Guest Commentary: Basic IT infrastructure key to healthcare's future

Our healthcare system is badly broken and in crisis. Study after study report the bad news: Up to 98,000 preventable accidental deaths in hospitals annually; getting research results into clinical practice takes an average of 17 years; up to $300 billion spent annually on treatments with no health yield; access to specialty care is highly dependent on geography; patients who are minimally involved in their own health decisions; public fear of identity theft and loss of privacy; fragmented and untimely public health surveillance; meaningful use of health information technology occurs in only a small proportion of clinical environments; and the litany goes on.

SEE OUR ESSAYS ON REFORM Healthcare reform cannot fix these problems without HIT, because the healthcare system is so complex and so information dependent. Without integrated HIT support, we clinicians are not humanly capable of practicing healthcare without killing people by accident. Although we blame—and sue—individual clinicians when things go wrong, as often as not, it is the “system” that is to blame, not the individual.

The quality and safety of healthcare delivery can be improved only at the point of service—reminding clinicians long after service delivery that their care did not meet a standard, when the clinicians are not given the data or the tools to help them make the right decisions, leads only to frustrated clinicians. We must direct the efforts of healthcare reform to fix the entire system so that it prevents these accidents while providing higher quality care and controlling cost.

Having an electronic health record, or EHR, system in every doctor's office is necessary, but not sufficient to solve the underlying problems. It would be like supplying the moon shot with a lunar lander; a necessary part, but one that cannot solve the problem at hand without the infrastructure and all the other parts integrated into a functioning whole system. Higher-quality, lower-cost healthcare can result only if we incorporate into the EHR system intelligent advice about what actually works. Using these “best practice” rules does not dictate how to practice medicine; it just means that each clinical decision can be informed by what has been shown on a national basis to have the best outcomes given what is known—so-called evidence-based medicine.

We cannot practice such medicine by relying on the—mostly verbal and from memory—interactions between a clinician and a patient in an examining room. We must change the way we practice medicine to include the computer so that clinical interactions involve a clinician, a patient and an information system that can provide advice to help both clinician and patient make better decisions, whether they are all in the same room or separated by space and time. This clinical decision-support component is the key, yet it cannot operate without two major inputs: data about the patient and rules about best practices that are used to interpret the patient's data; and the necessary standards and communications mechanisms to allow the secure exchange and sharing of the data and the rules.

The clinical decision-support engine cannot interpret the patient data unless the data are “computable,” i.e., expressed in a standardized way that conveys all the semantic content, including the context of the information. That requires the use of standard nomenclatures or vocabularies of clinical concepts and a standard mechanism for expressing those concepts and transporting them securely from their disparate sources into the EHR system. Such nomenclature and information exchange standards have been developed (e.g., SNOMED, LOINC, HL7, HITSP) but they require government funding, support and adoption through an agency with expertise in this area, such as the National Library of Medicine, before they will be broadly accepted and incorporated by vendors into such systems. The secure infrastructure by which such information is exchanged has yet to be created.

The clinical decision-support engine also cannot recommend the best clinical practices unless there are standard ways for expressing the rules and a trusted source for those rules. Currently, decision-support rules are developed by different vendors and expanded by clinicians locally, which typically makes them incomplete and inconsistent from one place to another. Optimally, such rules must be made available in standardized form for any EHR from any vendor to incorporate and use, but the only entity that has the capacity to serve as a common source for such rules across the country is the federal government.

There is also room here for a law that would protect a clinician from frivolous lawsuits about decisions that were based on such nationally vetted best practices.

Standards for how such rules should be expressed have been developed, but they require government funding and support through an agency with expertise in this area, such as the Agency for Health Research and Quality, before they will be broadly accepted and incorporated into HIT systems. This is one direct way to solve the problem of getting the results of clinical research into practice in a timely manner. In addition, a federally supported independent body, similar to the National Transportation Safety Board, must be created and funded to work with clinical experts to investigate medical errors and to design and promulgate the rule set for best practices to avoid them.

We need to build a new, secure “layer” on top of the Internet—a layer that handles the authentication, identity verification, audit logging, consent management, encryption and everything else that is required to manage the secure flow of sensitive information through the Internet. As with the original Internet, this model would require the federal government to fund the creation and adoption of the architecture and standards necessary to make it work and then build and maintain the underlying services necessary for it to function. Once the essential standards are set and the secure communications infrastructure is built and working, maintenance, expansion and refinement of the standards can be turned over to public-private boards similar to those that manage the Internet. The essential support services, however, need to be operated by a federal agency, much in the way the Federal Aviation Administration provides essential support services to the aviation industry, at least until (and if) the private sector can reliably earn the trust of the people and replace it.

Once the new secure information transport infrastructure is operational, health information exchange services would take advantage of it and add trusted exchange services on top of healthcare's “Secure Internet.” One component of trust in such an HIE mechanism is confidence that health information will be held, used, and shared securely in ways that ensure its confidentiality, integrity, and availability. Rather than having data security handled by myriad individual applications, these services built into a secure Internet would allow data owners (patients, providers, etc.) to control how and by whom their data are used and engender the trust in the system necessary to promote its use.

The recently passed Health Information Technology for Economic and Clinical Health Act, or HITECH, (part of the American Recovery and Reinvestment Act of 2009) contains the funding and the authority to support this approach. Although HHS will be in charge of implementing the national infrastructure, it will take much coordination and cooperative work on the part of many federal agencies and private organizations to succeed. As HHS parcels out the funding, it should dedicate a 10% “tax” from day one to support a robust, continuous quality evaluation and improvement process to independently validate and document the products and outcomes transparently, and to provide mid-course correction feedback. This will demonstrate the philosophy that providing good advice while decisions are being made is much superior to making critical judgments afterwards.

Medical errors kill as many people in a year as if a jumbo jet fell from the sky and killed all of its passengers every single day. We have known this for a decade, and we must now commit ourselves to stopping these accidents as a critical part of healthcare reform. We need to develop and appropriately apply the necessary resources to track, investigate and propose resolutions to the root causes of healthcare accidents so we can make some progress in healthcare toward the “over 2 years without an accidental fatality” that our commercial airlines have just experienced. The basic supporting infrastructure described above is as critical to the future of our healthcare industry as the national system of roads, tracks and airports is to our transportation industry, and only the federal government can lay down the foundation for all of us to build on.

William Braithwaite is chief medical officer of Washington-based Anakam, which provides identity authentication solutions.