John,
Thomas' future-proof paper is great. The mention of SNOMED and ontologies gets at the notion of "medical concepts", and how those concepts are utilized in an information system. Different people in the same company will interpret a particular illness differently. Ontology and a well defined lexicon helps standardize data use and remove ambiguity, besides promote good system design.
Another way to envision what Dr. Ho is referring to is to think of your medical data in small, well defined sets. A data model is a way to conceptually represent information in a computer system, so for a hospital, you might have the following data models:
- a patient: demographic information, diagnosis, treatment, billing information, insurance information.
- a clinician: nurses and doctors that care for patients
- support staff: people that help out (IS employees, janatorial)
- administration: finance, HR
- operations assets: heart monitors, needles
- consumables: medicines, bandages
You can try to define those things in a functioning system as a complete unit by combining them all together somehow (the old monolith approach), or you can define each thing in its own way (micro-models). The point is that your resulting software will be defined in terms of the data model, so you can associate and all-encompasing model with a mainframe and a multi-model approach with decentralized/distributed computing (just as an example).
Gage wrote:
> > I take it that, in general, data concepts are being
> proposed that fall into
> > two categories: general and specific. The general data
You can think of them that way if you want, but you might get farther if you think of data in terms of required and optional. Required are ICD-9 codes, billing information, and anything else you must keep to manage the patient without making mistakes or being sued/fined/banned. An EMR system at a minimum will have to have the required fields, and it needs to be handled in a certain way. For example, data cannot be destroyed.
You asked before about HL7. HL7 is a communications protocol developed from a health care data model, but it is not a data model in and of itself per-se. The HL7 group does define some great data models upon which to build HL7 though!
How's that?
Richard Schilling
Webmaster / Web Integration Programmer
Affiliated Health Services
http://www.affiliatedhealth.org
