Session: Stated Preferences
Room: Upson B17
Time: Wed 10:15-11:45

Presenter: Ateesha Mohamed (RTI Health Solutions. Health Preference Assessment)

Discussant: Semra OzdemirUniversity of North Carolina, Chapel Hill

Abstract

BACKGROUND: When assessing the value of a new drug relative to existing drugs, regulators have to determine if the improvements in efficacy are offset by potential side effects. An example is triptans for migraine treatment that have become excellent drugs, and improvements in new triptans can be small. The added value of new triptans is especially sensitive to tradeoffs between small changes in efficacy and adverse events. We evaluate how the method of conjoint analysis can be used to inform decisions by regulators and other stakeholders about patients’ relative preferences for new triptans.
OBJECTIVE: To quantify patient preferences for favorable and unfavorable outcomes of migraine treatments and to assess whether the information informs decisions of relative value to other treatments.
METHODS: US residents ≥18 years of age with a self-reported diagnosis of migraine and currently using triptans completed a web-enabled choice-format conjoint survey instrument that presented a series of trade-off questions, each including a pair of hypothetical migraine medication profiles. Each profile was defined by eight attributes—(1) pain and sensitivity to light and sound 1 hour after taking the medicine, (2) pain and sensitivity to light and sound 2 hours after taking the medicine, (3) nausea and/or vomiting 24 hours after taking the medicine, (4) chance that the migraine returns within 24 hours, (5) nervous system side effects 24 hours after taking the medicine, (6) chest-related side effects 5 minutes after taking the medicine, (7) limitations on ability to do daily activities 24 hours after taking the medicine, and (8) one-year, medication-related, heart-attack risk. Each subject answered 13 trade-off questions based on a predetermined experimental design with known statistical properties. Random-parameters logit was used to estimate effects-coded preference weights for each attribute level.
RESULTS: 201 US subjects completed the survey. Preference weights for all eight attributes were consistent with the natural ordering of the categories such that more severe (worse) levels of an attribute had lower preference weights. Relief from severe functional limitations was the most important improvement, with a change from severe to no limitations more than twice as large as the change from any degree of pain to no pain at one or two hours. The most important improvement in functional limitation was the change from severe to moderate pain. Patients were also willing to accept considerable risk to avoid pain. For example, an improvement from severe to mild pain and sensitivity 1 hour after taking the medication was as important to patients as eliminating the chest-related or nervous system side effects. The desire to reduce pain and sensitivity 2 hours after taking the medication from moderate to none is so great that patients on average are willing to accept a one-year medication-related heart-attack risk of 0.9%.
CONCLUSIONS: This study indicates that patients are willing to accept chest-related and nervous system side effects, but not activity limitations, in exchange for significant improvements in efficacy. Preference weights derived from conjoint analysis can provide relative values that patients have for favorable and unfavorable treatment outcomes to inform decision-making.

Key Terms

Conjoint analysis, migraine, patient preferences

Authors:

Ateesha Mohamed (RTI Health Solutions. Health Preference Assessment) , A. Brett Hauber (RTI Health Solutions. Health Preference Assessment) , F. Reed Johnson (RTI Health Solutions. Health Preference Assessment) , Bennett Levitan (ohnson & Johnson Pharmaceutical Services) and Paul Coplan (Purdue Pharma)