Database is

The recently released Commission on the Future of Health Care in Canada, better known as the Romanow report, has added fuel to the health care debate. Yet, at the same time, one of the report’s findings is being overlooked: namely, can a truly comprehensive, national electronic health records system be created?

The answer, according to a variety of experts ComputerWorld Canada spoke to, said the task at hand, though enormous, can be accomplished.

The report makes a series of recommendations, one of which is the creation of “a personal electronic health record for each Canadian that builds upon the work currently underway in the provinces and territories.”

“It takes time for people to sink into this, but it is absolutely doable,” said Dr. Thomas Jones, chief medical officer with Oracle Corp. in San Rafael, Calif. Though Jones no longer practices medicine, he understands the benefits of IT.

“The best way to reduce error is to have the computerized memory of the entire patient’s medical care,” he said.

This, in a nut shell, is the logic for electronic records. Though there are many other benefits and concerns, improving patient care is the foundation.

But building it? That’s where the fun starts. “The first thing you have to figure out is what to build,” Jones said.

Standards, and this includes the information model of how data is represented, are the critical starting place, Jones said. Long before any actual integrated IT system is built, a common platform will have to be chosen.

The global medical profession has something called Health Level 7 (HL7), which is a non-profit organisation responsible for the production and promotion of a series of IT health care standards. This includes something called clinical document architecture, which is an XML-based standard for clinical documentation.

This all “provides a very solid basis on which to implement the information model,” Jones said.

The information model (or patient data) can follow one of several standards for medical terms and items, he explained.

Two of the most common are the ICD system (international classification of disease) and SNOMED (systematized nomenclature of medicine).

“If you start with the notion of an application development framework that incorporates these sort of standards, then you have opened the door to a broad range of application developers,” Jones said. “It allows you to go across boundaries with great agreement…much like the Internet,” he said.

“You have to build this system on open standards,” agreed Sal Causi, business development executive for life sciences with IBM Canada Ltd. in Markham, Ont. There has to be commonality for data to be moved across the country, he said.

Various hospitals, counties and provinces are undoubtedly going to choose their own hardware and applications, so it is critical that the language between them is common.

build it and they will come

The next phase is figuring how all the data will be accessed. The idea of a national database of all Canadian’s medical records is viewed by most experts as inherently problematic. Not only does it create a problem with patient data synchronization (data would always have to be updated both locally and nationally), but it would also create the motherlode for hackers.

The consensus is that a distributed network should be created where individual doctor’s offices store patient information at centralized repositories – invariably at nearby hospitals. The hospitals themselves store their own information. Thus all patient data pertaining to John Doe would go to a predefined storage repository, regardless of where it comes from.

For a doctor to access a patient’s data, he or she would access a national directory which in turn would locate where the information is stored, not unlike the Internet.

Dr. Dana Hanson, the president of the Canadian Medical Association based in Fredericton, sees many potential benefits to a nationally accessible information system, not the least of which could occur during emergency situations when a doctor needs access to a patient’s charts which might be stored, in paper form, in an office miles away.

Another advantage is reducing the amount of tests done, since medical professionals will have access to the most recent data without having to reorder tests.

Everyone agrees that old, paper-based patient data is unlikely to be transferred to digital form except in rare circumstances since the cost would be too great.

“In my practice, I would have to [file] 25 years of charts,” Hanson said.

Privacy and Security

Of course with increased levels of access, there are increased potentials for abuse.

In one scenario, patients would control access, so that doctors would require their consent before reading a chart. For emergency situations there would be an override, Jones calls it “breaking the glass,” which would allow for immediate access to patient information but would automatically be flagged for an audit.

Regardless, random audits will be needed to make sure data is not being illegally accessed. Many Canadian hospitals already do this and are convinced it is a better method of protecting a patient’s privacy.

“In the electronic world one has a better chance of security (and) privacy than in the current environment,” Jones said. “I can tell you from personal experience that anybody with a white coat can get a hold of paper charts. It is just that simple.”

Causi said that all stored data should be encrypted and that only registered users could unencrypt the data. Jones goes a step further, he would like to see biometrics used to get at the encrypted data.