Manufacturing standards made your last driving vacation possible. How remarkable that gasoline refined in Oklahoma functions in an engine built in Canada, and that the fuel pump nozzle in Colorado fits a car purchased in Ohio. A reader in California captured the number from your New Jersey bank card and dispensed that emergency spending money. The cord on your notebook computer modem fit the phone jack in your hotel room and let you check up on your office e-mail even though you promised yourself this was going to be a vacation. And the car top luggage rack was just the right size to fit beneath the highway overpasses half a country away from home.
If you traveled to Europe or Africa, however, things didn't work as smoothly. Electrical power had a different voltage, frequency, and physical connector. The phone connector was different, too.
Most of us give little thought to technical standards, even though they affect our lives daily. Like electrical power, drinkable water, and pH balanced shampoo, standards are expected and when they function well we hardly notice them. We tend to pay attention only when the lack of standards endangers or inconveniences us.
Patient care depends upon detailed, timely information -- both specific information about the patient and general clinical knowledge. The lack of standard ways of collecting and sharing this information reduces efficiency and hinders collaboration as organizations expand. But some notable efforts to develop such standards are in progress.
Functions of Standards
Standards in healthcare informatics serve some of the same functions as standards in commerce and industry. Efficiency improves when uniform, interchangeable parts can be constructed and stockpiled in advance of need. Standard parts can be designed for multiple uses, so that one supply serves several needs. Standard components tend to become cheaper and more readily available as vendors compete to supply them.
The availability of standards for electronic funds transfer, electronic purchasing and invoicing have transformed commerce. An integrated information system allows a travel agent, or even a home computer user, to book a reservation on practically any airline flight. Similarly, a standard interface between clinical laboratory systems and clinical information systems means greater flexibility for the practice and a more timely way of receiving test data.
Types of Standards
- De Facto Standards
De facto standards are adopted by the marketplace, not by an official process. A de facto standard exists when many parties within an industry adopt a popular method or device. An example of a de facto standard is the bus structure of the IBM Personal Computer. Once several manufacturers offered expansion cards for the IBM PC, manufacturers of other computers adopted the same bus structure so that their computers could use the same cards.
- De jure Standards
De jure standards are enacted by law or regulation. For example, federal regulation mandates use of the UB-92 claim form by hospitals to submit claims for Medicare patients. The recent Health Insurance Portability and Accountability Act of 1996 (Kennedy-Kassebaum Bill) requires that HCFA adopt standards for several aspects of medical information. If suitable consensus standards are not available, the federal government may develop and mandate its own standards.
- Consensus Standards
Consensus standards are developed by agreement among a cooperating group of users of the proposed standards. Affected parties draft, refine, and vote on proposed standards until a given degree of consensus is reached. Unlike the case of de facto standards, such agreements are designed from the outset to create standards. Unlike de jure standards, consensus standards rely on voluntary adoption; they do not have the force of law. Consensus standards may be developed by industry or professional groups such as the American Dental Association or the American College of Radiology. Other organizations are formed specifically to create standards-- for example the American Society for Testing Materials (ASTM).
- Hybrid Types of Standards
One type of standard may acquire features of another type in time. A de facto standard such as the IBM AT expansion bus was formally described and adopted by an industry group -- it is now known as the ISA (industry standard architecture) bus. A de jure standard such as the International Classification of Diseases (ICD), developed for mortality reporting, may become adopted de facto for other purposes such as insurance claim diagnosis reporting. Eventually it may become mandated for this purpose as well, as in the case of Medicare claims. The Diagnosis Related Group (DRG) system was an ad-hoc method for measuring resource use that was adopted de jure for calculating hospital reimbursements.
Accreditation of Standards
Consensus standards can gain additional stature by being approved by organizations other than the one that developed the standard. Like the JCAHO which accredits hospitals, there are organizations that accredit standards organizations and their standards efforts. In the United States the principal accrediting organization is the American National Standards Institute (ANSI). Internationally, similar accreditation is done by the International Organization for Standardization (ISO).
Areas Requiring Healthcare Standards
Like nested Russian dolls, one set of standards provides the critical inner substrate for the next set and that next set is fundamental to the next, until the resulting complex set of standards finally permits full information transactions. In health care, the major areas -- from smallest and most basic to largest and most complex -- are as follows:
- Terminology
The "components" of healthcare information are clinical data. Is a patient encounter the same thing as an episode of care, or do they differ? Are crescendo angina and unstable angina the same or different symptoms? To make data useful in multiple settings, we must understand what the items mean and how they relate to each other. Hence we must agree on terminology, definitions, and clinical vocabularies. Classification systems such as ICD-9 allow us to fit clinical items into categories. Medical informatics likewise stands to benefit from standardization of materials and communication. For an example of one popular standardization of terminology, see NLM UMLS fact sheet.
- Information Model
But true understanding of clinical data involves agreement beyond the data items themselves, to the larger context not explicitly stated in an individual care record. We need accepted models for the structure of medical knowledge, for the structure and content of health records, for the health care system, and for the processes it uses. For example, assuming they all agree on the definition of unstable angina, how should we combine records on such a symptomatic patient when they may include data from physicians providing emergency care in one hospital, invasive cardiologists documenting angioplasty in another, and primary physicians' ongoing record keeping at the clinic or private office? Once everyone agrees on the definition of unstable angina, what of the standards on the care for patients with unstable angina in the ED, or the indications for angioplasty, or the guidelines for long term medical management? The HL7 web page at Duke University provides background on one popular information model.
- Connectivity
Once the models and meanings have been agreed upon we can expect a degree of understanding when information is shared, but we still need efficient ways of sharing the information. For example, the emergency department, cardiology suite, and primary physician all need access to the records of a patient with unstable angina, to obtain information and to record information for others. Standards ease the connection of medical instruments (e.g., serum analysis machine or CT scanner) to systems that store their data. Standards permit the transfer of health records from one provider or system to another, and aggregation into longitudinal records, community hospital information networks (CHINs), and knowledge bases.
- Policies
Agreement on concepts and the technical means to share information will not suffice unless appropriate policies for using the information are also in place. We need reliable ways of identifying the subjects, sources, and recipients of medical data so, for example, the emergency staff can have instant access to the records of a patient with unstable angina. Associating data with the wrong patient is potentially dangerous, and disclosing data to inappropriate recipients violates the patient's expectation of privacy and the professional's commitment to confidentiality. Education, research, and quality of care will benefit if we have the means to analyze patient information without disclosing the identities of the individuals involved. The goal is to provide ready access to information when it is needed for optimal care, yet at the same time protect patients' privacy and autonomy.
How do the Office of the Assistant Secretary for Health and the National Library of Medicine "put it all together" so advanced technology and informatics standards serve U.S. heathcare needs? One of the classic (and more readable) Standards article is on line and worth perusing: Making a Powerful Connection: The Health of the Public and the National Information Infrastructure.
Why Physicians Should Care About Standards
Should we who care for patients take the time and effort to be involved in creating standards? Isn't there enough to do already? Certainly. And that is part of the answer.
Standardizing the information tools dramatically improves efficiency, despite requiring the added task of updating and maintaining these tools. Frustrated clinicians may see these innovations as improving efficiency not for them, but for someone else; electronic claim submission has reduced effort and paperwork for both providers and insurers, and pharmacy benefits are now checked electronically and immediately for most retail prescriptions. With new standards in claims, for example, comes time spent by clerical staff dealing with requirements for more documentation, and continually changing lists of diagnosis and procedure codes. But the types and degree of detail of information gathered for claim submission are practically useless, if not burdensome, to a physician seeing the patient on a subsequent visit. We can send a prescription record across the country electronically and it will be understood, but not a medical record.
The reason our work remains inefficient is that ways to describe our patients and their needs are not yet standardized. Clinical standards are not easy to create; the information one needs to care for an individual patient with individual problems can be quite selective in scope and specific in content. The information that is critical for the physician and patient may be of little use for administrative purposes where the ability to aggregate patient encounters is paramount. The administrator wants to classify this patient and encounter to a grouping (ICD-9, DRG, etc.); the physician needs to understand how this patient is unique, and must have ready access to medical knowledge that fits the patient's circumstances.
Economic forces are driving the development of standards for administrative purposes because a clear financial benefit can be promised. This willl happen whether or not physicians are involved..But the benefit of better clinical information on improved quality of care -- for a patient, for the community or public health, for medical research and education -- is harder to quantitate, to forecast, and to realize. Without the participation of those of us who understand and demand the benefits of better information handling for improved patient care, and for medical practice in general, standards for medical data will remain focused on administrative uses and economic goals.
