Karsten Hilbert <[EMAIL PROTECTED]> writes: > A mere listing of available drugs is most likely relatively > easy to obtain. It'll more an order of magnitude more > difficult to get interaction data, dosage, packaging info etc.
For transcribing one doesn't need all of the information - just the spelling and possibly use (to see if the medication fits to context). This, however, is insufficient for a medical doctor. There are some of these online. They're basically the Prescribing Information html'ized (it looks verbatim or extremely close to it). Some of these are very thorough. What if each drug were described in XML and we built a schema that the information must validate against (it must contain such and such; it's manufacturers must match "Eli Lili" or "Roche", etc). This would provide some rudimentry checks though not thorough. OTOH, if a drug database was somehow linked to patient records it may be possible to allow a user, such as a nurse or a "trusted client" to enter the raw data in his/herself. I suppose, though, that a benefit for a doctor is if he or she could diagnose a patient with, say, an ICD9 code. Using this code then looks up all the medications in the DB whose indications are the same ICD9. So, the drug information itself would have to be ICD9-coded (I suppose; I have yet to take that course and know little of ICD9). Do you think that drug manufacturers would be willing to help us with the raw data? That is, if we explain to them the project and it's goals and just simply ask, "May we please have the raw prescribing information in an ASCII format?" that they may oblidge? Ask and you shall receive. Elizabeth
