The effect of Medicaid expansion on prescriptions for breast cancer hormonal therapy medications

Abstract

Objective: To quantify the effects of the Affordable Care Act Medicaid expansion on prescriptions for effective breast cancer hormonal therapies (tamoxifen and aromatase inhibitors) among Medicaid enrollees.

Data source/study setting: Medicaid State Drug Utilization Database (SDUD) 2011-2018, comprising the universe of outpatient prescription medications covered under the Medicaid program.

Study design: Differences-in-differences and event-study linear models compare population rates of tamoxifen and aromatase inhibitor (anastrozole, exemestane, and letrozole) use in expansion and nonexpansion states, controlling for population characteristics, state, and time.

Principal findings: Relative to nonexpansion states, Medicaid-financed hormonal therapy prescriptions increased by 27.2 per 100 000 nonelderly women in a state. This implies a 28.8 percent increase from the pre-expansion mean of 94.2 per 100 000 nonelderly women in expansion states. The event-study model reveals no evidence of differential pretrends in expansion and nonexpansion states and suggests use grew to 40 or more prescriptions per 100 000 nonelderly women 3-5 years postexpansion.

Conclusions: Medicaid expansion may have had a meaningful impact on the ability of lower-income women to access effective hormonal therapies used to treat breast cancer.

Keywords: Medicaid; breast cancer; differences-in-differences; prescription medications; public insurance.

Figure 1

Unadjusted trends in quarterly prescriptions per 100 000 nonelderly women for breast cancer hormonal therapy drugs: State Drug Utilization Database (SDUD) 2011 Q1 to 2018 Q3. Notes: Dataset is the State Drug Utilization Data (SDUD) 2011 Q1 to 2018 Q3. Breast cancer medications are tamoxifen, anastrozole, exemestane, and letrozole (brands and generics). The unit of observation is a state‐year‐quarter. Data are aggregated to the treatment‐quarter level. States with substantial Medicaid expansions prior to 2011 are excluded (see the online appendix). Data are centered around the expansion year for expansion states and are centered around Q1 2014 for nonexpansion states (referred to as the expansion date for nonexpansion states). Endpoint restrictions are applied: Periods more than 12 quarters in advance of the expansion date are excluded from the sample, and periods more than 18 quarters after the expansion date are excluded from the sample

Figure 2

Effect of Medicaid expansions on quarterly prescriptions per 100 000 nonelderly women for all included breast cancer hormonal therapy drugs using an event‐study: State Drug Utilization Database (SDUD) 2011 Q1 to 2018 Q3. Notes: Dataset is the State Drug Utilization Data (SDUD) 2011 Q1 to 2018 Q3. Breast cancer medications are tamoxifen, anastrozole, exemestane, and letrozole (brands and generics). The unit of observation is a state‐year‐quarter. The omitted category is one quarter prior to the expansion. All models are estimated with weighted least squares and control for state demographics, state fixed‐effects, and period fixed‐effects. State nonelderly female population serves as the weight. Data are centered around the expansion period (first full quarter) for expansion states. Nonexpansion states are coded as zero for all lead and lag indicators. Endpoint restrictions are applied: Periods more than 12 quarters in advance of the expansion date are excluded from the sample, and periods more than 18 quarters after the expansion date are excluded from the sample. States with substantial Medicaid expansions prior to 2011 are excluded (Table 1). 95% confidence intervals account for within‐state clustering