Health information exchange and Stage 2 EHRs

Health information exchange is the next big phase of Health IT. Stage 1 of Meaningful Use was focused mainly on getting computer screens (instead of paper charts) in front of clinicians’ eyes as part of day-to-day work. Several vestigial pieces were required for interoperability, such as exporting electronic documents, but these were not required to actually connect anywhere. At least not at first.

Stage 2 of Meaningful Use is supposed to be about actually interoperating. It is supposed to be focused on successfully exchanging data from one place to another – like from a lab to an EHR, or one EHR to another (which may or may not be with the same vendor), or from the hospital to the community doc’s EHR, or (most importantly) between a patient’s PHR and their doctor’s EHR.

Regardless of the business issues of how health information exchange is supposed to be organized on a regional and national level – whether it is supposed to be via locally-designated “official” Health Information Exchanges (the “public utility model”), or whether it will happen across already-existing networks that had carried limited types of data in the past (like Surescripts carrying prescription data, or Emdeon carrying billing data) and who now want to carry broader types of data – the principles of health information exchange will be central to the discussion in this next period.

Given that the Office of the National Coordinator (ONC) for Health IT has issued an Advance Notice of Proposed Rulemaking (ANPRM) around creating standards for healthcare metadata – the definition of how data is to be packaged and exchanged – it might be worthwhile to review some of the basics of health data exchange.

How is data exchanged?
In order for data to be exchanged between separate systems, there are several things that need to be specified. (1) who is the patient? (2) who is the recipient? (3) who is the sender? (4) what kind of data is it? (5) what is the timeframe of the data? (6) how sensitive is the information? (7) how is the data protected against risks to Confidentiality, Integrity and Availability?

Each of these issues needs to be addressed in order for data to be exchanged. And each issue is the subject of exhaustive discussion. The particulars of the discussion can rapidly become, not surprisingly, quite arcane to outside audiences. So, rather than delve into the pluses-and-minuses of different approaches to each of these questions, it might be worthwhile to take a step back and look at real-world scenarios.

How have we done this in the past?
Consider how health data is exchanged in a paper-based environment. A typical scenario would go like this: (1) medical data is requested; the patient usually signs a consent form to document that it’s ok to send medical records, and this request is then faxed to the outside data source (the other office, or the hospital); (2) relevant records are copied by the request recipient, and often a summary/header form is attached, which contains patient identifiers that are “good enough” for both parties (name, date of birth, maybe insurance information, address); (3) the records are then mailed or faxed (hopefully to the right fax number or the right address); (4) the received data is then reviewed by a clinician, and placed in the patient’s chart. This works more-or-less, most of the time.

Sometimes, for example with prescription data sent by phone, by handwritten Rx or by faxed Rx, the data is sent to a local pharmacy. If there is any question, the pharmacist can opt to verify that the Rx is legitimate – calling the sending clinic for confirmation is the usual method. Once there is a trust relationship between the sender and the receiver, then this is seldom subsequently needed .

The principles of “close enough patient matching” for local parties, and the trust relationship between known senders and recipients, are characteristics of data exchange using paper. Verification, usually by phone, serves as an audit or way to handle questions. A simple version of data exchange, the Direct Project, is based on this premise of trusted parties and simple data exchange. To date, it has been used “in the field” in limited settings, and is faster-to-market than a more robust full-fledged build-out as envisioned in the ANPRM for Metadata Standards.

Do we need to wait for everyone to agree on the same standards?
We need to get to data exchange by whatever routes work the best. The risk of large-enterprise approaches to information exchange has been that the process is so arcane, and requiring of so many standards-making bodies to agree on a best-method of proceeding, that the process may get bogged down in its own quicksand.

As an example, the Department of Defense and the VA have been working for two years on building a seamless, lifetime medical record for troops in the military, and may take up to 6 more years to get it finished. Much of this has been due to difficulty in agreeing on interoperability protocols.

Clearly, despite the lofty efforts of those engaged in Nationwide Health Information Network (NwHIN) activity, we may be waiting a long time for all this to get worked out. In the interim, quick-and-simple exchange has been carried out by the Direct Project – and this may be sufficient for the small-office and local data exchange between already-known and trusted parties. Simple pragmatism says that this may be a better approach to Stage 2 Meaningful Use, simply because it can be deployed now.

What is the impact of web technologies on all of this?
Web-based technologies, such as the Practice Fusion EHR, put a whole new layer onto the discussion. With very broad adoption of this EHR platform by a sizable segment of the small-sized independent practices, there are two kinds of data exchange to consider: (1) user-to-user exchange of data within the web platform, and (2) user-to-outside party exchange.

Within the Practice Fusion platform, user-to-user data exchange – chart sharing between different practices – is part of the near-term roadmap. As this capability becomes more robust, physicians can refer to other physicians on the same platform (or enroll them quickly – it’s fast, and it’s free), and basically share important elements of the same patient’s chart. Tying this into the physician-to-physician referral patterns makes those referrals better for all parties – the workup already done is visible to everyone sharing the chart, the thinking of other clinicians is documented, and the response to the referral goes back to the referring doc in a much better way (“closing the loop”).

Exchanging data with outside sources – the local hospital, labs, other docs using other systems – is different too. Some of the exchange might be local (point-to-point connection) – dropping an electronic output file, like a CCR or Immunization Reporting Record, onto the local doc’s computer, and then sending that document to an outside local destination (perhaps using Direct Project tools). Other exchange can be done centrally, like lab connectivity, relieving the local doc of having to worry about building that bridge him/herself – when a connection is made with a large lab (e.g. LabCorp), the bridge only needs to be built once; after that, it is mainly an administrative matter of matching clients on either end (a “toll plaza issue” rather than a “bridge building issue”).

Conclusions
Health information exchange (lower case) will the be central issue in the next stage of EHR maturation. Stage 2 Meaningful Use should keep this focus. Large enterprises (the Epic’s of the world) will struggle to establish high-level data exchange, and may get bogged down in the intricacies and politics of Metadata Standards definitions (not unlike the DOD and VA experience). The market will encourage other smaller, more lightweight solutions – the Direct Project may be among the early successes. The impact of widely-adopted web-based EHRs will also drive data exchange in ways not foreseen by the NwHIN vantage point, and will change the landscape rather quickly. This next year will be one where intense innovation will help “connect the dots” of health data, and begin to knit together the highly fragmented status-quo of health care.