Publish and Subscribe

The current published version of an ITI White Paper on this topic can be found at the IHE web site. The general ITI Publish Subscribe wiki page contains the most up-to-date information on this topic.

1. Proposed Workitem: Publish and Subscribe

• Proposal Editor: Floyd Eisenberg
• Editor: N/A
• Date: N/A (Wiki keeps history)
• Version: N/A (Wiki keeps history)
• Domain: IT Infrastructure

2. The Problem

Repurposing of health care data refers to any use of “data other than the primary purpose for which they are recorded” and can address a myriad of possibilities. Although not intended to be a comprehensive list of repurposed data uses, examples include biosurveillance, quality performance measurement, pharmacovigilence, research, and healthcare financial analysis.

Using biosurveillance use-cases as specific examples, aggregate data are needed for syndromic surveillance, situational awareness and notification. Health information exchange performed through the sharing of documents (e.g. CDA, CCD and IHE-XDS), uses document based transactions such as notification, retrieve and query. While these transactions may work well for individual physician-to-physician or patient-to-physician notification and retrieval of information, managing this information in aggregate poses some challenges. Document availability notifications for each encounter document, today would require the sending of messages directly to Public health authorities that need to process and aggregate all of the information to access what is useful for their purposes. With information shared in RHIOs, Public health authorities (and other potential users of repurposed data) should be able to avoid the need to duplicate some of the information stored, and rather take advantage of the ability to subscribe to certain specific filters detecting when certain information is present. Additionally, as Public health authorities cannot predefine all potential conditions of interest; some conditions may need to be identified through data mining at the receiving (Public health) end. Therefore, a publish and subscribe methodology, even in a document based sharing environment is required for appropriate and high performance aggregate and population-based data analysis functions.

3. Key Use Case

Data are shared among clinical care providers for the purpose of clinical care delivery to identify active, current and historical information about an individual patient. For purposes of Pubic Health, there are two essential functions regarding individual patient data:

1. First, is a case management function in which the Public Health practitioner acts as a case manager and has the authority to investigate the patient, identify contacts, and determine and/or provide appropriate care to the individual patient. This function is similar, in many respects to direct care delivery in which the Public Health practitioner assumes an adjunct, or care provider role. In the case of clinical trials and research, this adjunct provider role manages the data regarding each patient included in the research. With respect to quality, case manager clinicians may similarly manage individual patient data for individualized patient care interventions.

2. The second function for Public Health is for surveillance of all occurrences of disease in a defined population. Surveillance is important to determine prevalence (extent of disease in a population at a point in time) and incidence (newly developed cases of disease per a defined period of time). Determination of incidence is further defined for new or emerging symptom constellations (i.e., syndromic surveillance) and for situational awareness (the extent and rate of disease progression in the population). Public Health is not necessarily able to determine the exact nature of the disease entities that will present; however, types of data elements required for analysis can be predetermined. Public Health requires the ability to subscribe to required and available data types for analysis by the Public Health Biosurveillance system. When one of these occurrence is identified, the Public Health Biosurveillance system needs to be aware of it and capture it within the analysis set. Also required is the ability to collate multiple data elements for an individual patient to maintain appropriate counts of patients in the population with the disease entity. Once Public Health identifies specific additional information required on an individual patient (for case management or for surveillance purposes), that information is expected to be provided through case reporting. (Figure 1). In the case of clinical trials and research, this data miner / analyzer role manages the data to determine appropriate cohorts of patients to include in research protocol offerings such that consent can be requested. With respect to quality, there is a similar requirement to identify rates of adherence to protocols and guidelines, as well as to identify adverse events and outcomes of care for trending and attribution for performance improvement initiatives.

For this surveillance requirement, it is proposed to define an IHE profile for a Subscribe and Notify capability. This profile would allow an environment where document contributed by several document sources in an affinity domain may be analyzed against the “data types” subscribed by a Public Health Authority (or, other users of repurposed data). This would include some active monitoring capabilities related to the XDS document registry/repositories.

4. Standards & Systems

Stakeholders that could leverage such a “publish and subscribe” architecture include Public Health (Biosurveillance Information Systems - BIS), clinical and other research oriented organizations, Quality reporting organizations and insurers such as JCAHO, NCQA, AMA DOQ-IT, CMS, Public Health in its role as a quality performance monitoring agency, and others.

Relevant standards are identified in the HITSP constructs approved to date, including HL7 CDA-r2.

5. Discussion

IHE profiles are currently recommended as part of the HITSP Care Delivery (formerly Electronic Health Record <EHR>), Consumer Empowerment (CE) and Population Health (formerly Biosurveillance <BIO>) interoperability specifications and constructs. To avoid excessive implementation and coordination efforts from ambulatory physician offices, hospitals, and other clinical care facilities and settings, a single methodology for sharing patient information must also be usable to manage surveillance, reporting, and research needs. IHE requires a clear focus on aggregate, population-based data usage and methodology. The focus of health information exchange to date has primarily been related to sharing of individual patient data for individual patient care decisions. It is most appropriate and timely to explicitly address the aggregate and secondary use data issues using the same constructs, but without encumbering primary use and users of the clinical data systems.