New technology can have a significant impact on the day-to-day experience of providers and patients.
The American Medical Informatics Association is hosting its 2021 Annual Symposium, and, during its virtual day last week, researchers presented multiple studies on how technology is impacting both groups.
Here's a look at some of the findings:
Using VR to understand the challenges of home care
To study and better understand behaviors that affect health outcomes, the National Institute of Nursing Research (NINR) developed an immersive virtual reality tool to track participants as they engage in a common daily activity like food shopping. The tool simulates a typical grocery store and tests participants’ ability to follow low-sodium diets in the context of chronic illnesses like heart failure.
The benefit of using this technology, Denise Goldsmith, a nursing informatics consultant at NINR, explained during her panel, is the level of factors in the environment researchers could control. In fact, it would be difficult—if not impossible—to replicate in real life, she said.
Another appealing element is the fact that the study could be done remotely without being in the patient’s presence.
Participants were trained for a brief period beforehand on how to use the technology. Then, in the simulation, they were free to move about the grocery store, select food items, read nutrition labels and then check out at the cash register. At the end, each user received feedback on their food choices based on the total sodium tally.
“Using IVR in this way, we can observe real-time decision-making and activities of our users without needing to be present in their natural setting,” Goldsmith, who develops the clinical scenarios chosen for these studies at NINR, said.
The technology captures behaviors including frequency of label referencing and product comparisons and also tracks and plots movement activity throughout the virtual store. This helps researchers study cognitive patterns in home-care patients and potential challenges impacting the performance of such tasks.
The various self-care management scenarios NINR develops are identified by home-care nurses, according to its website.
Medical conditions dictate patients' tech savviness
To explore whether certain medical conditions lend themselves to better telemedicine preparedness among seniors, Jorge Rodriguez-Fernandez, a clinical informatics fellow at the University of Illinois at Chicago, analyzed data from the National Health and Aging Trends Study on nearly 4,000 seniors 65 years and older.
He and his team defined telemedicine readiness as being able to contact a provider, handle medical insurance or get medical information all online.
There were two other categories, one for telemedicine physical unreadiness, and another for telemedicine technical unreadiness. The former was defined by a physical barrier preventing a patient from using telemedicine; the latter was defined by a patient not owning a proper device, not knowing how to use it or not having recent email, text or internet access.
Cancer patients by far exhibited the most telemedicine readiness, followed by those with hypertension and arthritis.
“For cancer, which was the most striking result, there is the theory of the cancer continuum,” Rodriguez-Fernandez explained during the panel. An oncologist treats not only their remission but also focuses on follow-up care and their long-term well-being. Drug manufacturers also increasingly support and engage with patients across the oncology spectrum through their entire care and recovery journey, he noted.
In terms of technical unreadiness, conditions that were most impacted included depression, anxiety, stroke and diabetes. In contrast to cancer patients, people living with depression or anxiety may lack good support, Rodriguez-Fernandez said, may be isolated and may have difficulty trusting providers online.
It’s critical for providers to keep in mind that patients with certain conditions are harder to follow, harder to treat and harder to engage with, he said, and therefore may be less responsive to telemedicine than others, he said.
In terms of physical unreadiness, nearly every condition studied except for cancer was found to be significantly impacted. One helpful solution is making technology more accessible to patients with physical barriers, Rodriguez-Fernandez said, like improving touch features on devices for patients with diabetic neuropathy.
“We have to engage into a new design of telemedicine,” he explained, “that follows the senior technology acceptance model. The idea is that we have to design our apps better.”
Transcribing visit notes may save physicians energy
In a presentation on how patient visit notes can be automatically generated using technology, Kevin Lybarger, a postdoctoral fellow and engineer at the Department of Biomedical and Health Informatics at the University of Washington, explained the benefits and drawbacks of the developing approach.
“Creating clinical notes is known to be time-consuming, and in fact, this documentation burden negatively impacts the job satisfaction of clinicians,” he said. Notes may also at times omit critical information, he added.
With the aid of what’s known as a digital scribe, a visit dialogue can be recorded, transcribed using speech recognition and then produced in a note that can be edited and finalized by the clinician before being put into the electronic health record.
In his past work, Lybarger and his team had set up information extraction modules to draw out information on social determinants of health and symptom data from clinical texts. They then applied those extractors to the dialogue transcripts and evaluated their performance in a conversational environment.
What they found is that the visit transcripts were “rich” with symptom data, with roughly a dozen symptoms per note. However, the extractors still outperformed on clinical notes, for which they were originally designed.
One problem is that extractors have difficulty recognizing colloquial language used by patients during visits. Ultimately, while digital scribes have the potential to ease the documentation burden for clinicians and may benefit from these types of information extraction modules, “new extraction architectures are likely needed to extract medical concepts from the dialogues with high performance,” Lybarger said.