Understanding Valuation of Tyrosine Kinase Inhibitors Among Patients with Chronic Myeloid Leukemia Using a Novel Hypothetical Purchasing Task

Introduction: The discovery of tyrosine kinase inhibitors (TKIs) dramatically improved survival among patients with chronic myeloid leukemia (CML). Despite the benefits, concerns remain regarding the financial impact to patients who often cover significant co-pays for their TKI. Currently, imatinib is the only TKI available through generic formulation. However, as more TKIs become generic while others remain on-patent, patient preference for lower cost will become an increasingly important factor in drug choice. Hypothetical purchasing task questionnaires (HPTs) are used in behavioral economics to study motivation for consumption of products under rising cost constraints. These tasks provide insight into consumer demand. To date, this method has not been applied to assess patient demand for expensive cancer therapeutics. Purpose: The purpose of this study was to trial a novel, CML patient-directed HPT to study valuation (i.e. motivation to purchase) of TKI therapy. We hypothesized that socioeconomic factors (marital status, income, education level, race) and taking imatinib vs. later generation TKIs would be associated with TKI valuation. In an exploratory analysis, we assessed whether lower valuation of TKI therapy is associated with financial toxicity (FT) and TKI nonadherence by CML patients. Methods: Patients treated at Ohio State for a diagnosis of CML in chronic phase who had been prescribed TKI therapy for at least 1 year were recruited. Participants completed surveys assessing demographics, financial toxicity (Comprehensive Score for Financial Toxicity) and TKI adherence (Morisky Medication Adherence Scale). They then completed the novel CML HPT, ranking the likelihood of taking their TKI as prescribed on a scale of 0%-100%. This task was repeated 11 times, each with an increasing increment of out-of-pocket monthly cost (range $0-$2000). Patients were asked to assume their current TKI was the only available therapy. “Valuation” breakpoints were defined as the monthly, out-of-pocket cost at which patients reported themselves 0% (breakpoint0) and first time less than 50% (breakpoint50) likely to take their TKI as prescribed. Wilcoxon rank sum test or Fisher's exact test was used to evaluate the associations between variables of interest depending on data types. Correlation coefficient (r) was used to measure the relationship between TKI evaluation and financial toxicity values. Results: Forty-one patients (age 19-80, 51% male) participated. 85% of patients were taking 2nd or higher generation TKIs. Additional demographics were as follows: 66% married, 29% non-Caucasian, 63% had less than a 4-year college degree, 37% earned < $50,000 annually. Median breakpoint0 and breakpoint50 costs were $500 and $200, respectively. Lower annual income (<$50,000) and non-Caucasian race were associated with lower TKI valuation at which patients reported 0% likelihood of taking their TKI as prescribed ($110 vs. $1000, p= 0.007 and $35 vs. $1000, p=0.03, respectively). Marital status (p=0.06), < 4-years of college (p=0.053) and taking imatinib (p=0.16) showed no significant association. Income, race and < 4-years of college were associated with lower breakpoint50 (p=0.015, 0.02 and 0.048, respectively). Lower TKI valuation was not correlated with FT (r 0.17) or associated with TKI adherence (p=0.42). Conclusions: As the TKI market diversifies, HPTs may provide a unique tool to study patient demand for these life-saving drugs. Adapting HPTs to study the cancer populations requires continued refinement. In our cohort of CML patients, the median monthly out-of-pocket cost at which patients stated they would not take their TKI (despite no other available treatment) was $500, which is well within the co-pay range required for many patients. While limited by a small N, this study demonstrates that in current US healthcare system, even life-saving therapies have a price at which they lose their inherent value to the consumer- particularly among patients of lower income and non-Caucasian race. The lack of association between TKI valuation, FT and medication adherence may reflect some benefit of medication assistance programs, although the study of both FT and medication adherence is highly complex and lack robust measurement tools. Future studies will adapt the CML HPT to study patient preferences between cheaper and less effective vs. expensive but more potent/well-tolerated therapies.