Venue: The Fuqua School of Business, Duke University, 1 Towerview Drive, Durham, NC 27708-0120
Presentation
Are Asthma Patients' Patient-Reported Outcome (PRO) Importance Weights Consistent with Linear Scoring Rules? A Stated-Choice Approach
BACKGROUND: Patient-reported outcomes (PROs) are important measures of clinical benefit for many health-care interventions. The Onset of Effect Questionnaire (OEQ) is a PRO instrument used to assess outcomes in clinical trials of asthma patients on prescription combination maintenance therapy. The OEQ produces scores for five domains: subject could tell the medication was working, subject could feel the medication begin to work right away, subject felt physical sensations shortly after taking the medication that reassured him/her it was working, the medication worked as quickly as the subject's rescue medication, and subject was satisfied with how quickly he or she felt his/her medication begin to work.
PRO summary measures often score patient responses by averaging Likert responses over all the parts of the elicitation instrument. This calculation presumes that differences among Likert categories are equally important to patients within and across outcomes. If this assumption is incorrect, a weighted score would be more appropriate. The maintained hypothesis in most applications of economics and utility theory is that preferences are non-linear.
OBJECTIVE: The aim of this study is to compare a linear scoring rule with the subjective importance of different domain and categories of the OEQ among asthma patients on prescription combination maintenance therapy.
METHODS: We used a pretested stated-choice conjoint analysis survey instrument to elicit patient preferences for varying categories of the five domains measured in the OEQ. In addition to the five OEQ domains, we included a monthly out-of-pocket cost attribute that ranged from $35 to $70. A total of 50 asthma patients completed the pilot survey instrument. All subjects were at least 18 years of age, currently taking prescription combination maintenance medication, able to read and understand English, and provided informed consent. Survey subjects were randomly sampled from a large internet panel. We report the results of the pilot study here.
RESULTS: Our pilot study results indicate strong non-linearities both within and across domains and categories. An improvement from 'neither agree nor disagree' on 'During the past week, you were satisfied with how quickly you felt your maintenance medication begin to work' to 'strongly agree' is 2.5 times as important as the same improvement on 'During the past week you could tell that your maintenance medication was working'. We are currently analyzing the results from the full study of 509 asthma patients and will present these findings at a later date.
CONCLUSION: Our pilot results indicate large differences in the impact of individual domains and categories on patient perceptions of clinical outcomes. Thus, linear scoring rules may overstate or understate patients' perceived changes in outcomes. Our pilot results indicate that preference measures may represent a more accurate interval scale within PRO domains than the often standard assumption that the distances between categories are equal.