Brand versus generic drug choice for management of a chronic condition

BACKGROUND: Rate of generic drug use increased substantially after the Hatch-Waxman Act, from 19% in 1984 to 88% in 2014. While change in the drug approval process allowed greater generic availability, barriers to greater generic adoption remain. Generic drugs are approved based on equivalence with brand in terms of active ingredient, strength, dose, and route of administration, yet skepticism exists among physicians, pharmacists, and patients that they are perfect substitutes. Survey research has measured mistrust of generics as highest among patients, with some variation across demographic and socioeconomic characteristics.

OBJECTIVE: This study will explore what patient and insurance plan characteristics are associated with actual choice of brand versus generic drug for treatment of a chronic condition.

DATA: Health insurance claims from a random sample of patients in a large commercial claims database were utilized. Drugs were selected which (i) were filled in both brand and generic forms across at least 9 months in 2013, (ii) had generic substitution rates of at least 15% but no higher than 95%, and (iii) were indicated for management of chronic conditions. Patients who filled the study drugs in 2013 and were continuously enrolled January 1, 2012-December 31, 2013 were included.

METHODS: Probability of filling a brand was estimated with a logit model, conditional on copay for the chosen product; copay imputed for the alternative as mean copay across the same product, generic status, and mail indicator order status for people in the same plan type and geographic division; the product filled; age; sex; race; education level; household income; region; Charlson Comorbidity Index; whether the patient had a health reimbursement arrangement (HRA) or health savings account (HSA); whether the product dispensed was modified from that specified by prescriber; and the fill number.

RESULTS: Of 5,760 products filled by patients in the sample in 2013, 21 had at least 9 months of fills for brand and generic, and 3 met the additional criteria for generic substitution rate and chronic use. There were 8,765 fills across two insulin products: isophane human insulin and 70-30 isophane and regular human insulin, both injections of 100 units/mL.  There were only 395 fills of the third product, albuterol sulfate, so the insulin sample was used for estimation. Among those taking insulin, older patients (p<0.01) were more likely to fill a brand. Females were 41% more likely to fill a brand (p<.05) while Whites were 59% more likely to fill a brand compared to Blacks (p<.05). With each fill, patients were 7% less likely to choose the brand (p<0.01). Patients with HSAs filled brands at lower rates as brand copays rose (p<0.01), but this result did not hold for those without such accounts.

CONCLUSION: This study suggests differing preferences for brand versus generic drugs across age, sex, and race. Results reveal one way HSAs may reduce healthcare costs as intended—through greater generic adoption as brand price rises. Results, however, rely on imputation of copays not observed. Future research should explore whether results are consistent for other drugs and in other populations.