Evaluation of the Usefulness of Speech Recognition Tools in Creating Medication Therapy Management Records

Takuya MURAO; Akinori MARUO; Yohei MANABE; Takayoshi KAWABATA; Taro MIURA; Azusa NAGAHIRO; Takashi MAKITA; Hidenori SAGARA; Hirofumi HAMANO; Hideki MORI; Yoshito ZAMAMI

Return

Evaluation of the Usefulness of Speech Recognition Tools in Creating Medication Therapy Management Records

VernacularTitle:薬剤管理指導記録作成における音声認識ツールの有用性の評価
Author: Takuya MURAO ¹ ; Akinori MARUO ¹ ; Yohei MANABE ² ; Takayoshi KAWABATA ¹ ; Taro MIURA ² ; Azusa NAGAHIRO ¹ ; Takashi MAKITA ³ ; Hidenori SAGARA ¹ ; Hirofumi HAMANO ¹ ; Hideki MORI ¹ ; Yoshito ZAMAMI ⁴
Author Information

1. Department of Pharmacy, Okayama University Hospital
2. Equal contribution
3. Department of Clinical Pharmaceutics, Okayama University
4. Graduate School of Interdisciplinary Science and Engineering in Health Systems
Keywords: keyboard input; speech recognition tools; ward pharmacist; workload; medical record
From:Japanese Journal of Drug Informatics 2025;27(2):58-65
CountryJapan
Language:Japanese
Abstract: With the expanding responsibilities of pharmacists, the time required to create medical records has become a significant burden, highlighting the need for efficiency improvements. Among text input methods, speech recognition offers a faster alternative to keyboard input and is expected to reduce workload. This study evaluated the utility of a speech recognition tool in generating medication management instruction records. First, ward pharmacists rated subjective changes in their workload before and after using the tool on an 11-point scale (ranging from −5 for deterioration to ＋5 for improvement). The median score was 0 (range: −4 to ＋2), indicating no perceived improvement. Next, two sample texts (349 and 605 characters) were transcribed using both the speech recognition tool and keyboard input, with input times measured. Speech recognition required 186±15 and 395±24 for the first and second texts, respectively, compared to 252±19 and 511±26 s for keyboard input, demonstrating a significant reduction in time. Although no subjective decrease in workload was observed, these findings suggest that speech recognition tools may enhance efficiency in record creation. Further research in this area is warranted, encouraging audience engagement in the ongoing scientific process and fostering a sense of shared responsibility.