INFORMATION ARCHITECTURE

We have presented an expanded vision of how the organizational and technical infrastructure of HIOs (Health Information Organizations) could improve the efficiency and quality of public health reporting, facilitate public health investigation, improve emergency response, and enable public health to communicate information to the clinical community. We note that there are significant worries about the financial viability of some HIOs, some notable failures, and much concern about developing business models for HIO sustainability. These issues notwithstanding, there were 193 HIO projects at various stages of development in 48 states, including 42 that were operational at the time of a 2009 survey. In addition, new models for health information exchange are being developed to augment regional health information organization efforts, such as the Nationwide Health Information Network Direct Project (available at: http://nhindirect.org). The use cases described here provide clear advantages to public health, but each use case also provides benefits to clinical participants, whether by easing the burden of mandated reporting and responding to public health investigations, improved clinical decision-making that is based on epidemiologic data, or protecting the ability of the institution to function during public health emergencies.

Although clinical use cases often seem to be the initial motivators of HIOs, public health can and should get involved during the initial phases of development. Having appropriate public health agency representatives at the table early in the process may help influence governance issues and architectural design decisions so that the HIO project can support public health use cases.

Different technological approaches may be required to support public health use cases, depending on the system or network architectures for a given HIO. Examples of the varying architectures include centralized repositories, as in the case of large hospital networks with enterprise-wide electronic health record implementation (e.g., the Veterans Affairs or Kaiser Permanente health systems);  hybrid peer-to-peer file-sharing models, in which all clinical information is stored at the participant organization on edge servers that sit behind their firewalls but with patient demographics stored centrally to allow patient matching and retrieval of relevant clinical information; and patient-controlled health records, in which patients determine which data to deposit into their account and who has permission to view or change them (e.g., Google Health or Microsoft Health Vault). Obtaining the public health benefit envisioned here will require additional capabilities and functionality on the part of HIOsand a thorough understanding of pertinent legal and privacy issues.

Depending on regulations and the particular details of a given HIO implementation, the information relayed in the various use cases may be summarized counts that are fully identified, deidentified or anonymized — so that a patient cannot reasonably be identified individually — , or pseudonymized in instances where patient identifiers are not initially reported but a mechanism exists to allow rei-dentification if necessary (e.g., a clustered outbreak where confirmation and investigation are necessary). Although we have briefly mentioned the likely level of privacy necessary for each use case, the privacy implications of HIE are complex and are discussed in detail elsewhere.
Having a single point of contact on the clinical side for establishing, testing, and maintaining data flows would be invaluable to public health partners. What we have not discussed is the reciprocal need to consider how public health is organized to interface with clinical entities, both within public health agencies and across them. An assessment by the Council of State and Territorial Epidemiologists notes the limited progress made in integrating electronic disease surveillance systems, with only 13 of 48 states reporting interoperability between any surveillance modules. The Universal Public Health Node being developed in New York state is the latest in a series of attempts to accomplish this integration. The potential for harmonization of clinical reporting through HIE challenges public health officials to develop their own analog to the HIO, with parallel requirements for technology standards, staffing, governance, and trust.
Multiple parallel and discrete efforts are under way to institute electronic reporting from clinical providers to public health, covering such areas as electronic lab reporting, immunization and cancer registries, birth and death registration, adverse events, and syndromic surveillance, to name just a few. Unprecedented national investments in health information technology are poised to dramatically increase the amount of structured electronic data available and stimulate the advancement of multiple models for health information exchange. To maximize the benefits of these investments to public health, new HIE infrastructure must also demonstrate its ability to support these public health use cases, and health jurisdictions must be given the financial resources necessary to fully participate.