PII-185 - A QUANTITATIVE MODELING APPROACH TO PREDICT AVAILABILITY OF GENERIC ORPHAN DRUG PRODUCTS.

A. Tong¹, J. Wang¹, W. Sun², M. Hu¹, L. Fang¹, M. Kim², L. Zhao¹; ¹Division of Quantitative Methods and Modeling, Office of Research and Standards, Office of Generic Drugs, Center for Drug Evaluation and Research, U.S. Food and Drug Administration, Silver Spring, MD, USA, ²Division Of Therapeutic Performance II, Office of Research and Standards, Office of Generic Drugs, Center for Drug Evaluation and Research, U.S. Food and Drug Administration, Silver Spring, MD, USA.

Background: Orphan drugs refer to those intended for the treatment, diagnosis or prevention of rare diseases that affect fewer than 200,000 people in the U.S.. The high price of orphan drugs is often a barrier for patients to access care. Generic competition can help reduce drug price and enhance drug accessibility; thus mitigate the accessibility issue. This study aims to examine the status of generic availability of orphan drugs and identify impactful factors on generic availability using quantitative modeling approaches.

Methods: In this work, the data regarding drug product information (e.g., drug complexity, dosage form, administration route, approval dates), regulatory factors (e.g., product-specific guidance (PSG) recommendation dates, and orphan drug exclusivity expiration dates), and pharmacoeconomic factors (sales data between 2011 and 2022) were collected for analysis and model building. Our dataset included 29 potentially relevant variables for 209 orphan reference listed drugs (RLDs) and their related generic products (if available). Summary statistics were used to examine the status of generic availability for orphan drugs. In addition, the oblique random survival forests method was used to identify the highly impactful factors and predict the probability of the first generic product availability at any given time point(s) of a given orphan RLD. The out-of-bag Harrell’s C-statistic was used to evaluate the model performance.

Results: Among the 209 orphan RLDs (approval dates ranging from 2008 to 2019), 16% of them have at least one generic product available. The annual drug sales (e.g., at 3 years after the RLD approval), drug product complexity and PSG availability were identified as the most impactful variables. The results of Harrell’s C-statistic show that the constructed machine learning model is a promising tool for predicting availability of generic orphan drugs.

Conclusion: In this study, three most impactful factors (drug sales, drug complexity, and PSG availability) associated with generic orphan product availability were identified and the model performance of predicting availability of generic orphan drugs was assessed. The model-informed results may be utilized in future intervention strategies development to promote the availability of generic orphan drug products.