There are always myriad issues involved (from the technical to the philosophical) when it comes to moving content across disparate systems, but these problems come into glaring focus when the content includes confidential medical data. When the medical industry attempts to integrate records data, it faces a number of daunting hurdles—whether trying to move that information internally across a mix of systems or externally to insurance companies, state, and federal medical data sharing repositories, pharmacies, physicians, or other medical facilities. Of course, there are the expected technological obstacles faced by any industry when it comes to moving data across homogeneous systems, but there are also unique issues related to vocabulary (how different institutions label different conditions or different medications) and the fact that patients lack a single unifying medical record number.
Looming large are issues of privacy and security, which are of primary concern when working with this type of sensitive data. Perhaps more than any other type of content, it is absolutely essential that medical information be kept confidential and that only personnel authorized to see this information get to view it. As the industry moves toward a Web services model where information is made available on the World Wide Web, security and privacy take on even greater importance. This article looks at the issues the industry faces as it tries to move toward a more efficient means of integrating and distributing medical content, while maintaining patient privacy in the process.
Identifying the Patient
Dr. John Halamka, who is CIO of Harvard Medical School and CareGroup Healthcare and was recently named CEO of MA Share, an organization devoted to facilitating the sharing of medical data in Massachusetts, says the industry faces a number of key challenges, but perhaps the largest is finding a way to identify the patient when patients have different medical record numbers for each facility they visit.
"Recognize that regardless of whether it is hospital to hospital or region or state or nation, one of the great challenges is identifying the patient because we don't have a national medical record number in the United States. So if you go to a doctor's office in Boston and to a hospital in Boston, you are going to have two totally different medical record numbers. You have to build some kind of index, whether it is centralized in a hospital system or town or state or in the country, that interrelates all the medical record numbers you have been known by. That's certainly a first challenge," Halamka says.
Haim Nelken, manager of Health Care & Life Sciences for the IBM Integrated Medical Records (IMR) project based at IBM in Haifa, Israel, agrees that this is a major problem and that there needs to be a central place where all of the information on a patient is maintained. "Individual patients have lots of pieces of information in many places and we want to create an electronic health record which will be longitudinal, meaning it starts before the person is born and continues until the person is six feet under. Creating such an electronic health record has a lot of benefits," says Nelken.
Nelken says his organization has come up with the idea of an account, much like a bank account, which holds all of this information but that also can be divided into sections to show only the information any particular medical provider needs to do the job. "One of things we are trying to advocate is the notion of a bank, similar to a financial bank, which would be the repository of medical information for an individual. All pieces of information that would be accumulated for an individual will go into these accounts, and within those accounts an individual can create abstracts of information. When an individual wants to go to the dentist, he or she would get an abstract relevant to the dentist, along with universal information such as allergies, but the dentist is probably not interested in a broken arm or psychiatric episodes," Nelken says, pointing out that if you go to a general practitioner, they may not be interested in dental information but would be interested in just about everything else.
Halamka is working on a similar concept with MA Share to build a master patient index that provides a way to link a patient to whatever identifying numbers he or she has been assigned. "We created a master patient index for the state of Massachusetts, which today has 500,000 patients but by the end of the year will be hooked up statewide with about 3 million patients," Halamka says. He says that the index works in a similar fashion to the Internet, where you enter a URL, which gets translated into an IP address through the domain naming system, and then figures out where the server that corresponds to that URL lives. In a similar manner, entering certain data in the "Record Locator Service" will help locate the correct patient record. "Our notion was: why not create a domain naming system for patients where we cross-reference name, gender, date of birth, zip code, and the medical record and institutions you are known by?" Halamka says. In this way, he explains, you can see all of the places a patient received medical care, and they can link to all of the patient's medical information in this fashion.