National Patient Identifier with FHIR is the answer

Direct Secure messaging has been implemented across the country by many physicians and hospitals due to ONC Meaningful Use requirements. Direct is great for clinician to patient interaction or even consults in some cases. Is this really the way that we should be sharing patient information? Ideally, we want the information in the patient record not just between two clinicians out of band and not stored in the patient history. This way anyone entering later in the patient care scenario has access to the information should it become necessary and the patient has also consented to the release.

Last September, John Halamka posted an article about a new RESTful national provider directory that has been built with the idea that Healthcare Information Services Providers (HISP's) would contribute the known Direct addresses of their clinicians to this directory. While this is a great concept, it also promotes a point-to-point communication mechanism that is not a good fit for producing a longitudinal health record.

Industry really needs two things that John has talked about for years. A National Patient Identifier (NPI) and a patient consent system. These items, combined with the implementation of FHIR, could change everything. We don't need the Federal Government's help to put these things in place.

With a NPI, one could look up a patient’s information the moment they arrive in an emergency room. The system could then poll all known FHIR endpoints using the patient identifier and then aggregate the data returned into a singular view of the patient's history. The same technique of uploading known Direct addresses could be done for a NPI. With the patient's consent, a provider can search the NPI for a patient using the unique identifiers already available and in use today. Once a patient is located, a provider can update the information as necessary.

Successfully doing this requires trust relationships among the various users of the NPI. This could be executed in the same the way groups such as Commonwell function. Consent would be stored locally with the encounter data. If the patient has provided the consent to release their medical history, then the treating organization would return a response when queried via their FHIR endpoint. The same organization that stores the NPI could also maintain a list of all members’ FHIR endpoints. These endpoints would be queried when looking for patient’s histories.

The real goals are to look up patients, pull down all relevant data that patients have approved to send, and then aggregate that data into a single picture of the patient's health history. Does anyone really want another email account to monitor? The technology exists to achieve these goals. What is needed is cooperation and partnership among industry to make it happen.