Robert Rowley, MD
Twitter: @RRowleyMD

The effort to build health IT products that are patient-facing, and able to pull together all the health information about oneself into one unified dashboard, is well underway. It is the inevitable next step in the evolution of health care data – putting patients at the center of their own data, and having them be the shepherds of it. In order to really achieve this in a scalable way, however, there is one big foundation-piece dilemma that needs to be resolved: establishing a unique patient identity.

Many countries issue a national healthcare identity card, but this is not the case in the U.S. In fact the most widespread surrogate for a national identifier, the Social Security Number, is specifically dissuaded by numerous pieces of federal legislation for use as a unique healthcare identifier.

What we are left with, then, is that a given individual has a collection of unique identifier numbers – one for the (current) health insurance plan, one for the medical delivery system (if care is delivered through an integrated system with an enterprise-level unique patient record system), and one for each doctor’s office that functions independently. How can these be tied together?

The fragmented healthcare dilemma
Mapping patients in one system to the correct one in another system has been a challenge facing us for quite some time. When I was at Practice Fusion, helping build, implement and grow a large web-based EHR system – where a single system connected tens of thousands of clinicians and millions of patient records into a central data store – there was still the issue of identifying patients coming in from e-prescribing channels (Surescripts), as well as lab results coming in from major clinical laboratories (Quest, LabCorp and others).

Being separate entities, it was inconsistent that pharmacies and labs knew of the practice’s internal patient ID number. Therefore, a set of probabilistic matching algorithms needed to be built, which could identify incoming data with in-EHR patients fairly well, but not with 100% accuracy. When there was uncertainty, it required the user (the clinician) to hand-pick the correct patient with whom to link the external data.

This is the same challenge faced everywhere, and becomes particularly important when building networks (HIEs, ACOs) that try to link different systems together into one central hub. In an EHR deployment, at least there is the clinician who can hand-pick the correct link when there are ambiguities about incoming data from outside sources. But who does this for an HIE?

A patient-centered hub might solve this
Most of the solutions to this dilemma have been provider-system focused: how a hospital system can link its records with a clinic’s records, with a lab’s records, with a prescribing system records. The central “atom” of identification, however is simple: it’s the patient.

Let’s assume, for the moment, that in a patient-centered unified personal health record system, the patient’s identify is validated and authenticated as belonging to the right person. Records pushed out to the patient from healthcare providers (each doctor’s EHR systems, where they exist) in essence validate the individual’s identity by virtue of having an in-person relationship with the patient. The doctor says “here are the login credentials for you to see your information in our system,” based on knowing the patient first-hand. Each doctor who has such an EHR (and, in 2013, a majority of U.S. physicians have EHR systems) gives that same individual the login credentials to look into each doctor-specific patient portal. The patient now has a “set of keys.”

If the patient uses a universal personal health record system to connect with each of the doctors, using the keys they have been given to connect with each one, then that universal patient-centered portal now becomes the mapping path between each of the participating physician’s EHR systems.

Taking it a step further, if that universal personal health record system allowed for consumers (patients) to sign up themselves, and undergo a rigorous authentication process to make sure that the individual is who they say they are (and not, for example, a gossip reporter trying to spoof his way into a celebrity’s health records), then records can be (1) created by the patients themselves if they want to, and (2) links can be requested from each of their healthcare providers (including doctors, hospitals and health plans) and tied into their own unique record.

What does this mean?
A National Health ID system may never happen in this country (speculation on my part here). There may be too many political obstacles for this to work its way through Washington. Absent that, then there needs to be a way of connecting together patient records housed in separate, fragmented systems, in order that the goals of interconnected health care can be realized. The current system of probabilistic matching between systems has come a long way, and is pretty good, but still needs manual governors when there are ambiguities.

Though we are in its infancy, patient-centered health IT products may be the most likely way of linking it all together. Patients given access to their record from each one of their providers, and linking them together onto a common platform, creates a mapping strategy between systems that is highly accurate. The curator is not the providers, or hospital CMIOs, or other such provider-system-centered overseers. The curators are the patients themselves.

Dr. Robert Rowley is a practicing family physician and healthcare information technology consultant. This article first published on his blog. From its inception through 2012, Dr. Rowley had been Practice Fusion’s Chief Medical Officer, having created the underlying technology in his own practice, and using that as the original foundation of the Practice Fusion web-based EHR. Dr. Rowley brings a depth of experience and expertise in health care as well as health IT, having been in clinical practice for 30 years, including experience as a Medical Director with Hill Physicians Medical Group and as a developer of the early EMR system Medical ChartWizard. His family practice in Hayward, CA has functioned without paper charts since 2002.