Mandl and Kohane on Health Care Information Technology
Posted Jun 03 2009 4:52pm
In a March 26, 2009 article in the New England Journal of Medicine entitled "No Small Change for the Health Information Economy, Kenneth D. Mandl and Isaac S. Kohane emphasize that interoperability is not sufficient to achieve the results we need. What is required, they say, is "flexible information infrastructure that facilitates innovation in wellness, health care, and public health." Such flexibility is critical. If we are to provide massive incentives to implement health care technologies, do we want systems that merely preserve and actually strengthen our fragmented system or do we instead what systems that "function under new policies and in the service of new health care delivery mechanisms." Do we want to apply to the systems we use a certification process that is of necessity complex and slow so that by the time of certification are already "legacy" or do we want to create processes that "incorporate emerging information technologies on an ongoing basis?"
To achieve the aims of a "learning healthcare system" promulgated by the Institute of Medicine and others, such technologies must not just be interoperable, they must be composed of components that can be substituted one for another. They use as an exemplar the components of our ATM banking technology systems. The Apple iPhone with its myriad applications, is cited as another case. (As yet a another example - not the authors, - personal computer users can substitute one web browser for another, one news feed for another. What is important is that the functions of each component are provided reliably, not that the entire package is somehow certified).
What types of components can be substituted? The authors have a table of categories worth further study and examination. Each category is defined by its purpose. In their view, the "proposed platform would allow a clinical practice or hospital to select the combination of applications that are most useful for the local environment. As alternative applications are developed by competitors, the existing ones may be substituted, or new ones added." Examples cited are
Public health reporting
Data acquisition (subscription)
Each category has many examples. Readers are urged to study this table carefully and to refine it. These two authors create a structure that supports a stable "platform" upon which interoperable and substitutable components can be added. They claim that health care software must be
promote data liquidity by reducing "impediments to the transfer of data"
modular to ease susbsitutability of specific applications
limited in scope when need be (every system must not do everything just as "a customer cannot apply for a mortgage at an ATM.)
based on open standards
robust through diversity that allows choice and "natural selection...driven by successful, real-world innovations."
They join the growing chorus of leaders who argue that certifiying the entire package slows innovation and is doomed to a view of the world through the rear-view mirror. Functional requirements will change and should be not confused with certification. How functional requirements are met is the subject of the type of assurances that certification can provide for subsitutable components.
Their view - one this writer and many others share - required DHHS to take immediate action through regulation, incentives, and critical evaluation of results. The current approach, many argue, is to focused on static products and outcomes measured through the imperfect prism of claims data. As the authors state, one "positive step would be to reduce demands for excessive documentation to support billing and medicolegal defense, so that valuable data-entry efforts could serve nobler goals." DHHS could also introduce greater transparency into its evaluation processes. DHHS should draw on the experience of other fields. The authors cite the Stead and Lin National Academies report as one example.
One hopes informed and influential people will read this article, reflect on its common-sense approach, and stop the sincere but misguided proliferation of certification efforts and quality metrics that root us in the past and potentially making matters worse.
One also hopes that the same rigor encouraged of clinicians will be applied to other actors in the health care delivery industry. As a global metric of performance, should we consider thousands of formularies already prevalent in the US as a good thing? Where is the evidence for this? Where is the business purpose? Similarly, the wide variations in authorizations and processes administrative entities require. Why do they have to be so different? What is the global rationale of unnecessary complexity? How does this allow for meaningful differences in services that count?
We face great opportunity. ARRA requires decisions. The right ones - along the lines of Mandl and Kohane - will make an enormous positive impact. The wrong ones - including much of the status quo - will doom us to a static world that is inefficient and unresponsive to the tsunami of need our society faces as we age.