CS99I Freshman Seminar

Winter 1997/1998.

Traveling the Information Highways: Healthcare

Maps, Encounters, and Directions

Master copy on Birch
Draft 27Nov1993, rev 1feb94,4Apr94, 9May94, 30May94, 20Jun94, 19Oct1994
This material is

©Gio Wiederhold and CS99I students, Stanford University, 1998.

Chapter: Healthcare

Previous chapter: Digital Libraries -

Next chapter: Electronic Commerce

HEALTH.Intro

Healthcare information is taken to be a crucial component of the information flow along the digital highways. Healthcare today represents nearly 14% of the Gross National Product (GNP) and at least 20% of its cost is related to information collection and processing. The corresponding percentages are 6.6% in Great Britain and Japan, in Germany 8.5%, in Canada 9%, and in France 9.1%. The British National Health Service (NHS) is Britain's largest purchaser of telecommunications.

An estimate by the U.S. Department of Health and Human Services (DHHS) has used an estimate that a fully developed HealthCare Information Infrastructure (HCII) could save the U.S. economy $100 billion between 1992 and the year 2000 [HHS93].

The traffic pattern of health information differs greatly from that of other applications using digital highways. Much of the traffic is local. A great deal of medical care is delivered in the final years of ones life, and at that time one does not move around much. There are of course cases where rapid access to ones medical record at a remote location can be crucial, and being able to obtain a good record in digital form can make a big difference. That the record should be good must be stated explicitly. Many medical records being kept today are not good. They are often partial, because care was delivered at multiple sites, they are medically incomplete, because they focused on billing and legal accountability, and they are hard to interpret, because of diversity in coding schemes and habits.

After a brief history, focusing on desirable structure of the medical record, we will summarize the functions of the medical record with respect to demands being made for the digital highways.

HEALTH.History

Systematic collection of information about a patient was only advocated in !!1890>> by !!xx>>. Earlier records were sporadic, and a patient's disease would only be described if it was particularily interesting. The emphasis remained on patient's hospitalization records, and by 193!!x> all hospitals had medical records departments. keeping of an adequate medical record is a requirement for accreditation of hospitals. A hospital which is not accredited cannot receive reimbursements for its services from government funds, and will also have difficulties in settling claims with private insurers, so that hospital records are universal.

Private physicians have guidelines on the keeping of records on the patients they see, but there no oversight today. In a clinic or a group practice the records should be adequate so that a !!replacement> can provide care to a patient. Only the potential need to defend oneself legally is a direct inducement to keep careful records about ones' patients. Still many physicians keep careful records, following the habits acquired in medical school.

In 195x? Larry Weed presented a formal structure to collect medical records. The Problem Oriented Medical Record (POMR) became a model for the teaching of keeping patient records. The record for a patient visit is to consist of four entries for each problem:

S A Subjective statement of the problem, as expressed by the patient.
O A set of Objective diagnostic tests.
A The Assessment, based on the subjective and objective information.
P The final entry is the plan, including a the presecribed treatment and prognosis, a statement of the expected patient state at the next planned encounter.

Any new problem gets a new problem number for further reference. For a follow-up visit the earlier problem statement should be referenced. Sometimes subsequent tests may show that the perceived problem was actually a result of a pre-existing problem, or that one problem had two causes. Then problem assignments can be merged or split.

Having a prognosis enables feedback to occur about one's prognosis and plans. Unfortunately, the prognosis entry is the least consistently applied aspect of the POMR. Without being explicit about the expectations for treatmet it is difficult assess if the plan is successful.

HEALTH.Functions

The central information resource in health care is the * medical record. A medical record is supposed to document all interactions of an individual with the health care system. Its objective is to assure that any new action to be taken will be beneficial, and not cause harm due to interference with prior treatments, or particular patient !!status> as having allergies to a drug. We will briefly discuss the function of the medical record for individual care, but then move on to the functions of the record that interact with the topic of this book, the national information infrastructure.

HEALTH. Functions.patient-careXxx

HEALTH. Functions.care-monitoring

HEALTH. Functions alerts

Many tests to determine the health status of a patient are performed at laboratory remote from the patient or the provider. The laboratory reports the test result to the provider, for entry into the patients medical record. Increasingly, such results are reported electronically, reducing the delay and increasing the responsiveness of the health care system. A standard for electronic reporting of diagnostic results has been developed, so that one laboratory can serve many providers and a provider can use the most suitable laboratory without creating a confusion that mirrors the many different pieces of papers used now [ASTM!!1287>]. If a test indicates that there is a serious problem additional actions can be taken. An * alert can be generated, perhaps advising the provider to call on the patient or to advise the patient to make an appointment. To reach the provider or patient rapidly a telephone message may have to be created. A * mediating computer module in the providers office can scan incoming laboratory results, and, following rules given by the provider, determine the seriousness, formulate an appropriate message, and initiate any needed calls. A log should be generated so that one can review what phone calls were generated and if they were received. If *!!universal service> is available many alternatives to forwarding alerts become available. Rapid processing of alerts is likely to reduce both the cost and the pain suffered due to health problems.

Alerting can be improved by exploiting the * prognosis entry in the !!POMR>. A result value that might be serious for one patient may be expected for another. Since any alert will cause concern, the limits for alert generation are often set quite wide. For instance a weight gain that is in the normal range for a teenager may indicate a pulmonary embolism in a older patient with a heart problem. The use of the patient medical record, and specifically prognosis field greatly increase the relevance of alerting. Alerts can also be generated when expected events do not occur: no laboratory result was received when expected, or the patient failed to make an expected appointment.

HEALTH. Functions.clinic-management

HEALTH. Functions.HMO-management

Eligibility and Billing

Most patients are covered by some form of health insurance, but the extent of their coverage varies. It is important for providers, say a physicians office, to determine if a patient is covered fully, if some co-payment is required, and how to bill the insurer. Such services can be effectively provided over the Internet now. The office can either contact the presumed insurer directly, or an intermediate service can establish the connection.

Security and protection of privacy is a major concern here. The insurance company must be assured that the care provider and the patient are who they claim to be. The provider wants to be assured that a the bills will be paid eventually, and the patients do not want the record of their problems to become public knowledge.

HEALTH. Functions.epidemiology

When information from multiple sites is assembled we encounter the problems of having a consistent terminology. This problem !!bedevils> us throughout when traveling the digital highways. In Sect. Library.Technology.UMLS we describe a major effort to create a consistent * ontology for medicine.

HEALTH. Functions.research-support

HEALTH.Technologies

HEALTH.Technologies.UMLS

!referenced in health and mediator chapter. best place?
A major effort to create a consistent terminology has been supported since 1986 by the National Library of Medicine (NLM) as the Unified Medical Language System (UMLS) [Lindberg:93]. Medical institutions in all parts of the U.S., as well as some in Europe participate. The central part of UMLS is a * metathesaurus, which provides cross references to three major sources. The metathesaurus also provides the linkages, as shown in Fig.\terms, which is a critical addition, since a simple combined list would increase, rather than reduce confusion. In Chap. MEDIATORS.Ontologies we discuss how one can carry out algebraic operations on such thesauri.

Figure:. Example of an 'isa' inheritance (not yet in)

All terms are categorized into one or more of 132 semantic types. A total of 47 relationship types are recognized among them. The semantics of UMLS are much richer than those of the * structural model presented in Chap. DATABASES.Technologies.structural-model, making the definition of a suitable algebra that much harder. However, the primary link type is * ownership (or 'isa'), where a lower level term specializes a superior concept, and inherits by default the attributes of its superior, as shown in .smoking. Negative feedback, as implied there, is of course, crucial to achieving stability in control theory.

!!at Columbia [cimino in Scherrer:95 UMLS expanded for clinical processing has 108,000 strings with 145,000 relationships

Figure: Smoking: An illustration of negative feedback in a chain of causal relationships(not yet in)
smoking . increased blood pressure . heart attack . reduction of smoking

A total of 15 other thesauri and vocabularies contributed information to UMLS. Major, active sources are the * LoC list of Subject Headings (LCSH), and * MEDLINE, * MeSH \tm,.
!also used were AI/RHEUM, Physicians Data Query (PDQ), DXplain, Quick Medical Reference (QMR), Dorland's Illustrated Medical Dictionary, Online Mendelian Inheritance In Man (OMIM), ICD9-CM >

Like any ontology, UMLS requires long-term knowledge maintenance. It is now updated annually, but quarterly updates are contemplated. Some of the changes needed are to correct errors and omissions, but others will acount for the *evolution in the use of the terms. The existence and utilization of a sharable ontology brings the community and its use of terms together through practice. The latter is probably the most important result in the long run. By having tools more people will share concepts and naturally evolve towards common meanings for terms, just like the Kings James Bible provided a major impetus towards a consistent spelling in the <<16th>> century. The focus of UMLS research is indeed now the development of applications, to obtain user coherence and feedback. As such UMLS, in its domain, is a good precedent setter for HPCC and the general effort to establish highways of the future. The medical domain does have the advantage that its participants are well-educated and demanding of high quality services. Since healthcare delivery is costly, even relatively modest benefits can have a high financial impact.

HEALTH.Technologies.xxx

HEALTH.Alternatives

HEALTH.Alternatives paper

Paper allows for rapid scanning. Even if the information is electronically transmitted many health-care providers routinely print everything out, scan the data, and then file or discard the paper.

HEALTH.Alternatives.portable-record

A portable record is one that is carried by the person from one health-care facility to another. The U.S. military use such a record in paper form now, since it is more reliable than their communication systems and can accommodate all the media being used. A modern record medium could be floppy or a CD-ROM disk. Fot computer media standards for the format must be defined and and equipment supporting these standards would have to be widely distributed.

With portable media confidentiality remains a problem. Although the record does not go over multiple Internet nodes, it is likely that

The owner of the record can read it all. This is often not a problem, but at times physicians may want to withhold a depressing prognosis from a patient.
Each healthcare institution would be able to access the entire record, even if the information is not relevant to their task. Should the pharmacist have access to a past psychiatric record? Elaborate encodings could be developed, but would have to be widely standardized.

Since the general medical establishment will not entrust the patient with their medical records, there remains an important role for the interchange of medical records ovr the Internet. Organizations such as the Health Care Information Corporation (HDIC), a California non-profit group, are working at establishing secure Internet-based communication for such data.

Medical Systems Useful on the Network
Name	organization	field	when	[ref]	\| %source
ABDA-Pharma		drug		[Miller92]	\|
Ageline	AARP		sociology, psychology, services for elderly		[Miller92]	\|
AI/RHEUM		clinical, immunology		[Miller92]	\|
AIDSLINE		clinical, HIV		[Miller92] []\|
Cancer Weekly	NCI	clinical literature		[Miller92]	\|
CHEMLINE		biochemistry		[Miller92]	\|
DxPlain	Mass.Gen.Hospital	diagnoses			\|
EMI		environmental mutagens		[Miller92]	\|
FEDRIP		federal research in progress		[Miller92]	\|
HEALTH			health planning and administration		[Miller92]	\|
ICSD		inorganic crystal structures		[Miller92]	\|
ISHOW		hazardous organics in water		[Miller92]	\|
JICST	MITI, Japan		medical science literature abstracts		[Miller92]	\|
Lithium Library		psycho<> drug		[Miller92]	\|
Martindale Online		drug		[Miller92]	\|
Medical Forum	AAMSI	clinical articles, discussions			\|
MEDLINE	NLM		clinical		[Miller92]	\|
OMIM	Johns Hopkins		genetic diseases		[Miller92]	\|
PDQ	Physician Data Query	clinical		[Miller92]	\|
PDR		drug refernce		[Miller92]	\|
PHARma projects		drug		[Miller92]	\|
PharmLINE		drug		[Miller92]	\|
QMR	CAMDAT Corp.	Quick Medical Reference			\|
SEDLINE		drug		[Miller92]	\|
TOXNET	NLM, NIOSH	toxic substances, several files	\|