Peter Damoc wrote: > My intention is to develop a software that could potentially be used > in small clinics like our own but I dream about making it flexible > enough so that maybe it could be adapted to small hospitals. In Open Source it is sometimes better to take something designed for running eg a complete health service, and just use the bits you need, rather than re-instantiating those bits from scratch.
The concurrent entry in two copies problem is smaller if the medical record is designed properly, since almost all changes are actually additions. The number of changes which are edits - deletions of existing values or changes of them in place should be approximately zero. Identifying each computer, or each copy of the database, with a global unique ID and including that in the narrative should fix most of that. > - additional data for each intervention, for example some rhinoplasty > simulation (with patient consent via a fingerprint reader) You know your requirements, but I wonder whether recording sound and images of the patient giving consent is better than just a fingerprint? Have you looked at the use of cryptographic hashes and hash trees of records as a way of demonstrating at a later data that the element --> record --> complete collection of records is in a condition unchanged from the time it was made or hashed? It seems prudent. -- Adrian Midgley GP, Exeter, UK _______________________________________________ Gnumed-devel mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnumed-devel
