The CBC News reported on March 1, 2006, about a program designed by MedBridge to assist physicians and nurses in communicating with patients in different languages. The program had a trial run recently at Saint John, where hospital personnel saw it translate health terminology into languages such as French, Portuguese, Russian, Mandarin, and Cantonese.
The software lets healthcare workers ask questions in different languages by selecting from various possibilities on screen, and the patient then reads the questions, or sees them in an American Sign Language video (nice touch), and can respond accordingly.
The system runs on ordinary hardware and operating systems, so can be deployed now by anyone who wants it. And there is interest: hospitals in New York, Halifax, and Toronto are putting it to use.
Of course, similar systems have been developed for handheld computers and PDAs to allow tourists to communicate with locals, or soldiers in Iraq communicate with Iraqis, civilian or otherwise.
The limits of such systems are always the same. First, they are not true machine translation. Rather, they represent large databases of likely words and phrases for a particular context or situation, with a matching database for the target language. The advantages: speed, accuracy, and ease-of-use. The system doesn't have to accept dictation, process variations in pronunciation, idiom, dialect, and slang, or handle background noise, nor does it have to produce an audio response.
The disadvantages: if you step outside the context, the system fails completely. Also, the system creates an artificial barrier between patients and healthcare workers, cannot work with illiterate people, and may not, depending on the sophistication of the databases, have all the specialized terminology, idioms, or slang that people are prone to using.
Within the world of translation and interpretation, it is medical interpreters who are most likely to encounter this or similar systems. MedBridge points out that the interpreter can see a history of prior conversations with a patient, and so can seamlessly join the treatment process. At the same time, administrative staff may look upon such devices as replacements for human interpreters, when they are really just adjunct technology, a little help for when an interpreter is not available, or when the questions are simple enough that one is not needed.
For the intrepid tourist or a busy emergency room, such a device can be an invaluable aid. But it in no way will replace human interpreters in business negotiations, court depositions, or policy meetings, at least not for a long while. But MedBridge's system is an excellent application of existing technology, and should ease the burden of communication in hospitals in multilingual communities.