Studies reveal how AI-infused clinical decision support may change medicine – for the better

As trust in AI evolves – and adoption increases – the benefits are becoming tangible

By and Mehool Patel | 1 minute read | March 17, 2021

In many ways, AI has been applied in medicine, such as in drug development, business intelligence and patient care. It has also been applied to clinical decision support tools – where it can help rapidly surface insights from vast libraries of biomedical information to facilitate fast, informed decisions.

For clinicians to be comfortable with AI in supporting decision making, trust is essential. Trust can be influenced by several human factors, such as user education, past experiences, user biases and perception towards automation, as well as properties and the reliability of the technology itself.1 However, nothing builds trust like results. Over the past two decades of AI use in healthcare, a body of evidence is emerging on how AI can help in areas such as clinical decision support.

Put AI to work across your entire clinical enterprise

Emerging data: One client reports AI-assisted answers in less than one minute

An IBM Watson Health client, TidalHealth Peninsula Regional, recently integrated AI-powered clinical decision support into its EHR. TidalHealth clinicians used the solution 489 times in the first month, compared to just 275 searches during the previous six months – nearly a tenfold increase. Time studies also revealed that 73% of AI-assisted searches provided answers in less than a minute, in some cases saving up to three minutes over traditional search methods. Watch the video to learn more about this study:

AI-assisted searches are also increasing the breadth of returned information. A recent study2 examined clinician interactions within IBM Micromedex with Watson – an AI-enabled drug information reference database used by more than 4,500 healthcare organizations. The study looked at the differences between the solution’s AI-assisted searches and those conducted using its standard keyword-based interface. While dosing information was the most commonly-searched topic using both methods, the study3 found that AI-assisted searches mapped to a broader distribution of content than conversations using the standard interface. While future research will delve into these differences, these initial findings suggest that AI-assisted search may enable users to get richer answers to questions.

There is more to explore and learn about how AI can assist clinicians when making decisions. As adoption continues to increase, it is exciting to envision how the application of AI in clinical settings may lead to more efficient, evidence-based care.

 

Refences:
  1. Asan O, Bayrak AE, Choudhury A. Artificial Intelligence and Human Trust in Healthcare: Focus on Clinicians. J Med Internet Res. 2020;22(6):e15154. Published 2020 Jun 19. doi:10.2196/15154
  2. Informatics Summit; 2020 March 23-26.  Poster presentation; at Special Informatics Summit Poster Session at AMIA 2020 Annual Symposium, Chicago, IL, November 14-18,2020.
  3. Ibid.