and SW-DR methodologies.
• Attribute importance was analyzed using random parameter logit (DCE) and rank order centroid (SW-DR), with comparisons via Cohen's weighted kappa and Dirichlet regression.

Main Results:

• Most (>65%) patients found both DCE and SW-DR tasks easy to understand and complete.
• Survival was the most important attribute regardless of the method used.
• Significant differences were observed in aggregate attribute rankings (P < 0.01) and a greater spread in attribute weights from DCE compared to SW-DR (P < 0.01).
• Moderate agreement (weighted Kappa 0.53-0.55) was found in individual-level attribute rankings between the two methods.

Conclusions:

• Both DCE and SW-DR can determine attribute importance, but significant differences exist in their outcomes.
• While generally easy to use, the methods yield distinct results regarding attribute weight and ranking.
• Further research is needed to guide the selection of preference elicitation methods for various decision-making contexts.