Bibliometric Analysis, Case Analysis and Reappraisal on the Usability and Safety of Medical Devices

Ashwini Srinivasaprasad, Purdue University

Abstract

Healthcare is one of the most critical and valuable industries. This report is a systematic literature review of the usability and safety of medical devices used extensively for patient care. Safety has been broadly considered to mean in-patient safety and improving patient care. As part of the literature review and bibliometric analysis, a search of all the articles containing the keywords “medical device” AND “usability” AND “safety” was conducted on both Web of Science and Google Scholar (through Harzing’s Publish or Perish). The results that also contained the metadata were exported to VOS Viewer for cluster generation and co-citation analyses. As part of data mining, statistical tests such as regression analysis were conducted to find the causal relationship and significance between the variables. Following this, a core set of articles, some of which demonstrated strong ties to the topic at hand, were chosen from various databases and co-citation analyses. The core set also included a relevant chapter from the Handbook of Human Factors and Ergonomics, Fourth Editionby Salvendy. The articles were added to Mendeley and then exported to MAXQDA for the word-cloud generation. As part of the lexical search, Springer’s AuthorMapper and Digital Science’s Dimensions were used to extract trending keywords, author information, and metadata of relevant contributions made to the topic by institutions worldwide. The regression analysis revealed a significant tie between the database and the number of articles. Additionally, the regression showed that cites per year in Google Scholar had a significant effect on g-index. Finally, the discussion section included a case study of three different drug delivery systems. This aimed to perform a risk assessment of pre-existing “smart” medical devices. The analysis revealed that there is a high risk of malfunction in all the devices due to pre-programmed dose error. None of the devices were capable of “learning’ from patient behavior and adapting accordingly. Finally, the paper discusses Human-Computer Interaction and Human Factors Problems that exist in software medical devices that utilize AI/ML methodologies. Overall, the main keywords of human cognition, health informatics, precision medicine, applied ergonomics, and design safety showed high relevance within all the chosen articles. These keywords are highly reflective of the fields that are currently disrupting the conventional governance of healthcare.

Degree

M.Sc.

Advisors

Duffy, Purdue University.

Subject Area

Artificial intelligence|Cognitive psychology|Information science|Medicine|Psychology

COinS