Macular Degeneration Drug Prescribing Patterns After Step Therapy Introduction in Medicare Advantage

Abstract

Importance: In Medicare Advantage (MA), step therapy for physician-administered drugs is an approach to lowering drug spending. The impact of step therapy in MA on prescribing behavior and the magnitude of any changes has not been analyzed.

Objective: To evaluate the impact of step therapy on macular degeneration drug prescribing patterns for 3 large MA insurers.

Design, setting, and participants: This was a retrospective encounter-based analysis using 20% nationally representative MA outpatient and carrier encounter records for 2017 to 2019. Participants were MA beneficiaries who were 65 years or older and had received a macular degeneration drug administration. Macular degeneration drug administrations for beneficiaries of MA Aetna, Humana, and UnitedHealthcare (UHC) insurers were assessed. Humana implemented macular degeneration step therapy in 2019, setting bevacizumab as the plan-preferred drug, and aflibercept and ranibizumab as the plan-nonpreferred drugs. Aetna and UHC, which did not implement macular degeneration step therapy, served as the control group. Data analyses were performed from May 2024 to December 2024.

Exposures: A macular degeneration drug administration subject to a step therapy policy.

Main outcome and measures: A binary indicator of whether the drug administered was bevacizumab. Linear probability models and a difference-in-differences framework were used to quantify changes in prescribing patterns before and after the introduction of step therapy for MA insurers that did and did not implement step therapy. To empirically measure the impact of step therapy, the first administration of a treatment episode was assessed, followed by switching patterns.

Results: A total of 18 331 MA beneficiaries, 21 683 treatment episodes, and 171 985 drug administrations were included across the control and treatment groups. The difference-in-differences regressions found a 7.8% (95% CI, 4.9%-10.7%; P < .001) greater probability of being prescribed bevacizumab for the first administration due to step therapy. The predicted probabilities of preferred-drug administration in the treatment group increased from 0.61 to 0.70 between the periods before and after step therapy implementation for the first administration. Step therapy was not significantly associated with an increased rate of medication switching (hazard ratio, 0.86; 95% CI, 0.71-1.06; P = .15).

Conclusions and relevance: The findings of this retrospective encounter-based analysis indicate that step therapy is associated with a greater probability of prescribing the plan-preferred drug for the first administration. The analysis failed to find a statistically significant greater rate of medication switching within a treatment episode. Step therapy changed macular degeneration prescribing patterns, but step therapy alone did not transition all administrations to the plan-preferred drug.

Figure 1.. Unadjusted Percentage of Bevacizumab Administrations, 2017 to 2019

X-axis refers to the period from 2017 quarter (Q) 1 through 2019 quarter 4. The vertical line indicates the quarter for which step therapy was implemented for the treatment group, 2019 quarter 1. The labels indicate the percentage of bevacizumab drug administered for that quarter, defined as the number of bevacizumab administrations divided by the total number of macular degeneration drug administrations. The control group includes Aetna and UnitedHealthcare drug administrations, and the treatment group includes Humana drug administrations.

Figure 2.. Event Study of Step Therapy on Bevacizumab Prescribing at the Annual Quarter Level

Excludes 2018 quarter (Q) 4 because that time period directly preceded the implementation of step therapy. In 2017 and 2018, prescribing of bevacizumab was relatively constant, with the exception of 2017 quarter 4. After implementation of step therapy, the prescribing of bevacizumab increased.

Figure 3.. Coefficient Plot for Primary Specification and Sensitivity Checks

Coefficient on the interaction term for the primary specification, which was an unadjusted model with treatment episodes specified with a 150-day washout period. Results of the sensitivity checks, including treatment episodes specified with a 90-day washout period, 365-day washout period, the primary specification with facility fixed effects included, the primary specification with covariates included (Supplement 1 provides covariate information), and when subset only to data with high levels of data reporting. The last specification included only years 2016 through 2018, when step therapy was not applied.