Yesterday’s blog post focused on the value and future of comparative effectiveness analysis; today's second part of our interview with Lynn explores how the Archimedes predictive model has implications for comparative effectiveness research. You’ll remember that last October, Pioneer awarded an $15.6 million grant to Archimedes to support the development of ARCHeS, (ARChimedes Health care Simulator), an online interface and delivery system that will enable many more health experts to use the Archimedes model to answer health care questions with greater speed and precision.
How does ARCHeS and the Archimedes predictive model relate to comparative effectiveness research?
I see three major advances from Archimedes and its efforts to expand access to the model through the ARCHeS online delivery platform. First, more widespread use of the model will allow us, quickly and efficiently, to make confident, science-based predictions about treatment options. Archimedes founder David Eddy says it’s like building a house. If we can establish strong foundations at key knowledge points (e.g., through clinical trials), evidence-based predictive models can reliably fill in what hasn’t been studied directly.
Second, Archimedes can incorporate vastly more information—from physiology, systems biology, clinical studies and databases—about patients with complex, multiple conditions, and often on multiple treatment regimens, than even the best clinician can have in his or her mind and use well.
Third, Archimedes achieves a quantum advance in clinical care as a science. The leading-edge sciences, such as physics, start with careful experimental studies (like clinical trials in medicine) – and then use these results to build predictive models. They then use the models to predict observational data, and learn from how well these models work to design new experiments and develop better theories and models for prediction. The ultimate goal of other sciences is the ability to predict observed reality.
In the same way, accurate predictions of treatment options for patients should be the goal of clinical science. But clinical research is now focused mostly on perfecting and using the “randomized clinical trial” method, i.e., collecting careful experimental data. To move forward as a science, clinical research needs to use these understandings to develop useful predictive models that are tested and refined using large observational databases. Archimedes fast-forwards clinical care as science. And by using the new large EHR databases, it will put comparative effectiveness research on turbo.
We know Archimedes can provide reliable answers to a range of key health care and health policy questions – why is this area particularly important?
Expanding coverage for 47 million uninsured persons will soon be, we can hope, on the top of the national agenda. It will be important to be able to afford the expenses that come with that, particularly with forecasts of large deficits. Several of the leading health care proposals look to new “best practice” initiatives that will advance and – and apply – evidence-based health care to achieve economies. Effectiveness research is the foundation for evidence-based health care, although it will have to be supplemented by quality measures (identifying who provides effective care) and payment reforms.
What impact will the IOM’s recent recommendation to develop an independent entity dedicated to clinical effectiveness research and recent legislative proposals to set up a comparative effectiveness research body have on the future of this field?
The IOM and the new legislative proposals recognize that there must be broad political support for a national comparative effectiveness initiative, that this kind of initiative needs to address national priorities and to be generously funded, and that it is critical that comparative effectiveness research be first-rate science and guided by a trusted organization. The IOM report brings leadership and strong support from one of the most respected national health organizations.
What else needs to happen now to continue to make comparative effectiveness research an accepted part of health care?
We need the clinical research community to enthusiastically adopt these new “rapid learning” approaches. This is starting to happen, for example, with the Cancer Biomedical Informatics Grid (CaBIG), a World Wide Web specifically designed for health research, that enables researchers to access and share very large amounts of data. We also see this with the Cancer Research Network – it creates a new “virtual research organization” among 10 of the nation’s HMOs with large electronic health records databases. We also need many more physicians and hospitals to adopt electronic health records and make rapid learning from tens of millions of patients annually an integral part of the healthcare system.