Patient preferences and linear scoring rules for patient-reported outcomes

Abstract

Background: Many patient-reported outcomes (PRO) instruments are scored by averaging or summing Likert category values over all items or domains of the elicitation instrument, yielding domain-specific scores or a total score for the entire instrument.

Objective: To evaluate differences between conventional linear and preference-weighted scores for PRO instruments used in asthma, oncology, and obesity.

Methods: We estimated preference-weighted scores for all the items and response categories in the Onset-of-Effect Questionnaire (OEQ), the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ)-C30, and the Impact of Weight on Quality of Life Questionnaire-Lite version (IWQOL-Lite) using choice-format conjoint analysis, known also as discrete-choice experiments.

Results: Conventional linear scoring rules can overstate the relative importance to patients of improvements in some domains and understate the relative importance of improvements in other domains.

Conclusions: Patient preference-weighted scores estimated by conjoint-analysis methods allow for non-linearities and account for the relative contribution of individual items and domains to patient well-being. Conventional linear scores and preference-weighted scores can result in different conclusions about the size of patient-reported treatment effects.