Exploring new digital therapeutics technologies for psychiatric disorders using BERTopic and PatentSBERTa

Eunji Jeon, Naeun Yoon, So Young Sohn

Research output: Contribution to journalArticlepeer-review

17 Citations (Scopus)


Coronavirus disease 2019 (COVID-19) has accelerated the growth of the digital therapeutics (DTx) market; therefore, development strategies for new DTx products are necessary to satisfy market needs. However, data-driven methods for recommending digital healthcare technologies for novel DTx applications are scarce. We propose a technology opportunity discovery framework that recommends 1) potential technologies as new DTx products, and 2) the applicable target disorders. We applied BERTopic and PatentSBERTa to patents filed with the United States Patent and Trademark Office and calculated the score of potential technologies on the basis of their thematic characteristics with respect to their digital capabilities and similarity to DTx technologies. By identifying the target disorder of similar technologies, specific disorders were proposed that can be treated with the proposed technique. By applying the proposed framework to psychiatric disorders—one of the largest therapeutic areas of DTx, we recommend digital monitoring technologies applicable to poor breathing or sleeping patterns for cognitive impairment. Furthermore, we provide strategies to utilize the recommended digital technologies for DTx for specific disorders to facilitate a direct intervention or treatment, which can contribute to the planning of roadmaps for DTx.

Original languageEnglish
Article number122130
JournalTechnological Forecasting and Social Change
Publication statusPublished - 2023 Jan

Bibliographical note

Publisher Copyright:
© 2022 Elsevier Inc.

All Science Journal Classification (ASJC) codes

  • Business and International Management
  • Applied Psychology
  • Management of Technology and Innovation


Dive into the research topics of 'Exploring new digital therapeutics technologies for psychiatric disorders using BERTopic and PatentSBERTa'. Together they form a unique fingerprint.

Cite this