As speculation continues to swirl around AI’s forthcoming transformation of healthcare, fueling boom times in AI research, a review of the literature has turned up scant evidence the technology is benefiting patients at the consumer level.
“Without a clear understanding on why patients and consumers need AI in the first place, or how AI could support individuals with their healthcare needs, it is difficult to imagine the kinds of AI applications that would have meaningful and sustainable impact on individual daily lives,” write the authors, who were led in the project by Annie Lau, PhD, of Macquarie University in Australia.
Lau and team arrived at their conclusions after systematically searching PubMed for studies published in calendar year 2018.
The full study is running online in Yearbook of Medical Informatics.
From an initial crop of 99 papers meeting their inclusion criteria, the researchers culled 14 as “best paper” candidates and presented them to a panel of international experts for full review and scoring.
Three papers emerged as the best papers published in 2018 on AI in healthcare for patients and consumers. All three had in common the use of data-driven algorithms and insight-led approaches, the aim of which was to reveal patients’ experiences with online health information.
The three top studies were:
- Abdellaoui et al., “Detection of Cases of Noncompliance to Drug Treatment in Patient Forum Posts: Topic Model Approach” (Journal of Medical Internet Research);
- Jones et al., “Novel Approach to Cluster Patient-Generated Data Into Actionable Topics: Case Study of a Web-Based Breast Cancer Forum” (JMIR Medical Informatics); and
- Park et al., “Examining Thematic Similarity, Difference and Membership in Three Online Mental Health Communities from Reddit: A Text Mining and Visualization Approach” (Computers in Human Behavior).
En route to identifying these papers as the best of the bunch, Lau and co-authors found no studies on AI applications designed specifically for patients or consumers.
Likewise, their searches turned up no studies that sought and obtained patient and consumer input on AI.
Going by the literature, the most common use of AI for patients and consumers is finding and perusing materials online, the authors conclude.
At the same time, it’s unlikely many are aware of the technology’s role in the process.
“For patients and consumers to truly benefit from AI, the design of the technology may need to be embedded deeply in their environment or perhaps even invisibly in their daily routine,” Lau et al. comment in their discussion section. “In addition, real-life decision support for patients and consumers remains an open opportunity, provided the right problem, use case and interaction mode are identified.”
The authors close their comments with a quote from a Nov. 2018 Nature article on managing expectations around AI:
“The public’s view of artificial intelligence may not be accurate, but that doesn’t mean that those developing new technologies can afford to ignore it.”