Publication
Article
We must overcome all of the technical, political, and economical obstacles NOW. The need for system efficiency and patient safety demands it.
Imagine the following scenario: You’re on a vacation and need cash. You go to the nearest ATM and insert your card. Th e machine replies with, “I’m sorry. I do not know you. And, I have no idea how much money you have to withdraw.” You try another and another ATM with the same results and come to the inevitable conclusion that “someone has to do something about this.” We have become so accustomed to the ease of electronic banking that we would find the above scenario extremely frustrating and demand change. And yet, we continue to tolerate similar episodes in healthcare on a daily basis. We act on incomplete information, we duplicate testing performed elsewhere, and we spend an inordinate amount of time collecting information that others have already collected. We must overcome all of the technical, political, and economical obstacles NOW. The need for system efficiency and patient safety demands it.
Now imagine a vision in which the patient, and when appropriate the patient’s family, as well as any physician, allied health professional, hospital, or laboratory that have appropriately delegated access to that patient’s medical records, can input data to contribute to his or her management. Is it possible? Of course it is. Think about the ATM machine in our example: It is relatively agnostic about where its data is coming from; it functions because agreedupon verification processes exist, and there are industry-standard defi nitions for “balance,” “withdrawal,” “deposit,” and other terms.
Progress is being made in healthcare, albeit slowly. Many justify this by pointing to the relative complexity of medical data versus that used in the banking or airline industries. Is this really true? Or is it an excuse for some other agenda? Is our data really that much more complex?
Models for Unification
There are basically three models that have been developed that fulfi ll the requirements of a portable, unified medical record:
Model #1
There are totally independent information systems, from which defined baskets of information are downloaded to some portable storage device. The radiology industry was one of the first to adopt this technology, burning CDs with MRI, CT, and other digital images. The patient is the primary “owner” of the information and carries it from provider to provider, who try to read and incorporate it, as best they can, into their native systems. Obviously, this was, and still is, a reasonable method of achieving some degree of portability of what has become relatively standardized information, although the amount and type of information conveyed are rather limited. Some patients have taken things a step further by manually converting their lab data into impressive spreadsheets, graphs, and MS Word documents. Patients were the first to recognize that, in the absence of system solutions, they would need to take control of data organization and portability. Source data integrity and cross-compatibility between systems generally remains problematic, except for radiology images.
Model #2
Vendors develop centralized, usually Web-based, repositories to house relevant clinical information organized on a per-patient basis. With proper access security, any given provider can go to a centralized database and access pertinent patient information. Depending upon the systems utilized by the provider, some of this information may be amenable to importation into the provider’s native electronic medical record. The Web-basis of this design facilitates portability and relieves the patient of the responsibility of being the most current source of data.
Data integrity is generally improved, especially for those elements with relatively standardized vocabularies—such as demographics, laboratory, and radiology—but falls off when data is less standardized, such as with dictated reports. Nonetheless, this industry continues to grow and evolve, awaiting more formalized and complete data definitions. Despite significant efforts toward secure systems, numerous polls have shown patients remain reluctant to have their medical data “out there” on the Web.
Model #3
The assumption is that the majority of the care for any given patient is delivered at one primary location. Generally, this applies to large, multi-specialty centers. It is less true for community-based providers, given that the balance of care is delivered in outpatient, smaller clinics as compared to care delivered in an inpatient setting. In that situation, hospitals continue to investigate and navigate the many obstacles associated with sharing electronic medical record systems with their physician staff. Patient input and access has been slowly adopted.
Model #3 is the approach we have decided to pursue at MD Anderson, where we are in the process of enhancing our patient-centered EMR. A guiding philosophy, however, has been to include both community physicians and the patients themselves as part of the input and output system. Patients who are technologically savvy now input the information about which they know the most (eg, demographics, medical history, family history), and this is “imported” into our EMR, after being verified by the provider. Additionally, patients have access to defi ned elements of the EMR, such as scheduling and secure e-mail to communicate with select providers. Other accesses are in the development stage. To complete the circle, outside involved providers are given access to the EMRs of those patients with whom they have a direct relationship, granted by the patient.
Technology and policy drives the particular aspects for which access is available. The days of waiting for a summary letter about a recent discharge or treatment plan have been shortened potentially to hours or even minutes, thus facilitating communication and increasing patient safety. None of these models are the perfect solution for all circumstances. Each meets certain needs and requirements, given the limitations in current systems integration. Nonetheless, the goal should be the same: unifi cation of the patient and all providers to provide the highest quality and safest medical care.
Dr. Walters is the associate vice president of medical operations and informatics at the University of Texas MD Anderson Cancer Center.