Science & Research

Comparative Surveillance of Generic Drugs by Machine Learning

Performer: Marshfield Clinic Research Foundation

Principal Investigator: Peggy Peissig

Project Duration: 9/30/2015 – 9/29/2018

Regulatory Science Challenge

Generic pharmaceutical products play an important role in controlling health care costs. However, patients and physicians are concerned about potential changes in clinical outcomes and/or adverse drug events (ADEs) resulting from differences between branded drugs and their generic equivalents. Although FDA has rigorous processes to ensure that a generic drug is chemically and biologically equivalent to a safe and effective branded reference product prior to its being released to the market, these studies are not designed to detect differences with respect to clinical outcomes or adverse events. Requiring such clinical studies for the approval of generic drugs would increase costs associated with bringing a generic drug to the market and ultimately increase generic drug prices.

Although FDA has a diversified approach to ADE detection, it remains unclear whether events involving branded versus generic products can be distinguished using existing methodologies. The MedWatch Program is a voluntary reporting system for health professionals and the public to report serious reactions and problems with medical products. However, this reporting only captures a small percentage of the actual ADEs that are believed to occur.

A more efficient alternative solution for assessing possible differences in the efficacy or incidence of ADEs between generic and branded drugs is needed.

Project Description & Goals

Newer strategies for monitoring medical product safety have focused on existing data sources such as insurance claims data and electronic health record (EHR) data, which is a detailed digital version of a patient’s medical history. However, these approaches have not been used to find differences in ADEs associated with branded and generic products.

Our proposed approach will analyze data from EHRs and a corresponding research data warehouse that integrates hospital (inpatient) and insurance data on over 1.5 million active patients In the Marshfield Clinic healthcare system. The large patient numbers and drug exposure times will provide the sensitivity to detect ADE associations. This study's goals are to develop a surveillance system that compares generic drug experience to branded drug experience for early detection of any differences between them.

Innovative computer science approaches, such as machine learning (ML), will be applied to analyze text contained in the patient records and predict which patients may be at risk of suffering known ADEs following exposure to a generic medication. Machine learning approaches will also be used to detect ADEs and differences in drug effectiveness between populations taking generic and branded drugs.

Page Last Updated: 12/05/2017
Note: If you need help accessing information in different file formats, see Instructions for Downloading Viewers and Players.
Language Assistance Available: Español | 繁體中文 | Tiếng Việt | 한국어 | Tagalog | Русский | العربية | Kreyòl Ayisyen | Français | Polski | Português | Italiano | Deutsch | 日本語 | فارسی | English