One concept for accelerating the development of drugs for neglected diseases (i.e., those whose treatment is not reimbursed by insurance companies) is application of the “open source” innovation model which originated the software industry. This model is based on easy access to source code, distributed work among unaffiliated programmers, and rights to use (and sell) derivative products to produce usable and useful programs quickly and at low-cost (c.f., Open Source Initiative). Last week, the folks at the Center for Global Health R and D Policy Assessment, a subgroup of the policy contract firm, Results for Development Institute, made public their draft study of the topic, “Open Source for Neglected Diseases” (CGHRDPA Report) and invited comment. While the report is a good start, I opine that it doesn’t address the fundamental limitations of the open source approach as applied to drug development and, as I have written about the Center’s previous two draft assessments (my posting of 12/9/10), doesn’t reflect the years of experience of the biotech/pharma industry in playing the drug development game.
First, the good start is that the authors do a laudable job of covering a large and diverse field of ideas and efforts, and their summary of the open source “initiatives” on pages 6-9 and 19-20 is comprehensive but lacks, as the authors later note (page 17), an assessment of the initiatives’ output and cost effectiveness. The authors also missed two software platforms that are available and are being used for neglected disease drug development, specifically, HEOS by Scynexis (Scynexis and my posting of 11/4/10) and OpenClinica by Akaza Research (OpenClinica) and an initiative, the Distributed Drug Discovery Project at the University of Indiana/Purdue University (D3). The report also dwells too much on the supposed barriers of patents in early stage drug development, although the authors do note that there is considerable debate on this topic and the IP challenge may be a “theoretical problem.” I think patents on research tools and methods are unenforceable in early-sage research and, if the patent-holder attempts enforcement on products, may be found invalid (as with the University of Rochester vs. Pfizer, c.f., Science article). Moreover, I think that, if the patent was based on government-funded research, it should be licensed royalty-free for all neglected disease drug discovery and, if not, it should be ignored until public opinion stimulates reasonable licensing.
I disagree with the authors noting the utility of an open source approach to the later stages of drug development is “not clear.” I’d say “not relevant.” The open source approach is really only applicable to the discovery of drug candidates, or molecules that may result, after lots more testing in animal and humans, manufacturing process development, and regulatory approval of both, in a drug. Clearly this is the most expensive part of drug development and therefore requires the greatest investment and at least break-even economics and the open source approach accounts for neither. The authors mention a role of the approach in decreasing overall drug development costs (“A factor of five decrease in new drug costs, for example, shifts the landscape – and therefore the viability of collaborative and open source approaches“ on page 14), but do not offer any substantiation. Their analysis of these economics would be helpful. They also point out that Mauer et al. (Mauer et al. 2004) propose that the results of open source early-stage drug candidate discovery could feed into the neglected disease product development programs (PDPs), a logical idea but one that the PDPs haven’t supported over the intervening years, preferring, less effectively in my mind, to give grants to a small number of academic groups, typically those headed by members of the PDPs’ science advisory boards.
As an aside, I got a sense of the cost-efficiency of the tools available for drug discovery last fall when I met young man at the Medicines for Neglected Diseases Workshop who had used online structural data bases of targets and compounds to identify a series of candidates for treating dengue fever (my post of 9/16/10). Andrew Navia, still a junior at Lexington High school, did the work for a state science project competition (“Molecular Modeling of Viral Protease Inhibitors: Focus on Dengue Fever,” Mass S and E Fair), but then he may have had help from his father, Manuel Navia, a renowned structural biologist and founder of several companies (Navio Bio). Hey- I helped my kids with their science projects, too.
Another weakness of the open source approach that the authors do not address is that in effective drug development, at least as I experienced it, the early stages are informed and guided by continuous input from those responsible for the later stages, a positive feed back loop. In open source, the distributed workers are likely to follow their own direction since they are not really invested in the final outcome and so may be highly ineffective. Finally, as I mentioned above, I wonder if the opinions of industry drug developers may have been helpful. The authors point out there is a need for identifying “clear metrics of ‘value’” and a “business case” for the open source approach (page 16), but, other than referring to the now-dated and limited case studies by BIOVentures for Global Health, give no insights that may have been provided by those whose paychecks depend on finding drugs. I noted that of the 10 interviewees quoted in the report only two had industry affiliations. Finally, I found the report’s recommended next steps pretty timid. The first two (evaluating existing initiatives and developing value propositions) should be within the scope of the report and the third (a website for sharing) seems duplicative. My opinion is that the open source approach for neglected disease drug discovery has merit, in part because it piggy-backs on the tremendous public investment in biomedical research and enables participation in neglected disease research by many scientists who want to see their work result in social good but who are not rewarded for it. I also think the open source approach should be encouraged and supported by those that may benefit from its output (the PDPs and companies looking for new drug candidates) but need help to figure out how to make this happen.