The American Recovery and Reinvestment Act represents a significant investment in support of our nation’s infrastructure for health care information. In particular, the Health Information Technology for Economic and Clinical Health, or HITECH, Act provisions have been key to advancing health care delivery and outcomes via the use of health IT. Exchanging health care information across settings and among all providers is a central vehicle for health care improvement. Such sharing promotes patient-centered care, improved outcomes and enhanced efficiency. A major component of the federal infrastructure to support health information exchange is the Nationwide Health Information Network (NHIN). NHIN was initiated to provide linkages that would facilitate information sharing across organizational and geographical boundaries, as well as among health information organizations, delivery systems, federal agencies, health plans, providers, pharmacies, laboratories and other health care stakeholders. To augment NHIN functionality, the Office of the National Coordinator for Health IT initiated the creation of NHIN Direct. Its goal is to support simple point-to-point data exchange between two known entities. Exchange Versus Direct: A Short Summary For the purposes of this Perspective, Direct will be used to refer to NHIN Direct, while Exchange will refer to broader NHIN activities. ONC provides the following definition for Exchange: a set of standards, services and policies that enable secure HIE over the Internet. While Direct also fits these definitional constructs, there is a key difference between the two. Direct supports only “push” use cases, in which a discrete document, template or set of information regarding a specific patient is sent, or pushed, to a known entity or person. These types of transfers would otherwise take place through other means, such as facsimile, mail or email. In contrast, Exchange was designed for “pull” use cases, where a query for information can be sent and, where available, the relevant information is pulled back into a system as a response. Exchange supports many types of inquiries, whether the entities are known to one another. Exchange could be used by a provider querying to find all incidents of care a patient received, when, for example, a provider had a new patient and needed a more complete medical history. If a patient presented to an emergency department with complications from a chronic disease, such as diabetes, Exchange could also be used by the ED to query the most recent labs or test results ordered by any provider and performed by any entity. In contrast, Direct’s goal is to improve the exchange of health information by beginning with more immediate and easily implemented transactions. Direct would enable point-to-point exchange between two entities. For example, a family physician referring a patient to an endocrinologist could use Direct to send a summary of care record. Similarly, a laboratory could send results to an ordering provider via Direct. Both Exchange and Direct can be used in concert; use would be governed by the type of information exchange required and the relationship between the exchanging entities. According to ONC, Direct is not intended to address every type of information exchange. Rather, Direct is designed as a way to initially satisfy “meaningful use” requirements. In addition, Direct may be of more immediate benefit for information exchange at the community level, where providers have established professional relationships and referral patterns. In contrast, Exchange has broader applications, pulling information from any relevant source regardless of whether or not providers are known to one another. NHIN Exchange Overview Exchange can be considered a “network of networks” designed to connect a diverse range of entities that need to exchange health information (e.g. state and regional health information organizations, integrated delivery systems, health plans, federal agencies). It is a set of standards and services that supports interoperability and provides a framework for exchange, including:
■An Internet-based network for secure transport; ■Assurance that only valid, trusted entities participate; ■Legal agreements to protect privacy and security; and ■Governance activities to structure and define activities, roles and responsibilities. ONC initiated NHIN development in 2004 as a successor to the 2002 National Health Information Infrastructure (NHII) project. Since then, a number of stakeholders (e.g. federal agencies, regional and state health information organizations) have supported NHIN activities. In its first year, four standards-based prototype architectures were designed. The prototypes demonstrated a number of functions, including: ■Ability to find and retrieve health information within and among health information organizations; ■Delivery of new data to appropriate recipients; ■User identity proofing, authentication and authorization; and ■Methods to match patients to their data without a national patient identifier. The second phase of NHIN activities focused on developing specifications, services and working constructs to support priority HIE areas, including: ■Emergency responder electronic health records; ■Lab results; ■Medication management; ■Consumer access to clinical information; ■Registration and medication history; ■Quality; and ■Bio-surveillance. This phase also included trial implementations, tested by health information organizations and federal agencies, from June 2007 to August 2008. In 2009, MedVirginia became the first organization to use Exchange to support disability determinations made by the Social Security Administration. In September 2010, Indiana was the first state to use Exchange for public health purposes. Indiana’s use consisted of sending de-identified influenza and pneumonia data from 76 EDs to CDC. By the end of 2010, it is expected that upwards of 10 agencies and organizations — including the Department of Veterans Affairs, provider practices and others — will use such standards to exchange information. Exchange has a number of core capabilities, including functionality to: ■Query, find and retrieve health information and documentation among diverse entities; ■Deliver a summarized patient record; ■Support consumer preferences regarding the exchange of personal health information, including opting out of NHIN participation; ■Support a common trust agreement; ■Match patients to their data without a national patient identifier; and ■Support standards developed by consensus bodies. Exchange functionality has been used by federal agencies and contractors as part of the rollout process. Participants include Beacon communities, CDC, CMS and others. NHIN Direct Overview Formally announced in 2010, the goal of Direct is to support a simple, standards-based way to send encrypted health information directly to known, trusted recipients. The Direct concept was predicated on the idea that existing Internet protocols could be leveraged for basic exchange activities. Through a consensus process, Direct proposed an approach to information exchange based on current e-mail standards and processes, with added security and other features to ensure information could be transported from a sender to the anticipated destination. Practically, Direct can be considered a replacement for communications that would currently be handled via fax, postal mail or e-mail. For example, rather than faxing or e-mailing a lab result to another provider, a physician could use Direct as a secure way to transfer information. Direct activities are focused on drafting specifications and services, including tools to guide development, from a protocol-based approach. Such tools encompass descriptions of standards, services and policies that enable secure health data transmission. Once prototype code has been developed and tested, the resulting standards and specifications will need to be adopted by a wide range of trading partners. As noted above, once complete, Direct could be used to push data from one health IT system to another. Transport level specifications will handle different types of information, including unstructured, semi-structured and fully structured content. Examples include: ■Unstructured: text, PDF of lab result; ■Semi-structured: clinical document architecture templates, medical document management; and ■Fully structured: Continuity of Care Record, Continuity of Care Document. Unlike Exchange users, Direct users would have to establish policies and standards for deciding which other Direct addresses to trust. This is because Exchange users are required to sign a data use and reciprocal service agreement that essentially establishes a code of conduct and legal framework to enable trust among users, while Direct users are not. Direct is developing a reference implementation to establish the necessary technical components. Direct working groups are addressing documentation development for various audiences and the requirements that support a range of transmittal functions, such as sending lab results to the ordering provider, preventive care reminders to patients or immunization data to public health entities. Direct development has been based on the concept of open government, which has allowed a broad array of stakeholders to participate. While Direct was primarily created as a way to support Stage 1 meaningful use requirements, Direct may accelerate development of the broader NHIN. Policy Issues Under HITECH, ONC is required to establish a governance mechanism for NHIN. ONC recently established a governance work group to fulfill this requirement. So far, the work group’s activity has been focused on issues related to trust and state compliance issues. On Oct. 20, the work group presented initial recommendations to the Health IT Policy Committee, and on Nov. 19, the work group will submit final recommendations. In addition, it is likely that minimum standards will be established for Exchange participation. ONC still must complete the regulatory process to establish more specific rules regarding participation and utilization. As Exchange and Direct mature, a number of issues — such as the eventual convergence of the two exchange methods — will need to be explored. In addition, if and/or when Exchange and Direct are used to support the same clinical encounter, data discrepancies may require reconciliation. For example, during the physician referral process, Physician A could refer a patient to Physician B and use Direct to send care summary information. Meanwhile, Physician B could use Exchange to query for incidents of care relating to that patient. Any discrepancy between the data Physician B obtains from these two sources would need to be reconciled. Another outstanding policy issue is the question of how and by whom Exchange and Direct adoption will be supported. Among other things, the role that states will play in adoption efforts remains to be determined. Finally, a key open question is whether Direct development will advance at a pace adequate to support Stage 1 meaningful use exchange needs. Exchange and Direct will play a central role in the advancement of health data exchange activities. While Direct has a more limited scope, both will contribute to more immediate needs vis-à-vis meaningful use, as well as the longer-term and broader scope requirements of a diverse range of entities and to meet a wide range of transactional needs.