A Global (Biomedical) Research System

Add New Comment

An International Rapid-Learning System: Biomedical Research

     Several countries are beginning to organize huge, online, research databases including hundreds of millions of electronic health records, with sub-sets of these records to include genomic, protein, environmental, behavioral, and other data for each patient.<1> For example, the US government recently appropriated $19 billion.

     The World Academy may be able to play a useful role to network discussions and evolve individual national systems into a new global research system for rapid-learning biomedical research. I suggest that the President designate a Working Group to consult with colleagues and identify useful projects. 
     Two key steps are: 1.) the decision of national governments to participate; and 2.) "Grand Strategy," designating scientific priorities and including useful benefits for many groups - for example, researchers in academic and corporate settings as well as physicians and patients who want to query the databases to assist with such medical decisions as the improved matching of drugs and patients.
     - It would be attractive to design an adaptive global system, open to multi-disciplinary input and that also can be used by advocacy groups. For example, reference databases being developed by the Kaiser system (California) and the Robert Wood Johnson Foundation will include data on the proximity of public parks: there are initial studies that parks encourage exercise and significantly reduce stress; with 50% of humankind now living in urban areas, such data could be useful and informative.
     It would be attractive for the global research system for rapid learning to include online analysis software and free computer time for individuals and small companies from all countries.
<1> The national systems also have protocols and standards for individual privacy rights. (The Wikipedia entry for "Electronic Health Records" has references.) The HMO Research Network has developed software for translation and querying across large databases (www.hmoresearchnetwork.org): Uniform data systems are not necessary for a global database.


6-9 months v. 8-10 years

 "Orphan" Diseases: 6-9 months v. 8-10 years
     There is an emerging unanswered question about how much different nations should spend for electronic health record (EHR) systems for biomedical research, and with what strategies and priorities? One purpose of a WAAS working group would be to move this discussion forward. Two examples:

1.) Orphan Diseases

    There are about 6,000 "orphan" or rare diseases (defined as an estimated prevalence of about 70,000 cases or less per 100 million). Typically the number of patients is too small to interest drug companies to develop new drugs and pay for clinical trials. Patients usually are treated by off-label use of medicines, with physicians guessing about a range of treatments.

     These patients can have new hope and benefit directly from a new international research system using electronic health records (EHRs). The new database will be large enough for refined statistical analysis for many, and perhaps most, of these diseases. (Extrapolating from US data (5%) suggests that early participation by China and India can add 120 million patients in these categories.]<1>

     Even at the first step (as soon as their EHR identification numbers for diseases are established, in 6-9 months), physicians or patients (or the families of patients) in participating countries can activate pop-up menus with information about international networks of researchers, research programs, and physicians; the range of treatment options that have occurred to other physicians; and the emerging statistical evidence of comparative effectiveness.

2.) 6-9 Months v. 8-10 Years.

      It will be very helpful to have international requests for a public database and free online analysis software and computer time (available to anyone, worldwide).

      There are high-level current discussions of "8-10 years for useful results" from EHR systems, based on established national research systems - i.e., x months to write and circulate Requests for Proposals; application processing with competitive peer review processes; ramp-up time for new research projects and teams; the research time; lags for article preparation, peer review, and publication; time for articles (or the preprints) to reach other researchers and start new cycles of grants for new hypotheses or replications, etc.! [And this 8-10 years estimate is only for known, major diseases, with their large N of full-time researchers.]

     By contrast, with a light touch of leadership, a freshly-thought international strategy for in silico capabilities (open database, free online software and computer time for all science-trained people in every country who want to test hypotheses and explore data, 24x7, at the speed of thought) will move everything much faster, for everyone.

     - International requests for US government cooperation are processed in different bureaucratic channels, with additional participants who are likely to want a shift from 8-10 years to 6-9 months. For patients, speed also is more important than the minimal cost of computer processing time, even (perhaps) supercomputer time, which continues to fall in cost.


<1> It is risky to extrapolate without knowing the causes of these rare diseases. The US definition of a rare/orphan disease is an estimated prevalence of less than 200,000 in a US population of 300 million. The US cumulative estimate, which includes conditions that may not be diseases (e.g., genetic conditions) is about 8%. I have used 5% in this example, because the populations of China and India (2.4 billion) are younger and possibly healthier. For current knowledge available to US researchers at NIH see http://rarediseases.info.nih.gov.