Virtual patients: assessment of synthesized versus recorded speech

Virtual patients have great potential for training patient-doctor communication skills. There are two approaches to producing the virtual human speech: synthesized speech or recorded speech. The tradeoffs in flexibility, fidelity, and cost raise an interesting development decision: which speech approach is most appropriate for virtual patients? Two groups of medical students participated in a user study interviewing a virtual patient under each condition. We found no significant differences in the overall impression, speech intelligibility, and task performance. Our conclusion is that if the goal is to train students of which questions to ask, synthesized speech is just as effective as recorded speech. However, if the goal is to teach the student how to ask the correct questions, a high level of expressiveness in the virtual patient is needed. This in turn necessitates the higher cost - even with the lower flexibility - of recorded speech.