Thanks, Tim, for these detailed responses. First a note to Adrian: you and I will be long dead before English/French/German/etc. text is eliminated from the medical record. Progress notes are the fundamental, the essential way through which clinicians communicate. It will not change. These notes are *messages* to other members of the clinical team. We don't speak or write in ICD-9, HL-7, SNOMED, etc., etc., etc. We never will. In addition, the practice of medicine is characterized by nothing so much as it is by uncertainty. Progress notes are the perfect forum to capture our uncertainty. Finally, our duty to communicate with our patients can *only* be accomplished through natural language. One of the attachments to a typical message would be patient instructions.
Within an institution, I don't know whether encryption is really needed. And most communications will be within institutions: we rarely write notes on charts in other hospitals (unless we have privileges in several, but, once again, those notes are usually written on site). I originally thought that encryption would cover most security issues, for example between a practitioner and a pharmacy, but I soon realized that even knowing where the message originated from was a breech of security by knowing the message headers. I reiterate that the answer to this is placing a mail transfer agent on each client. At that point, even the message hearders can be encrypted, and someone evesdropping will get nothing but garbage. So I'm really not proposing special routing tables although that will be ideal in certain situations, perhaps generally, I'm really proposing a mechanism wherebye the sender and the receiver are the only ones who have a prayer of decrypting the message. Now, I know that this ignores the problems of key theft etc., but that is a different problem. Seeing that many mail clients have automatic encryption, I don't see encryption as a tremendous problem. What I see as a tremendous problem is that the is no EMR in 2002. John
