To the Editor:

These remarks, in response to the letter “Until the lungs are removed …” [October 1999], are neither an endorsement nor a defense of voice-recognition software. Rather, they're a reminder that as bad as any of these systems may be, the physician's input is a key factor in the level of success.

Physicians are highly intelligent, well-educated people, but very often their writing and speaking skills don't reflect these attributes. These basic skills are a prerequisite for good records, whether those records are transcribed traditionally or by using voice-recognition software.

Too often, my transcriptionist colleagues and I must make judgment calls because a physician speaks too fast, doesn't pronounce words correctly, doesn't enunciate clearly and/or doesn't prevent background noises on the tape. Machines can't make judgment decisions, and transcriptionists shouldn't have to make them as often as they do.

Effective dictation training for physicians could prevent many problems, but I have yet to find a medical school or residency program that teaches this.

As we wait for all of the bugs to be removed from voice-recognition software, maybe medical educators can begin training physicians in the art and skill of medical record dictation as well as the nitty-gritty of the English language.