RESUMEN
Emergency conditions can inhibit the use of clinical data, even when it has been professionally collected and structured. Our aim is to improve the assessment of patients' medical histories in emergency situations. Under emergency conditions, when it may be impractical for health care providers to interact with a visual display of patient data, a "speaking" medical record may be useful. We are investigating the use of a voice synthesizer in conjunction with MyAngelWeb, an Italian web-based medical record service. Only the textual subset of patient medical records was considered and restructured according to the needs imposed by voice communication. The quality of the received messages was tested. Some quantitative features, including the number of words and time durations, were considered together with other subjective features, including intelligibility of single words and overall significance of the voice messages. Provided that the linguistic architecture of a medical records' text is arranged to minimize the number of choices presented to the user, and phrases are kept short with few acronyms, health care providers can satisfactorily interact with the service.. Audible medical record delivery may be considered as an effective enhancement to those datasets needed in emergencies.