Severity or Incidence: Labeling Side-Effect Attributes in Discrete Choice Experiments

Area of Science:

• Health Economics
• Patient Preference Research
• Risk Communication

Background:

• Medical regulations often require incidence-based side effect labeling, but discrete choice experiments (DCEs) commonly use severity-based labels.
• This discrepancy may bias patient risk preferences and limit the external validity of DCE findings.

Purpose of the Study:

• To investigate how using incidence versus severity labels for side effects in DCEs affects patient preferences and the reliability of the experiments.
• To determine if label type influences maximum acceptable risks (MAR) and preference heterogeneity.

Main Methods:

• A discrete choice experiment (DCE) involving 1105 Dutch adults aged 60+ on influenza vaccination uptake.
• Randomized assignment of attribute labels to either severity (mild/severe) or incidence (very rare/very common) for side effects.
• Analysis using mixed and heteroscedastic logit models to assess MAR, preference heterogeneity, reliability, choice consistency, and predicted uptake.

Main Results:

• Severity labels, compared to incidence labels, increased MAR and preference heterogeneity for severe/very rare side effects.
• No significant differences were found for mild/very common side effects, reliability, or choice consistency.
• Respondents' perception of side effect severity did not always align with clinical severity labels.

Conclusions:

• Attribute label type (severity vs. incidence) influences MAR estimates and preference heterogeneity in DCEs, indicating potential priming effects.
• Predicted vaccination uptake was not significantly impacted by label type.
• Incidence-based labeling may more accurately represent real-world patient choice contexts, given potential misalignments in perceived severity.