Artificial intelligence has already been working behind the scenes in healthcare to assist physicians in providing patient care. For example, AI can help interpret MRI scans and detect possible cancer in imaging tests. Additionally, this month Amazon Web Services unveiled Transcribe Medical, its HIPAA-compliant speech recognition application program interface that developers can integrate into applications to enable physicians to do clinical documentation efficiently.
Modern Healthcare recently reported that AI-powered technology is now advancing from being behind the scenes to interface directly with patients to collect initial information about their symptoms and vitals, a process normally performed by clinicians during a patient encounter.
For example, research professor Albert Rizzo and his team from the University of Southern California have been developing a virtual reality character named “Ellie” that uses AI to lead a person through initial questions. Ellie was originally designed to determine whether veterans returning from a deployment might need therapy. Appearing on a computer monitor, Ellie can make eye contact, nod, use hand gestures and even pause to encourage a person to elaborate more on an answer.
Although consumers are increasingly accustomed to using technology for travel and shopping, there continue to be limits to how much technology can be used in the patient care continuum. Because AI has not yet mastered important care elements such as the ability to feel compassion or use common sense, Ellie is not intended to replace the therapists that are still needed to diagnose and provide treatment. Instead, the AI is used to better predict who is most likely to be suffering.
Another example of AI-powered technology that interfaces directly with patients are the self-service kiosks found in the walk-in Akos Med Clinics in Safeway stores in Phoenix. The kiosks guide patients through a series of questions to identify medical history, current medications, allergies and symptoms. Patients also conduct self-examinations using equipment such as a stethoscope, temporal thermometer and blood pressure machine. Images of a patient’s throat, ears and nose can be captured by a camera. The clinics have a medical assistant on stand-by, but there are no on-site physicians or nurses. Instead, a provider comes on-screen to speak with the patient as a live virtual exam room experience.
Using AI, patient information collected by the kiosk generates a set of probable diagnoses, but the on-screen provider ultimately makes the diagnostic decision. The patient process typically takes less than 25 minutes and costs $75 per visit for those without insurance. Patients can then walk next door to the pharmacy to fill medications prescribed. James Bates, CEO of AdviNOW Medical, which provides the underlying AI software to Akos Med Clinics, expects the program to expand into 1,000 clinics by the end of 2020. The company also envisions using AI to diagnose and treat some minor illnesses in the future.
In general, US consumers continue to adopt telemedicine and digital health technology. A recent survey of 4,000 US adults by Rock Health and the Stanford Medicine Center for Digital Health found that in 2019, one in four respondents reported using live video telemedicine over the past 12 months. Additionally, 44% of respondents reported tracking a health indicator or activity using digital tools in 2019. Of the respondents who reported using digital health tracking tools in 2019, 75% reported sharing data with a medical professional. Common health metrics tracked included heart rate, blood pressure, blood sugar, weight, and food or diet.
Another example of further telemedicine adoption is the significant growth in usage reported by the US Department of Veterans Affairs (VA). In the fiscal year 2019, usage of VA telemedicine services by veterans increased 17% due to the May 2018 enactment of the VA’s Anywhere to Anywhere initiative, which lifted state licensing requirements for VA providers practicing virtual care. State licensing restrictions have been a barrier for national and regional health systems to implement telemedicine programs in the private sector, and the VA’s policy is a first step toward demonstrating the efficacy of a national medical practice licensing concept.
Over the next decade ending 2028, the US Bureau of Labor Statistics projects physician employment will grow 7% to 812,200, while registered nurse employment will grow 12% to 3.4 million. With the low unemployment rates of 0.4% for physicians and 1.1% for registered nurses during the third quarter of 2019, the healthcare industry will likely continue to adopt telemedicine, AI and digital health technology to ensure adequate clinician resources exist to cover the volume and complexity of patient care needs across communities.
