Background: : With the increasing availability of biosimilars, understanding utilization dynamics is essential to optimize healthcare costs and outcomes. Trastuzumab biosimilars have shown significant potential to lower treatment costs while maintaining clinical efficacy for oncology patients. However, adoption rates vary considerably across different patient populations, healthcare organizations (HCOs), and geographic regions. Traditional analytics identify correlations but cannot determine which factors causally drive biosimilar selection, limiting actionable insights for managed care decision-makers.
Objective: : To establish causal relationships for drivers of biosimilar trastuzumab adoption patterns using advanced causal inference methodologies.
Methods: : We analyzed medical and pharmacy claims for patients initiating trastuzumab therapy between 2022 and 2024. Causal discovery algorithms identified relationships between patient characteristics (age, gender, comorbidities), treatment setting (duration, switching rates, dosing frequency, adherence, out-of-pocket (OOP)), total aggregated healthcare costs (inclusive of drug, hospital visits, ER visits, and physician fees), payer factors (formulary coverage, copay differentials), provider attributes (specialty, practice setting), margin percentage (reimbursed amount vs. Average Sales Price ASP), and biosimilar utilization and adoption. Treatment duration and adherence were also measured.
Results: : Among 26,547 eligible patients, 83% initiated biosimilars, with notable variation across payer channels (Originator vs Biosimilar—Commercial: 43% vs 48%, Medicare: 44% vs 34%). Descriptive analyses suggest that provider economics may be a key driver of biosimilar adoption. Preliminary findings show a median treatment duration of 5.1 months for biosimilar users versus 6.3 months for originator users. Adherence rates were 67% among biosimilar users compared to 55% for originator users.
Conclusions: : Causal AI analysis of real-world claims data highlights critical factors driving biosimilar utilization, such as formulary coverage, geographic and organizational practices, and reimbursement policies. Insights derived from causal modeling can inform targeted strategies to improve data-driven adoption rates, patient outcomes and optimize healthcare expenditure.
