epocrates

Exploring the evolving role of AI chatbots in medical practice

May 18, 2023

"The future of healthcare" is an oft-used term. As of late, one would be hard-pressed to find a headline that doesn't combine this phrase with another popular term: ChatGPT. With many academic debates centering around the value of AI and AI chatbots in healthcare and clinical practice, the potential of this technology to simplify medical administrative tasks is also being explored. As the application of AI in clinical practice evolves, current research is focused on its benefits and potential drawbacks.

Benefits of AI in healthcare administration

With the recent acceleration of AI technology, software engineers and software-as-a service providers are rapidly developing new chatbot programs and options; therefore, it’s reasonable to conclude that this technology is here to stay. Accordingly, healthcare businesses must remain current and knowledgeable about advancements in AI and determine whether the benefits of using this technology outweigh any potential adverse outcomes. Among the benefits of AI's application in healthcare are real-time clinical decision support, increased face-to-face time with patients, and improvements in clinician burnout.

Clinical decision support

Even with vast foundational knowledge and technical skills, it’s unrealistic to assume that clinicians know the intricacies of each medical condition, especially as they relate to diverse patient populations and practice settings. As a result, AI can be beneficial in clinical decision-making at the point of care. An article in NPJ Digital Medicine describes the application of computerized clinical decision support systems (CDSS). When questioning a diagnosis or the recommended management of a particular condition, clinicians can use AI to quickly analyze individual patient characteristics and search large digital clinical databases to assist in the decision-making process and plan of care development (Sutton et al., 2020).

More time spent providing patient care

When asked about the most cumbersome part of the day-to-day job, clinicians frequently mention charting and writing patient notes in the electronic health record (EHR). Statistically speaking, studies have shown that bedside nurses spend anywhere from 26.2-41% of their time completing documentation (Yen et al., 2018). This result is significant, considering that nurses spend nearly a quarter to half of their time on tasks that take them away from providing direct patient care. By applying AI to automate some of the administrative tasks related to EHR documentation, healthcare professionals could spend more time taking care of patients, thus improving overall patient outcomes.

Improved clinician burnout

A common theme in healthcare, particularly stemming from the COVID-19 pandemic, is clinician burnout. The American Medical Association (2023) reports that 60% of physicians experience emotional exhaustion and other physical and physiological signs of burnout. Additionally, more than half of nurses intend to quit their jobs in the next two years, further contributing to staff shortages and low workplace morale. A study in the Journal of General Internal Medicine explains that healthcare professionals relate most of this to work overload (Rotenstein et al., 2023). Theoretically, by automating tasks such as charting in the EHR, clinicians may have a more realistic daily workload, which can improve the negative feelings of burnout and fatigue.

Drawbacks of AI in healthcare administration

While the benefits of AI make it theoretically appealing, organizations should take specific care before deciding its role and value in healthcare. Most of these concerns are directly related to patient safety and privacy.

Patient safety concerns

Because medical advances often follow technological progress, AI may not accurately reflect the most appropriate clinical information available. Regardless of the robustness of AI technology, fact-checking will be required to ensure the accuracy of the provided information before using it in clinical practice. In most cases, the fact-checking process is straightforward and not too time-consuming, but it is an important consideration to ensure that using AI is still the most efficient use of time and the safest option for patients.

Patient privacy concerns

Because AI is digital software, its use in healthcare significantly increases potential access to private patient information. Patient information obtained from AI is often tracked, stored, and used in large digital databases (U.S. Government Accountability Office, 2020). Though most AI companies have measures to ensure consumer privacy, there’s still a significant risk of data leaks, hacking, or other security breaches. This factor is a significant consideration in the preservation of both patient confidentiality and privacy.

There are many examples of AI chatbots implemented into healthcare practice. One popular software is Ada Health, a digital application that collects consumer information to offer medical advice and strategies for achieving optimal health (Vanguard X, 2023). Another AI chatbot, Florence, acts as a "nurse," working on digital communication platforms like Facebook or Skype to track consumer health data and provide personalized medical information.

Most current examples of AI chatbots used in medical practice focus on the individual consumer and are not necessarily integrated systemically into the larger healthcare organization. However, recent advancements in medicine and technology indicate that the integration of AI on a larger scale in healthcare administration is not only on the horizon but also not far from becoming standard practice in patient care.

References

American Medical Association. (2023). Physician burnout. https://www.ama-assn.org/topics/physician-burnout

Rotenstein, L.S., Brown, R., Sinsky, C., & Linzer M. (2023). The association of work overload with burnout and intent to leave the job across the healthcare workforce during COVID-19. Journal of General Internal Medicine. https://doi.org/10.1007/s11606-023-08153-z

Sutton, R.T., Pincock, D., Baumgart, D.C., Sadowski, D.C, Fedorak, R.N., & Kroeker, K.I. (2020). An overview of clinical decision support systems: Benefits, risks, and strategies for success. NPJ Digital Medicine, 3(17). https://doi.org/10.1038/s41746-020-0221-y

U.S. Government Accountability Office. (2020, November 30). Artificial intelligence in health care: Benefits and challenges of technologies to augment patient care. https://www.gao.gov/products/gao-21-7sp

Vanguard X. (2023). 8 examples of AI chatbots that are changing healthcare experience. https://vanguard-x.com/ai/ai-chatbot-healthcare/

Yen, P.Y., Kellye, M., Lopetegui, M., Saha, A., Loversidge, J., Chipps, E.M., Gallagher-Ford, L., &Buck, J. (2018). Nurses' time allocation and multitasking of nursing activities: A time motion study. AMIA Annual Symposium Proceedings Archive, 1137-1146. https://pubmed.ncbi.nlm.nih.gov/30815156/