Code-Switching Automatic Speech Recognition for Nursing Record Documentation: Development and Evaluation (Preprint)

Abstract

BACKGROUND Taiwan has insufficient nursing resources due to the high turnover rate of nursing personnel. Therefore, reducing the heavy workload of these employees is essential. Herein, speech transcription, which has various potential clinical applications, was employed for the documentation of nursing records. The requirement of only one speaker facilitates data collection and system development. Moreover, authorization from patients is unnecessary. OBJECTIVE A speech recognition system for nursing records was constructed such that medical personnel can complete nursing records without typing or with only a few edits. METHODS Nursing records in Taiwan are mainly written in Mandarin, with technical terms and abbreviations presented in both Mandarin and English. Therefore, the training set consisted of English code-switching (CS) information. Next, transfer learning (TL) and meta-transfer learning (MTL) methods, which perform favorably in CS scenarios, were applied. RESULTS The word error rate (WER) of the benchmark model of syllables-based TL and the proposed model of syllables-based MTL was 29.54% and 22.20% WER in code-switching, respectively. The test set comprised 17,247 words. Moreover, in a clinical case, the proposed model of syllables-based MTL yielded a WER of 31.06% WER in code-switching. The clinical test set contained 1,159 words. CONCLUSIONS Medical personnel in Taiwan are often compelled to use a mixture of Mandarin and English in nursing records. Therefore, a Mandarin–English CS speech recognition system for nursing documentation was developed. The proposed data set has two characteristics, namely the medical field and CS, and lightens the workload of medical personnel.