America’s drug-approval process has set the global standard for rigorous safety and effectiveness review, but even with clinical trials and other safeguards, it is impossible to fully understand the impact of any particular medical intervention until it is widely used. Once on the market, drugs are further studied by pharmacoepidemiologists and other researchers, who work diligently to identify safety issues and potential unanticipated benefits. Many important discoveries have been made through this process, but researchers were always hampered by a reliance on voluntary reporting of problems and side effects as the primary source of data. As a result, there is a growing interest in the use of other sources of data generated within the healthcare setting. Unlike clinical trials, the use of observational data and methods to monitor medical product safety is challenged by the lack of accepted research methods and practices, leading to the well-known phenomenon of multiple studies of the same intervention yielding vastly different results.
In 2007, recognizing that the increased use of electronic health records (EHR) and availability of other large sets of marketplace health data provided new learning opportunities, Congress directed the FDA to create a new drug surveillance program to more aggressively identify potential safety issues. The FDA launched several initiatives to achieve that goal, including the well-known Sentinel program to create a nationwide data network.
In partnership with PhRMA and the FDA, the Foundation for the National Institutes of Health launched the Observational Medical Outcomes Partnership (OMOP), a public-private partnership. This interdisciplinary research group tackled a surprisingly difficult task that is critical to the research community’s broader aims: identifying the most reliable methods for analyzing huge volumes of data drawn from heterogeneous sources.
Employing a variety of approaches from the fields of epidemiology, statistics, computer science and elsewhere, OMOP took on a critical challenge: what can medical researchers learn from assessing these new health databases, could a single approach be applied to multiple diseases and could their findings be proven? Success would mean the opportunity for the medical research community to do more studies in less time, using fewer resources and achieving more consistent results. In the end, it would mean a better system for monitoring drugs, devices and procedures so that the healthcare community can reliably identify risks and opportunities to improve patient care.
Officially launched in late 2008 as a two-year pilot, OMOP worked to design experiments testing a variety of analytical methodologies in a range of data types to look for drug impacts that are already well known. The initial findings were inconclusive, and the team found more challenges than answers.
From this initial base, the project moved forward with additional research, and found greater success. Over the course of 2011 and 2012, research yielded greater confidence that particular methods used with particular types of data can reliably identify correlations between individual medical interventions and specific health outcomes. While there is still work to be done, the findings suggest meaningful progress toward the ultimate goal.
OMOP Web RL
The OMOP experimental process required an accessible but highly secure large scale computing environment to support the management and analysis of data accumulated from 100 + million patients. To address these requirements, the FNIH funded the development of the OMOP Research Lab, which was based on Amazon’s Elastic Cloud Computing (EC2) technology. The resulting OMOP Web RL Platform provided a mechanism to configure, deploy, and manage secure technical computing environments for organizations and individuals within the Amazon Elastic Computing Cloud (EC2) service.
The OMOP Web RL Platform consists of a series of software tools that enable user administration, security management, and secure user access to computing and data resources. The OMOP Web RL provided the OMOP team with a highly elastic but secure collaboration and data management platform to enable their research. The FNIH has placed the OMOP Web RL into the public domain. For more information on the OMOP Web RL Platform, please contact Ann Ashby at firstname.lastname@example.org or (301) 435-1641.
Having achieved its mission as envisioned by the founding members of the partnership, the OMOP program concluded at FNIH in June, 2013. FNIH placed all of the OMOP developed intellectual property into the public domain and transferred certain infrastructure and data assets resulting from the partnership to the Reagan-Udall Foundation for the FDA to be leveraged within initiatives supporting the ongoing development of the FDA Sentinel Initiative.
This includes the materials from OMOP’s methodological and data research, the OMOP Common Framework (Common Data Model and Vocabulary Specifications), and the OMOP Web Research Lab (Web RL) Cloud Provisioning Platform Software. The Common Framework continues to be used within the comparative effectiveness and drug safety research communities for building networks of observational healthcare data. The WebRL has been put into the public domain to help organizations build cloud-based data management and computationally rich computing environments for open science research initiatives.
Access to the Archived research materials can be found on this website and the open source WebRL software can be found within the OMOP Github repository. Additional information about Reagan-Udall Foundation and ongoing work in this arena can be found at www.reaganudall.org.