Re: design dilemma... 1 or several tables?

You might not remember the earlier ICD-8 codes, but 000.000 was Cholera. In FORTRAN IV, if you left a field blank it was read as all zeros. So as a result, about 1/3 of all hospital admissiosn were not "undiagnosed", they were cholera patients! This made doing statistics a lot of fun ...

I would do two things first. (1) Convert all the ICD-9 codes that map to a clear ICD-10 code with a program of some sort. They used to publish a conversion table and it should be available on mag tape or a CD. (2) Convert what you can of the ICD-9 codes that do not map to a clear ICD-10 code with a human being helping. Hopefully, there will be a relatively small number of cases that cannot be converted. Optimist that I am. But the doctor's rule about "look for a horse before a zebra" applies -- the common diseases are well known and I would guess they will map easily to the new codes. The new exotic ones will not map backwards.

I would go with the flag column telling you if the code next to it is ICD-9 or ICD-10 and a DEFAULT of ICD-10 on that column. Your goal will be to drop this column once the whole world is on ICD-10. Doing stats while you are in transition will be a bit of an adventure, but it is better than having to get the same information from two sources. Received on Thu Nov 22 2001 - 19:09:19 CET