Southern Health uses voice recognition technology for all dictated reports.
If the voice recognition engine is not adequately trained, then recognition accuracy can be very poor and, in some cases, dangerous.
Here are some examples:
"analgesics" recognised as "anal cheese sticks"
Endoscopic "retrograde cholangiography" recognised as "retrograde colon jogger"
Appendicitis, but "no evidence of a peri-appendiceal abscess" recognised as "no evidence of a para-penis or abscess"
"Dr. Flagiello" recognised as "Dr. Food Jello"
"Portable chest 12/15/06 8:16 AM" recognised as "Testicle taking Vicodin 816 am". When this one occurred on the first day of our changeover from human transcriptionists to computerised voice recognition, I knew I was in trouble.
"Patient was prepped and draped in sterile fashion" recognised as "patient was prepped and raped in sterile fashion"
"results were communicated by the resident to the Emergency Department" recognised as "results were communicated by the rectum to the Emergency Department"
"lungs are clear" recognised as "lungs are queer"
"History: Odynophagia" recognised as "History: Vagina dysphasia"
For a barium enema: "A large ulcerated cavity was identified..." recognised as "A large balls radiated cavity was identified..."
Remember to carefully read all your reports before send them off!