International Journal of Medical Informatics 48 1998 5 12 What is done what is needed and what is realistic to expect from medical informatics standards Clement J McDonald J Marc Overhage Paul Dexter Blaine Takesue Jeffrey G Suico Indiana Uni6ersity School of Medicine Regenstrief Institute for Health Care 1001 W 10th Street RG 5th Floor Indianapolis Indiana 46202 USA Abstract Medical informatic experts have made considerable progress in the development of standards for orders and clinical results CEN HL7 ASTM EKG tracings CEN diagnostic images DICOM claims processing X12 and EDIFAC and in vocabulary and codes SNOMED Read Codes the MED LOINC Considerable work still remains to be carried out Abstract models of health care information have to be created to cover the necessary domain and yet be simple enough to assimilate implement and manage This requires a high degree of abstraction Enormous amounts to develop standardized vocabulary are still required to complement such a model and to define the subsets that apply to given contexts 1998 Elsevier Science Ireland Ltd All rights reserved Keywords Message standards Vocabulary standards Computer automation Electronic medical records 1 What is done The word done is used loosely We are never really done with anything especially as it relates to information systems The clinical systems world is like a fractal and when we look very closely at the layer that we understand we see another layer of complexity just below the current one to digest and standardize Corresponding author I recall a document called Current Medical Terminology published by the American Medical Association in the early 1980s It bragged about having completely catalogued all of the terms in medicine It contained a mere 5000 entries Today s medical dictionaries such as Stedman s 1 contains 80 000 or more entries The current releases of SNOMED 2 systemized nomenclature of human and veterinary medicine and Read Codes 3 contain many hundreds of thousands of concepts and are still growing 1386 5056 98 19 00 1998 Elsevier Science Ireland Ltd All rights reserved PII S 1 3 8 6 5 0 5 6 9 7 0 0 1 0 2 0 6 C J McDonald et al International Journal of Medical Informatics 48 1998 5 12 ICD10 PCS is a set of tables 4 being developed for health care financing administration HCFA by 3M Corporation and contains more than a million procedural codes Nathan Myhrvold s assertion that there will always be a software crisis because ambition absorbs all advances 5 applies equally to standards As soon as we sniff success in one domain of standards we find other targets of opportunity and extend our expectations Yet medical informatics and the industry at large has or is well on the way to solving many standardization problems 1 1 Message standards I think we have solved the standardization problem of transmitting patient clinical information reports from one system to another Laboratory results can be sent via Health Level Seven HL7 ORU 6 American Society for Testing and Materials ASTM 1238 7 and Comite Europe en de Normalisation CEN ENV 1613 8 messages radiology and many kinds of clinical observations and x ray reports via HL7 and ENV 12539 9 messages diagnostic images themselves via Digital Imaging and Communications in Medicine DICOM 10 messages and EKG tracings via a CEN standard 11 that I like to call the Jos Willems standard due to him being the main force behind it It is of interest to note that our last Burdick EKG cart came with the Jos Willems standard installed Most of the above standards have been in active use for many years For example in the US the vendors of just about every type of clinical system from GI endoscopy OB ultrasound cardiac echo systems laboratory systems radiology information systems pharmacy systems to nurse triage systems advertise support for HL7 standards although the fidelity to the standard varies Similarly the distribution of ADT registration information has been mastered via HL7 and CEN ENV 12538 12 the transmission of community pharmacy information messages via HL7 and National Council for Prescription Drug Programs NCPDP 13 messages as well as the transmission of simple claims via Accredited Standards Committee ASC X12 14 messages to name a few The health care industry understands these problems has substantial experience with the standards and can show hosts of successful implementations Orders messages have also been standardized and many successful implementations exist However it is not certain as to whether we have mastered the communication of clinical orders to the same degree as clinical observations In part this may be due to orders being more complicated beasts than clinical observations Orders mutate in some institutions radiologists change as many as 30 of ordered procedures to different and more appropriate procedures They explode a request for routines becomes a request for urinalysis complete blood count and electrolytes They repeat an order for 9 am glucose is repeated each day They cascade an abnormally low hemoglobin may trigger tests for iron if the median cell volume is low and other tests if it is high However the real challenge of orders may arise from a more fundamental difference between orders and observations Observation messages interact mostly with the receiving computer and can consider themselves successful if they load properly into the receiving clinical application An order message interacts with the receiving system but it must also interact with the people who fulfill the orders This implies a much larger set of assumptions about behaviors and processes in the receiving system and may require the standardization of these processes The fact C J McDonald et al International Journal of Medical Informatics 48 1998 5 12 that people are in the loop adds to the challenges Wicked Problems Righteous Solutions 15 is an informative book about software development strategies It is easy reading well referenced and provides many good insights One of its axioms is that the computerization of any system that depends upon human behavior for its success is automatically a wicked very difficult problem Due to the fact that we cannot be sure as to how the human will behave within the system the development must include some trial and error cycles 1 2 Terminology Successful has also been achieved with the development of standard codes vocabularies for base clinical concepts mostly nouns and simple noun phrases that might be stored in the