OR WAIT null SECS
We all know clinical research is complicated, and it’s not just the science. Phase III therapeutic trials at NCI cooperative groups require on average 769 steps, 36 approvals, and a median of approximately 2.5 years from formal concept review to study opening.1 Furthermore, trials requiring a development time of over a year are significantly less likely to achieve enrollment goals than those which start within 12 months of concept formulation.2
The scientific, regulatory, contractual, and patient safety obstacles to start a clinical study are difficult enough–technology should not be part of that list. All too often, a contributing factor to the delay in getting to first patient in (FPI) is the selection, validation, training, and implementation of electronic software systems for the study.
Open source software, software that is distributed freely and can be used with few restrictions, can help to ease the burden of conducting clinical research by:
What is open source?
Fundamentally, open source is a type of software license. While there are many different “flavors” of open source licenses, common characteristics include allowing people to freely review, modify, and distribute an application's underlying source code. Chances are, you use open source. If you have ever used an Android smartphone, Firefox web browser, the Mac operating system, Google, Amazon.com, or Facebook, you have used open source.
Large, active communities form around successful open source projects, making the software often times more secure and better performing that their proprietary alternatives. A study conducted by Accenture identified organizations using open source as listing quality (76%), improved reliability (70%), and security (69%) as the three primary benefits to open source. With so many developers constantly scrutinizing an application's underlying source code, it’s hard for bugs to hide. Since the code is shared among numerous parties it istypically well structured, avoiding the spaghetti-code syndrome that often arises when a single company develops a code base.
Organizations are often attracted to open source software because of the benefits of flexibility, increased innovation, shorter development times, and faster procurement processes. A recent Gartner survey of corporate IT groups found open source software and internally developed software constitutes 58 percent of a company’s software portfolio, representing a near tripling in enterprise open source adoption in the last five years. The increase in open source software has also resulted in an equivalent proportional decrease in the use of proprietary software.
Open source in clinical research
While open source is prevalent in many industries, it is still emerging in the field of clinical trials. One notable success is the OpenClinica electronic data capture platform, which is used worldwide by hundreds of sponsors, CROs, academic, and government institutes for all types of clinical research. According to Cal Collins, who helped initiate the OpenClinica project, “OpenClinica was born out of frustration with existing proprietary EDC systems that are inflexible, expensive, and impose obstacles to fast, efficient study start-up.”
Open source is also being used to bridge the gap between healthcare and clinical research. The software program Mirth functions as middleware that allows healthcare applications that “speak” HL7 to communicate with other systems that use other standards and protocols, such as CDISC.
The NIH’s National Cancer Institute has also made a significant commitment to developing open source software through their caBIG (cancer Biomedical Informatics Grid) program. The caBIG Clinical Trials Suite is a set of interoperable open source tools that cover a gamut of clinical research functions (lab, AE reporting, study calendaring, and patient registry).
Some projects such as Open Data Kit (mobile data collection on Android),11 (mobile case management for community health workers), and OpenXData (mobile CommCare data collection) are leveraging open source as a way to bring research data capture and public health surveillance technology to low resource settings. According to Jon Jackson, CEO of Dimagi, a company that has developed such research systems, “The advantages of open source to large organizations are often around flexibility. To smaller organizations, and research groups in resource constrained settings, the primary benefit is often one of accessibility and customizability.”
Open source in clinical research often exists in mixed environments. For instance, an open source application may run a proprietary database or operating system (e.g., Oracle or Windows) or a proprietary application may be run on top of open source databases and operating systems (e.g., MySQL, Postgres, or Linux).
Obstacles to the productive use of clinical trial software often begin with the evaluation process. The most valuable methods of software evaluation often involve experimentation and prototyping, which can be a challenge with proprietary software. Community forums and mailing lists also can also serve as a valuable way to obtain peer feedback and minor technical assistance while evaluating the technology.
The increased transparency of open source software promotes a thorough, unbiased evaluation of the technology, revealing both the good and the bad. This helps to know what you’re getting. While no software is perfect for every circumstance, knowing more about what you’re adopting before you commit increases the likelihood of a successful implementation and happy users.
A positive byproduct of easy evaluation and widespread experimentation is that it can lead to higher quality software. Allowing people to fully inspect the technology and put it through whatever tests they choose, provides invaluable feedback, which, when communicated back to the community, fuels improvements to the software.
Another significant barrier to adopting e-clinical software is the ability to efficiently configure and implement a study. Success here requires four key elements:
Many open source adopters build and validate their studies by reading and generating documentation and gathering piecemeal assistance from their fellow community members. While this can be a viable and cost effective approach for some, many organizations wish to rely on assistance from a commercial entity to minimize risk and maximize success when implementing open source software. As Red Hat, Inc. provides support for the Linux operating system, OpenClinica, LLC (my company) provides a commercially supported OpenClinica “Enterprise” edition. Other companies provide support for open source software used in clinical research, such as Recombinant Data (i2b2/ transmart), Mirth Corporation (Mirth), and Revolution Analytics for R-based pre-clinical and clinical applications including chemometrics, medical image analysis, pharmacokinetics, statistical genetics, and survival analysis.
It is typical for any widely adopted open source system to have multiple vendors and consultants supporting it. This creates a competitive services marketplace, helps to reduce vendor lock-in, and creates strong incentives for commercial open source vendors to deliver high quality services and support. Even without the assistance of commercial entities, “tech transfer”—as it’s known in the industry—can often be accomplished in just a few days.
Scaling and evolving
Most research teams operate more than a single study. Each study invariably has unique characteristics, and the needs of most teams constantly evolve. Open source licensing provides guaranteed freedom from becoming locked-in to a rigid operational model controlled by a single vendor. By increasing flexibility and choice, open source adopters ensure their ability to make the right decision for a given study and its circumstances.
The freedom and control benefits of open source are widely recognized. Even very large organizations, like the United States government, recognize the advantages of open source software over proprietary software. Casey Coleman, CIO of the US General Services Administration, cites some key reasons why they deploy open source software:
"By using open source, the agency won't be locked in to using a proprietary software program, at least for the duration of the contract. Not having sunk costs in a commercial software program also means the agency can move to a new program more quickly should its needs change. The general openness also means the agency could become a collaborator in the further development of the software itself. You get much more transparency and interoperability, and that reduces your risk."3
It can be particularly difficult for organizations operating smaller, shorter studies to achieve positive return-on-investment with proprietary software. For example, while it is generally agreed that EDC technology reduces the effort expended per patient by speeding up data entry and query resolution, the per-patient savings often do not offset the high up-front cost and overhead of traditional, proprietary EDC system in small and medium sized studies because there are fewer patients. High fixed costs are an obstacle to justify use of proprietary EDC and, as a result, these trials are stuck using slow, error-prone paper-based methods and legacy processes. Since the underlying technology is free to use and deploy however the sponsor or CRO sees fit, open source software offers adopters greater control over how they allocate their resources.
For example, having the ability to match the appropriate method of deployment and level of support with each study provides advantages for scaling. A research group doing a low budget Phase I study could use existing open source software for data collection and management and have the systems be rapidly procured, configured, and maintained with internal resources (study configuration, hosting, and system maintenance). Toward the other end of the spectrum, a larger, pivotal trial may pursue a “best-of-breed” services strategy, relying more on outside vendors who can deliver greater systems expertise, service level agreements, and more robust technical infrastructure around the open source software, while maintaining other key functions in-house.
Being able to leverage the same software across a spectrum of clinical trials (even though the way it is serviced and supported may change) offers clear advantages, allowing organizations to reap greater productivity from the knowledge, skills, SOPs, etc. they have developed.
This flexibility is possible because of the symbiotic relationship that exists in a professional open source model between the community, vendors, and customers.
The open source community of users and developers acts as a forge for continually refining and improving the platform. Commercial support makes the open source platform more easily consumable and appropriate for mission-critical operations, delivering packaged deployments, seasoned expertise, enhanced delivery capacity, and guaranteed service. Customers benefit by obtaining a well vetted technology and more flexible infrastructure while maintaining greater control over their destiny.
Open source and open standards
The combination of open source and open standards can be a powerful way to deliver improved flexibility, quality, and efficiency. Data standards provide uniform ways to represent information or processes according to a detailed specification. A standard is “open” when it is not encumbered by patent, cost, or usage restrictions, and, similar to open source software, open standards have fewer barriers to adoption and ongoing improvement than do proprietary standards.
Clinical trials are conducted within a complex, heterogeneous universe of healthcare and research data and systems. The Electronic Case Report Forms (eCRFs) housed and managed by an EDC system is at the hub of this universe. The FDA in its recent draft guidance, “Electronic Source Documentation in Clinical Investigation”, has defined the eCRF as the central system for integrating multiple feeds of data.
This FDA guidance, combined with increasing adoption of electronic health records (EHRs) and increasing demands on businesses to be more automated, accurate, and efficient, are driving the evolution in how we integrate data. The merging disparate datasets using SAS at the end of the study are slowly coming to an end, as are the days of days of point-to-point, proprietary interfaces for integrating applications. These approaches are brittle, costly, and do not scale well as third- or fourth-party systems that need to be added to the transaction. Secretive interfaces and file formats are quite valuable to proprietary vendors who require their consulting services be used every time a connection is made, but impose a huge obstacle for users and introduce unnecessary business risk. Even if these interfaces are included as part of the basic product, they still are not accessible to third-party developers, so it is impossible for a rich ecosystem of proven integrations to emerge.
Open source software leveraging both domain standards (e.g., CDISC) as well as technology standards (e.g., SOAP, REST) can provide a fresh and useful alternative. Any developer can use these specifications to perform integration, and often these integrations are shared freely with the rest of the community. The open source development and licensing model encourages experimentation, reduces “reinvention of the wheel,” and allows otherwise unaffiliated parties to build on the work of others. In this way, a community-driven, open source offering—harnessing open standards—can produce a more robust, innovative technology solution. The result is that open source becomes a key driver of increased IT efficiency, creating reusable processes and wringing out unnecessary costs. In many cases, users can have the best of all worlds: the flexibility to develop and extend their systems as they choose, the knowledge they are part of a open source community forging the next generation of clinical trial technologies, and the ability to reduce risk by obtaining proactive, commercial support.
Many of the principles of open source are manifest in open standards. However, is it possible for the principles of open source to extend even further, into the greater pharmaceutical R&D process? Industry luminary, Ken Getz, believes that “open innovation” may help to meet many of the most significant industry challenges, such as increasing development costs and times and decreased success rates. Getz believes that “Successful open innovation is about sourcing and leveraging expertise wherever it can be found and sharing knowledge and information as freely as possible.”
Even though you likely already use open source software on your smartphone or laptop, there are still many within the clinical trials industry that are just beginning to understand the advantages that the principles of open source can have on drug discovery. Those who recognize the significance of open source will be a step ahead of their competition and have the opportunity to lead in the race to produce new therapies, diagnostics, and knowledge that improves human health.
Benjamin Baumann is Co-Founder and Director of Business Development at OpenClinica.