|
|
| |||||||||||||||||||||||||||||||||||
Flexibility is critical, since the system will have to function under new policies and in the service of new health care delivery mechanisms, and it will need to incorporate emerging information technologies on an ongoing basis. As we seek to design a system that will constantly evolve and encourage innovation, we can glean lessons from large-scale information-technology successes in other fields. An essential first lesson is that ideally, system components should be not only interoperable but also substitutable.
The Apple iPhone, for example, uses a software platform with a published interface that allows software developers outside Apple to create applications; there are now nearly 10,000 applications that consumers can download and use with the common phone interface. The platform separates the system from the functionality provided by the applications. And the applications are substitutable: a consumer can download a calendar reminder system, reject it, and then download another one. The consumer is committed to the platform, but the applications compete on value and cost. The Facebook social networking site is another example: it allows users to connect their core accounts to applications that add value for them, from family-tree builders to programs that track flulike symptoms or encourage blood donation. And of course, there is the Web itself, which supports myriad applications — some proprietary, some not — many of which interoperate. For example, users with a personalized Google home page can populate it with widgets from Yahoo. Again, all these programs compete with each other and can be substituted for one another, entirely and modularly.
The platform approach to software design can be used to create and sustain an extensible ecosystem of applications and to stimulate a market for competition on value and price. We believe that the Department of Health and Human Services (DHHS) should encourage the development of such a platform for health care — one that will support applications for communication and computation that span the domains of clinical care, public health, and research. There are early-stage examples of platforms in health care already. For instance, the emerging model of personally controlled health records (PCHRs) is based on a platform that has been adopted by Microsoft and Google, as well as by the Dossia consortium of large employers for its rollout of the Indivo product to employees of consortium companies.1 There is now an active marketplace of enterprises building PCHR applications.
We take it as a given that health care software must be interoperable and secure and must protect patient privacy. But these qualities are not sufficient to produce an optimal system, which must evolve on a health care platform that extends beyond PCHRs to include other critical infrastructural components, such as medical-practice-based electronic health records (EHRs) and applications that support the complex communication required in health care. We believe that such a platform should have a number of additional key characteristics.
First, there should be liquidity of data. The platform and its applications should reduce impediments to the transfer of data, in an agreed-upon form, from one system to another. In the banking industry, the automatic teller machine (ATM) is predicated on highly standardized, simple operations. Participation requires at least a minimal amount of data liquidity — ATMs enable consumers from virtually any bank to withdraw money, although only some ATMs can provide a given consumer with his or her account balance.
Second, there should be substitutability of applications (see table). The system should be sufficiently modular and interoperable so that a primary care provider could readily use a billing system from one vendor, a prescription-writing program from another, and a laboratory information system from yet another. Individual systems do not need to perform all functions. (Analogously, a customer cannot apply for a mortgage at an ATM.)
|
Third, the platform should be built to open standards, accommodating both open-source and closed-source software. Though installation of open-source software is not free, its use decouples the software code from implementation and integration tasks and facilitates customization, extension, and innovation.
Finally, just as evolution requires variety in order to create ecosystem niches, a platform that supports diverse applications will lead to a robust health information economy. This architecture reduces dependence on individual systems by allowing competition and "natural selection" for high-value, low-cost products. This approach contrasts sharply with design of a national system by committee. Like standards, system design must be driven by successful, real-world innovations; an incremental and iterative process is more likely than a wholesale advance prescription to be successful.3 The platform model allows disruptive technologies to emerge and enables evolution to proceed organically. New companies and players that will contribute to transformation must be recognized and welcomed.
The DHHS could promote the creation of such a system by taking certain actions in terms of regulation, the creation of incentives, and the evaluation of results. Although the platform we envision would support a free marketplace of products and ideas, oversight and regulation are important. The DHHS should ensure that the dominant driving force is the maximization of health and that adequate privacy protection is in place. We must decide as a society what kinds of transactions such a system would be permitted to support. For example, should the platform permit direct-to-consumer advertising or procurement of samples for research?
At the same time, federal incentives should be offered to providers to make use of health information technology in clinical decision making and in efforts to improve the quality of care and acquire population data for public health. The design of incentives should be built on a realistic respect for physicians' time and effort in order to avoid turning physicians into scribes. A 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.
In addition, the DHHS should institute a transparent process for evaluation of the platform, individual applications, and the effects of the system on outcomes (health, patient safety, and public health), process measures (physicians' workflow), and costs (of the information technology as well as the provision of health care that relies on it). Ideally, the platform would support applications that would readily allow trials and observational studies of the technology and of therapies and delivery models, promoting what the Institute of Medicine calls a learning health care system.
To get started on this platform, the DHHS should demonstrate at least the kind of interoperability and substitutability that banks have instituted with ATMs. For example, can we produce a medication list for every American that can be obtained through standards-based,4 interoperable, substitutable applications? It would be a catalytic investment on the part of the government to ensure that such functionality is comfortably seated on a platform that stimulates evolution and competition among contending, substitutable applications.
Medicine is increasingly becoming a knowledge and information industry, but it did not invent information technology or the Web.5 It makes sense to draw on other sectors' successes in making this type of transition, and they teach us that if we are to use information technology to improve health care, the variety of practice sizes and styles needs to be complemented by collections of information functions that are packaged on a consistent platform. The applications enabling these functions should be as substitutable as different stethoscopes in a doctor's office.
Drs. Mandl and Kohane are the developers of Indivo, an open-source personally controlled health record that has been deployed at multiple locations, including through the nonprofit entity Dossia. In the past, Dr. Mandl received support from Children's Hospital Boston to guide the translation of Indivo technology to a Dossia environment. No other potential conflict of interest relevant to this article was reported.
Source Information
Dr. Mandl is an associate professor of pediatrics, and Dr. Kohane a professor of pediatrics, at Children's Hospital and Harvard Medical School — both in Boston.
References
| |||||||||||||||||||||||||||||||||||
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
