Speech-based Documentation in Emergency Medical Services with the Electronic Language Interface for Ambulance Services

Emergency medical services (EMS) assume a leading
role in case of an emergency: They provide the professional
preclinical care on site and decide on the further treatment
of the patient. In order to trace the treatment process of
the patient after a rescue operation, information such as drug
administration are documented in an operation protocol. Such
documentation is often done following the rescue operation. Due
to the high psychological strain and stress, the retrospective
documentation is error-prone. In Switzerland, about 80% of EMS
use paper-based protocols. While electronic protocols ensure
data sharing to third-party systems, no process improvements
are guaranteed. In a Wizard-of-Oz experiment with speculative
design, we investigated how media can support paramedics in
the documentation process. Since voice records were evaluated
as a promising tool in that study, ELIAS, the Electronic Language
Interface for Ambulance Services was developed. ELIAS is based
on a digital emergency protocol and allows paramedics to use
a speech-based user interface to document measurements at the
time when they are performed. We run tests to study the word
error rate during transcription in different settings. Furthermore,
an analysis with ambient noise has been performed. The results
let us conclude that once implemented in daily practice, ELIAS
has the potential of increasing the semantic and time-related
accuracy of emergency protocol data.