The ICD-9-CM has been derived from the little changed statistical billing which has been this way for the last three centuries. The ICD-9-CM depends mostly upon the tabular data structure. This standard has no mechanism for the synonym, value restrictions, and inheritance or semantic and non-semantic it has got limited concept hierarchies.
This medical record system, which, should be integrated, has the major problem of the difference between the controlled vocabularies of the two systems. For this reason new standard is need to be adopted. Although the system developers find it difficult to adopt new standard as it does not fulfill there need.
ICD-9-CM codes are produced by the United States government. All right, title and interest in and to the copyrights and patents embodied in the Software are owned by BioSoftWorld or its licensor. License grants the single individual End User grants the purchaser permission to use the software on a network on a single site, with an unlimited number of end users. After installation User shall be permitted to make backup copy of the Software for archival purposes. BioSoftWorld maintains the software by taking feed back from the users
The characteristics of the system have been analyzed based on the various characteristics defined by Cimino & Campbell et al (Cimino, 1998; Campbell et al, 1999) and evaluation is given. The main features of the Cimino & Campbell et al (Cimino, 1998; Campbell et al, 1999) are:
It should be multipurpose-The vocabulary should be multipurpose i.e. it should be able to take care of the many purpose for which it may be intended to be used. There may be several purposes for which the vocabulary may be put to use. Some of them are stated as, for elaborating the clinical discoveries, for indexing the clinical records, for language processing. There can be many more purpose which can be added by the other readers
Comprehensive Content- The content of the vocabulary is the most important aspect. As for the real estate the most important thing is the location similarly for the vocabulary it is the content. The need for more term coverage is experienced everywhere. In an example which can be taken is the recent publication for the nursing terminology (McCloskey and Bulechek). In it the complete focus was on what can be deduced. But the solution of problem is beyond the simple addition of more terms.
As we can see that one solution can be to add more terms. Another solution can be enumeration of the simple units of a terminology and allowing the user to develop his or her own terms according to there usage.
Concept Orientation- The concept Orientation means that terms must correspond to at least one meaning (“nonvaugeness”) and no more than one meaning (“nonambiguity”), and that meanings correspond to no more than one term, “nonredundancy” (Cimino et al, 1994). On reading the medical informatics research, we shall find that the report using the controlled vocabulary is actually concept notations. It is now stated by the authors more explicitly that what they need are vocabularies that have more units of symbolic processing i.e. an embodiment of particular meaning (Evans, 1998).
On reviewing we see that there is some argument around the issue of ambiguity. As argued by Blois that as against the lower level concepts (such as protons) which can be exactly defined, high level concepts cannot be defined by necessary attributes but they need contingent ones, like the presence of chest pain in myocardial infraction (Blois, 1986). It should be understood that ay concept, no matter how fine grained, will always have a finer-grained concept into it. This context sensitive ambiguity is a different case from concept independent ambiguity (Cimino et al, 1989).
Concept Permanence- This is a derived concept from concept orientation. When a certain meaning has been given to a concept it must remain. The meaning should not change, although it may be flagged inactive. For example a word may be renamed by adding a certain word or character to it but the meaning should remain the same.
Non-semantic Concept Identifier-The concept must have a unique identifier so that each word may be associated with the concept. This can be easily done by giving a name to a concept. With reduction in the cost of storage this has however become less compelling. There has been an irresistible temptation of putting the unique codes in a hierarchical order. This makes it easy to be read by a human and it also makes to understand the relation s easier.
Certain problems also arise while using a concept identifier. These problems are basically of two types, the first one arise when we have to use a decimal coding. The second one arises when the same concept arises in many different places.
Poly-hierarchy- The hierarchical arrangement of the controlled medical vocabulary seems to be because of some universal agreement. This helps in grouping similar concepts, locating concepts, and conveys meaning. This can be explained in other words as if we see a concept cell in a certain concept anatomic entity we will understand a totally different meaning as against when we see this concept under the concept room.
The different users of concept demand for different hierarchy, which are equally correct, and this gives rise to the demand for multiple hierarchies (Cimino et al, 1989). Zweigenbaum and his colleagues believe that concept classification should be based upon essence of the concept, rather then arbitrary descriptive knowledge (Zweigenbaum et al, 1995). However this agreement about what the appropriate hierarchy should be can not be over, until there is a conscience about what the essences of concept should be.
Formal Definition- Many researchers and developers have indicated their need for a formal definition of the controlled vocabulary, in any form (Gabrieli, 1984). These formal definitions are usually expressed through the collection of links to other concepts in the vocabulary. This is explained by an example the concept “Pneumococal Pneumonia” can be defined by a hierarchical (“is a”) link to the concept “Pneumonia” and a “caused by” link to the concept “streptococcus pneumonia”. If “lung” has a site relationship with “pneumonia” then “pneumococal pneumonia” will also have relationship with “lung”. This data can be expressed in a number of ways which can be in the form of frame-based semantic network and graphs (Bernauer, 1991).
Reject “Not Elsewhere Classified”-A catch-all term is included to enclose all the information that has been left out, as no vocabulary can be complete for sure each time.
The phrase which usually accompanies these words is “Not Elsewhere Classified”. The problem with such words is that these terms do not have there own formal definition, rather their definition is based upon the definition of the concept with which they are used.
Multiple Granularities- With a particular process in mind each author demand for a different controlled vocabulary. With this each author also has some perceived levels of granularity up to which the concept must be expressed. This shows the fact that multiple granularities are needed for multiple purpose vocabularies. This is also proved by the fact that the vocabularies which will operate on a single level of granularity will be proved inadequate when the change in the granularity is required by the author.
Multiple Consistent View- For a vocabulary which is having multiple granularities, and is used for multiple functions there is a need of a multiple view s of the vocabulary, suitable for a spectrum of purposes. Although it should be kept in mind that we limit the ability to provide multiple consistent view so as to avoid inconsistent views. For example if we have concepts with multiple parents appear in several places, care should be taken that each concept is taken with a right view.
Representation of Context- It is difficult to use a standard controlled vocabulary, partly because of its standard nature it does not contain the specific concepts which have explicit meaning of the context in which it has to be used.
A need to have a controlled vocabulary to contain contextual information about the way concepts are used.
Evolve gracefully- The vocabulary has to change with the growth of the medical technology. This shows that even a perfect vocabulary was made; it will have to be changed with time. The desideration is that those involved with the controlled vocabulary should gracefully accommodate the evolution. By understanding the need to change the vocabulary, and also generating the ability to differentiate the good reason for the change from the bad, the graceful evolution can be easily done.
Recognize Redundancy- Redundancy is a condition which occurs in a controlled vocabulary when there are two terms which explain the same thing. In a controlled vocabulary redundancy is desirable. As vocabulary evolve gracefully or not they will be having this kind of redundancy
Conclusion
Cimino and Campbell have an excellent attempt to list out the indicators, which in general shall be considered in evolving the controlled vocabulary, which is multipurpose, shareable and uses a large set of medical terms. The paper of Cimino is not clear on the account of defining the term “Not Elsewhere Classified”. The coding shall be so evolved that it encompasses all medical terms so that singularities are avoided. ICD-9CM generally meets the Cimino and Campbell characteristics. There is a set official guideline for coding and reporting of ICD-9CM. This has made the software gracefully evolved and scientifically valid.
The ICD-9CM software right now is the most excepted software for representation of the condition of the patient. On account of the information mentioned in the above table, this software mostly fulfill the requirements. It is multipurpose, it has a multiple granulites, it can evolve gracefully, it sustainable it has redundancy, it is scientifically valid, it is well maintained and mostly all other characteristics are meet by this software.
References
- Bernauer J. (1991). Conceptual graphs as an operational model for descriptive findings. Clayton PD, ed. Proceedings of the Fifteenth Annual Symposium on Computer Applications in Medical Care (SCAMC). Washington, DC; New York: McGraw- HIll, 214-218.
- Blois M.S. (1986). The effect of hierarchy on the encoding of meaning. Orthner HF, ed. Proceedings of the Tenth Annual Symposium on Computer Applications in Medical Care (SCAMC). Washington, DC; New York: IEEE Computer Society Press, 73.
- Cimino, J.J. et al. (1989). Designing an introspective, controlled medical vocabulary. Kingsland LC, ed. Proceedings of the Thirteenth Annual Symposium on Computer Applications in Medical Care (SCAMC). Washington, DC; New York: IEEE Computer Society Press, 513-518.
- Cimino, J.J. et al. (1994). Toward a medical-concept representation language. JAMIA.1, 207-217.
- Cimino, J.J. (1998). Desiderata for controlled medical vocabularies in the twenty-first century. Methods of Information in Medicine. KFUPM.
- Campbell, K.E. et al. (1999). Enterprise Issues Pertaining to Implementing Controlled Terminologies. IMIA WG6: Conference on Natural Language and Medical Concept.
- Evans D.A. (1998). Pragmatically-structured, lexical-semantic knowledge bases for a Unified Medical Language System. Greenes RA, ed. Proceedings of the Twelfth Annual Symposium on Computer Applications in Medical Care (SCAMC). Washington, DC; New York: IEEE Computer Society Press, 169-173.
- Gabrieli E. R. (1984). Interface problems between medicine and computers. Cohen GS, ed. Proceedings of the Eighth Annual Symposium on Computer Applications in Medical Care (SCAMC). Washington, DC; New York: IEEE Computer Society Press, 93-95.
- McCloskey J. and Bulechek G. (1994). Letter: Toward data standards for nursing information. JAMIA: 1, 469-471.
- Zweigenbaum, P. et al. (1995). Issues in the structuring and acquisition of ontology for medical language understanding. Meth Inform Med: 34, 15-24.