Advances in Artificial Intelligence Help Evolve and Support Care Delivery

As organizations across the country continue to develop and refine their virtual health programs to ensure a continued return on investments made in the spring, emerging technologies continue to advance and gain traction. The emerging technology of artificial intelligence (AI), in particular, has been delivering value for health systems.

There have been several well-publicized reports on AI applications that have contributed to the delivery of high-quality care; however, the ways in which AI can reshape the health care landscape are often misconstrued.

Frequently, AI is touted as a replacement for physicians and a technology that will disrupt care delivery overall. Consistently though, organizations on the bleeding edge of AI exploration are finding that the most impressive outcomes are not driven by AI or physicians alone. Instead, those pushing the boundaries of AI use in health care have demonstrated that it is the combination of AI and clinicians that delivers significant results.

While not new, voice assistant technology has been the AI solution that has made one of the biggest impacts in recent years. Organizations who have adopted and deployed this technology have shown its value to be impressive; however, widespread adoption has been slow. With a new development, though, one organization may be moving this technology from innovative to expected.

Epic, in partnership with Nuance, has developed and deployed the Hey Epic! digital assistant that allows physicians to interact with their EMRs using only voice commands to accomplish specific tasks. Epic has been touting their work with Nuance for quite a while—this capability has been rolled out for use on Epic’s mobile apps. But the significant change is that this capability can now be incorporated into their desktop user interface.

Because the desktop interface is the most widely utilized version of Epic, this availability will likely make this capability an integral asset in the support of care delivery. With this solution, clinicians can conversationally navigate the EMR and search for information such as lab results, patient records and visit summaries. Additionally, clinicians will be able to place orders and switch to voice-enabled, hands-free charting.

This move will likely lead to more efficient visits and a continued advancement toward technology supporting clinicians in delivering high-quality care in a patient-focused interaction. In addition to efficiency, though, AI is being developed to enhance clinician accuracy as well.

In collaboration with the US Department of Defense (DoD), Google is developing a prototypical artificial intelligence digital pathology system. The innovative partnership is designed to support physician accuracy of cancer diagnoses by providing DoD medical centers with augmented reality microscopes overlaid with AI technology. While this technology is initially only being piloted at select medical facilities, the DoD has future ambitions to scale the platform across the broader US Military Health System. This solution will first be aimed at assisting researchers focused on prostate cancer, breast cancer, colon cancer, cervical dysplasia and lymph node metastasis. However, this type of capability may also enhance research on COVID-19 treatments.

As organizations continue to explore AI and consider new applications for this impactful technology, it is important to recognize that AI is not magic. As is the case with so many other technologies, it can be utilized effectively when designed in ways that augment or enhance existing routines and capabilities. However, when not structured appropriately, these technologies can simply add cost without delivering meaningful benefits. They can do harm as well, as AI can be impacted by implicit biases in the data they are utilizing or in their design. One widely discussed example is the fact that skin condition diagnostic apps may be more accurate for lighter skin based on the lack of diversity in the data used to build the system.

As you invest in your digital health program, consider the following recommendations for long-term success:

  • Seek opportunities to pilot AI, prioritizing areas that are aligned with your enterprise-level strategy and have provided impact across the country.
  • Collaborate with early-stage AI vendors to codevelop capabilities that are specifically designed to meet the challenges your organization is facing and that take into account the workflows and clinical protocols of your organization.
  • Explore modifying intake and selection processes to ensure that AI vendors or solutions offered have not inappropriately flagged any specific patient populations in previous studies or pilots. It is as challenging as it is important to understand the difference between recognition that there are clinical variances across patient populations, as opposed to biases built into the underlying algorithms.

Related resources

  • Share
  • Follow Sg2 on Twitter
  • Connect with Sg2 on LinkedIn

Tags: , , ,