Speech Recognition in Healthcare: A Comparison of Different Speech Recognition Input Interactions

Speech recognition (SR) in healthcare has the potential to improve clinical documentation and communication among healthcare providers. In this respect, there are different ways of voice interaction. The purpose of this study is to compare four different speech recognition input interactions for healthcare form filling, without implying a specific SR algorithm, and thus to measure which interaction is the best in terms of time spent, error rate, and participant preference. We conducted a user study involving 15 participants, divided into three groups according to their computer usage expertise. During the experiment, participants filled out forms using four different SR interactions and traditional typing. The results showed that the interaction in which participants say all the data in a single sentence has the lowest time spent by participants, with an 83% advantage over the traditional keyboard/mouse interaction. Despite this, the System Usability Scale questionnaire revealed that the participants preferred the interaction in which they converse with a vocal assistant. Our results also found that SR interaction reduces the difference in input time between occasional and expert users.