BACKGROUND: Over 15,000 new cases of CLL are diagnosed each year in the U.S. Different treatment options exist, which vary in efficacy, side effects, and mode of administration. Little is known about the value patients place upon the different attributes of available CLL treatments.
OBJECTIVES: To estimate patient preferences for CLL treatments and investigate the relationship between treatment preferences and cost.
METHODS: Patients with a self-reported physician diagnosis of CLL were recruited through a patient advocacy group to complete an online discrete-choice experiment (DCE) survey. The survey included 8 DCE questions in which respondents chose between pairs of hypothetical treatments for CLL. Each hypothetical treatment was defined by 5 attributes with several predefined levels; the attributes were progression-free survival (PFS; 10-60 months), diarrhea (none to severe), chance of severe infection (0-30%), chance of organ damage (0-8%), and mode and schedule of administration (pill versus intravenous administration). Treatment duration was not addressed independently and treatment cost was not included in the DCE. Random-parameters logit (RPL) was used to analyze the data and estimate patient preference weights for each attribute level in the survey. An importance score and minimum acceptable benefit score were calculated using results from the RPL model. Later, a cost attribute was added that included 2 hypothetical treatment profiles (Treatment A: 26 months of PFS, no diarrhea, 15% chance of serious infection, 8% chance of organ damage, IV once a month for 6 months, and lower cost; Treatment B: 60 months of PFS, mild/moderate diarrhea, 30% chance of serious infection, no chance of organ damage, daily pill for 60 months, and higher cost). Results from the RPL model were used to predict the share of respondents who would choose each treatment based on the DCE questions when cost was not a consideration, as compared with respondents' choices when cost was included.
RESULTS: 384 respondents recruited through LLS: 53% were female, 94% were white, and the mean age was 65 years. 57% had Medicare insurance, 53% of respondents had received financial aid and 40% reported difficulty in paying out-of-pocket costs for their medicines. In the DCE results, the attribute levels were ordered as expected, with respondents preferring better outcomes to worse. Within each attribute, all levels were statistically significantly different from each other (P :5 0.05). Attributes in the order of importance to patients as determined by DCE were PFS, chance of infection, chance of organ damage and the occurrence and severity of diarrhea. Mode of administration was the least important attribute. Although PFS had the highest importance score, the risk of adverse events was also highly important to patients; significant additional PFS was needed to offset patients' acceptance of worsening adverse events. For example, an increase in the chance of infection from 0% to 30% required an offset of 36 months of additional PFS. An increase in the chance of organ damage from 0% to 8% required an additional 26 months of PFS by the respondents. A tradeoff of 22 additional months in PFS is needed for a change from no diarrhea to severe diarrhea. Using results from the RPL model, it was predicted that 91% of respondents would choose Treatment B when cost is excluded from the treatment profile description. This was mainly driven by the better PFS attribute of Treatment B. When a modest cost difference was introduced ($75 more per month for treatment B), half of the patients chose treatment A. When this difference became larger ($400 more per month), the change in patient preference was even more pronounced; 74% chose option A.
CONCLUSIONS: The most important attribute in the survey was PFS, however patients also indicated that increases in the risk of adverse events would require significant increase in PFS to offset that risk. Patients with CLL are sensitive to treatment costs and many reported difficulty in paying for their medications. Given the choice of treatments available to patients and the high cost of some treatments, physicians may want to explore their patients' preferences for different treatment features, including benefit-risk tradeoffs and out-of-pocket cost when selecting the best treatment strategies for patients.
