This article originally appeared on the BeyeNETWORK
In my last article, I outlined some of the opportunities for implementing business intelligence in BioPharma and discussed whether or not business intelligence (BI) and clinical research are ready for one another. In this article, we explore the world of open source software and whether or not the industry is ready to commit to non-commercial software in its quest to create value.
There are a number of reasons that companies adopt open source technologies. These reasons include lower up-front costs, ability to see (and change) the source code, and better quality code because there is a community of developers continually solving bugs and fixing vulnerabilities and flexibility. However, let’s face it, in BioPharma the cost of the software is rarely the top concern of decision makers or influencers. In my experience, the top concerns for anyone involved in software selection are:
- Fitness of use (ability to meet business goals)
- Support (both commercial/vendor support and a community of users)
- Availability of service providers and/or skilled resources
- Flexibility in terms of its ability to easily integrate with other architectures and data systems.
In other words, what matters is the problem-solving ability of the solution, not whether it is open source or proprietary.
Definition of Open Source
Open source differs from commercial, off-the-shelf software (COTS) in that it is governed by a specific licensing model whereby the author(s) of the software freely distribute the source code along with the rights to redistribute it. The details of various open source licenses are more complicated, and there are literally hundreds of different licensing models out there, but generally open source means that the code is open and freely distributable. For more information on the meaning of "open source," see theOpen Source Initiative (OSI) definition. Examples of true open source software include the Linux operating system and the Apache web server.
BioPharma’s Open Source Roots
BioPharma also has a rich history of academic linkages, and many of its scientists hail from universities and even have joint appointments with these organizations. We know also that community-based innovations and open source consumers are common in both academia and BioPharma R&D.
There seems to be mounting evidence that U.S. Government agencies are adopting open source. The National Cancer Institute is adopting an open source stance for one of its largest projects for everything from clinical trials management, biospecimens, imaging, genome annotation, proteomics, microarrays, pathways, data analysis and statistical tools to data sharing, infrastructure, vocabularies and translational research. All of this is being built on an open, collaborative set of tools and technologies. Finally, the National Institutes of Health has adopted the perspective in their funding protocols that the software should indeed be open, freely distributable and the source code should be modifiable (PAR-07-344: Innovations in Biomedical Computational Science and Technology (R01) but make the point that they have no specific policy on open source per se.
Availability of Open Source Business Intelligence
Recently, we have seen the emergence of viable open source tools in the business intelligence space such as:
- Reporting/visualization: BIRT, Jasper, Pentaho
- Analytics: R, Weka
- Data Integration: Apatar, Pentaho Data Integration (KETTLE), Talend
From a total cost of ownership perspective, open source software can be very cost effective. Because of the lower up-front and lower residual fees, companies can spend more of their total budget on integration and customization, putting their value add into the solution. This also translates to fewer obstacles in getting started, since there is typically no up-front fee. This affords more consumers the opportunity to try it out in a proof of concept and actually see how well the software will perform in actual conditions and not rely on vendor promises.
Proponents of open source claim that feature/function advancement is quicker because there is a community of people providing code changes and because that community cares about solving the problems and wants to make sure that the software is right. Potentially, there are fewer bugs because the code and problems are more visible and vulnerabilities can be fixed more quickly. There is certainly a sense of community – the concept of a massive influx of ideas and contributions – that makes open source hard to ignore.
Others that have adopted this notion also speak to its flexibility and “no surprises” because there are no complex usage restrictions in the contracts, you don’t have to buy more modules if you want to integrate with other software, and many vertical software companies have appreciated the ability to OEM the open source code and focus on their specific value add.
Potential Objection: Validation
So what are the downsides of open source in BioPharma? Most business or clinically focused people, if asked, would say that open source software is not “validated” (if they even knew what open source was).
So let’s talk about the validation. As we know, BioPharma is highly regulated and, as an industry, is often reticent to adopt anything that others haven’t first adopted. Regulated companies follow others in making dramatic changes to their processes. Furthermore, they often look to vendors who are willing to stand up to the rigors of scrutiny by the FDA (i.e., validation) and prove that their software performs as expected. Software vendors need to have well documented software development life cycle (SDLC) processes and be able to demonstrate that their software has been formally validated and has withstood third-party audits.
When we talk about validation, what do we really mean? In most cases, it means that the vendor is willing to provide a guarantee that software works as advertised. The FDA defines validation as:
Confirmation by examination and provision of objective evidence that software specifications conform to user needs and intended uses, and that particular requirements implemented through software can be consistently fulfilled.”
There are no regulations against open source; and, as cited above, there are a number of FDA and NIH sources that are openly supporting the innovations created by the collaborative, community-based projects. We know that organizations will continue to seek the most cost-effective solutions and are now using a risk-based approach to determine whether or not open source represents greater value or risk in their programs.
Potential Objection: Support
In contrast to business and clinical stakeholders, on the IT side they usually know what open source is and would say – “As long as we can support it, it’s fine.”
From a support perspective, there are a number of companies that provide commercial support in the form of technical support, development and custom integration services around open source technologies. RedHat did this for Linux, and we’ve seen a growing number of companies step up and support open source including Pentaho (business intelligence, data mining, ETL), Apache, JBoss and REvolution. In fact, there are a number of these firms that provide validation instructions and/or guidelines for using these applications in a validated environment (see for example, Tomcat and R).
Potential Objection: Maturity and Stability
Another area of concern for any BioPharma is the level of maturity of the software as well as its overall stability. After all, we cannot have software that changes every time someone contributes to the code based on SourceForge. Again, commercial vendors play an important role in stabilizing the software and providing us with flexible release schedules.
Anyone who has had to deliver a validated software system into production will attest to the fact that we cannot cost effectively validate every time there is a bug fix or new enhancement in the software. But if we could ensure that the software was stable, mature, well documented and automated test suites have been run to ensure it’s stability and reliability (both integration and regression testing), then we should consider open source software just as viable an option as anything else on the market. Managing upgrades, hot fixes and patches are a management challenge even if the software is a commercial package.
Potential Objection: Intellectual Property Issues
Another concern that many companies have with open source is a belief that you cannot be assured that the source code is completely legitimate – that someone hasn’t borrowed from their company store to fit a piece of code or a developer has taken from another open source project that has a more restrictive licensing agreement. Most companies – BioPharma or not – want assurances that the code is legally acceptable through indemnification. Some may be concerned that open source may open up door for litigation in the case of problems – it’s easy to point fingers at open source. Commercial open source vendors offer support agreements today to address these concerns.
As mentioned earlier, the top concerns for BioPharma companies involved in software selection are fitness of use (ability to meet business goals), validation/certification, support (both commercial/vendor support and a community of users), availability of service providers and/or skilled resources and a flexible architecture and its ability to integrate with other systems to play the most important roles for decision makers. We know that open source is here to stay and recent examples of major investments in open source initiatives like MySQL and REvolution make it difficult to say that these companies will just go away. Most major companies use it, and the government is encouraging its use to drive innovation and to reduce costs. For those of us in BioPharma, the opportunities exist to take advantage of the innovation that is occurring. More choices for buyers and competition for vendors solving business problems will help drive down costs and improve operating efficiencies and effectiveness.
The concept of “open source” software has been around since the late 1980s and has become mainstream in back-office/data center/infrastructure-based operations. Industry analysts suggest that open source is now entering its next wave of adoption by coming to the forefront of business operations, moving from infrastructure to the application level (Gartner Open Source Summit, 2007, as cited in InfoWorld, 9/20/07). One of these critical applications is business intelligence. There are dozens of open source business intelligence projects.
So the challenge for us in BioPharma is to understand the opportunities and whether or not technologies such as open source business intelligence (BI) apply to our work in clinical research. Can and should BioPharma use open source BI now that it has entered mainstream and commercial support models are in place?