AI Is Transforming Medicine for the Better

Doctors could be forgiven for wondering if robots are taking over their jobs with all we hear of the reports and studies about machine learning and computers doing the jobs of nurses and physicians.

They don’t need to worry. Artificial intelligence (AI) may be new and spreading at a dizzying speed, but in the end will make the lives of both providers and their patients better.

Here’s how.

AI Is Fostering Widespread Benefits in Health Care

AI is already proving its mettle in several key disciplines. The examples below are just a smattering of how AI is transforming and improving health care for both patients and providers.

General Medicine: In the United Kingdom, more than 1.2 million people are using a chatbot to answer their medical questions on demand and at home. The National Health Service (NHS) expects the app to lower its costs; a typical bot-human interaction takes only 12 back-and-forth messages and is faster than speaking with a health care professional on the phone.

Cardiology: A 2019 article on an AI-driven computer that “studied” more than 90,000 EKG recordings said that the machine was able to learn to recognize patterns, make rules, and apply them accurately to future EKG readings. The machine quickly was able to recognize and classify 10 types of irregular heart rhythms. After seven months, the computer algorithm was as good, and in some cases better than, experts in making an accurate diagnosis.

A 2019 meta-analysis of the use of AI machine learning in cardiology concluded: “All the studies mentioned in this review have clearly demonstrated superior results of ML in relation to conventional approaches for identifying obstructions or predicting major adverse events in reference to conventional approaches.”

Cancer: A retrospective study of missed cancer diagnoses showed that cmAssist, an artificial intelligence-based computer-aided detection algorithm, can be used to improve radiologists’ sensitivity in breast cancer screening and detection. “There was less than 1 percent increase in the [radiologist] readers’ false-positive recalls with use of cmAssist,” the study authors wrote. “With the use of cmAssist TM, there was a substantial and statistically significant improvement in radiologists’ accuracy and sensitivity for detection of cancers that were originally missed.”

Psychology: A 2019 study assessed the utility of Tess, an AI tool for improving mental health services. Tess delivers on-demand support for caregiving professionals, patients, and family caregivers. The authors describe Tess as a “low-cost, user friendly, and highly customizable” service, noting that it provided notable improvements in both patients with mental health issues and their caregivers.

“There is evidence that using psychological artificial intelligence to provide customized support for caregiving professionals, patients, and family caregiver is a feasible service delivery method,” the authors noted. “This report suggests that the Tess service may offer an affordable and scalable solution that accommodates the busy schedules of caregivers while helping them reduce burnout and improve resilience.”

The researchers also noted that Tess’ capacity to expand support to patients reduced caregiver burden, and showed potential for relieving feelings of depression, anxiety, and loneliness. The use of Tess also allowed “emotional support to be scaled to thousands of people at a single time.”

Ophthalmology: A recent article noted that several studies have shown that AI is accurate and effective in detecting diabetic retinopathy, glaucoma, age-related macular degeneration (AMD), retinopathy of prematurity, refractive error and in identifying cardiovascular risk factors and diseases, from digital fundus photographs. In addition, the authors noted, AI is reliable for, “Identifying disease features, progression and treatment response for retinal diseases such as neovascular AMD and diabetic macular edema using optical coherence tomography (OCT). Additionally, the application of machine learning to visual fields may be useful in detecting glaucoma progression.”

Dermatology: In a 2019 study, researchers used 12,378 open-source dermascopic images to train an AI network to perform dermatologist-level classification of suspicious lesions. The network’s accuracy of diagnosis was compared to that of 157 dermatologists from 12 university hospitals in Germany. The AI-driven application “exclusively outperformed 136 of the 157 dermatologists and all the different levels of experience (from junior to chief physicians) in terms of average specificity and sensitivity,” the authors concluded.

Pediatrics: Research has recently been compiled on Google’s wearable glasses, called Superpower Glass, an artificial intelligence-driven, wearable behavioral intervention for improving social outcomes of children with autism spectrum disorder (ASD).

The device aims to help children read and interpret people’s facial expressions and improve socialization of children with ASD. In a randomized clinical trial, participants received one of two types of therapy. The first group received the Superpower Glass intervention and an accompanying smartphone app to promote facial engagement and emotion recognition by detecting facial expressions and providing reinforcing social cues. They also received behavioral analysis therapy. The control group received only applied behavioral analysis therapy.

The study showed that “Children receiving (Superpower Glass) showed significant improvements on the Vineland Adaptive Behaviors Scale socialization subscale compared with treatment as usual controls.” In addition, Superpower Glass “reinforces facial engagement and emotion recognition, suggesting either or both could be a mechanism of action driving the observed improvement,” the authors wrote. “This study underscores the potential of digital home therapy to augment the standard of care.”

As Sunnie Southern, health and life sciences lead at Onix, understands the concerns that physicians may have about widespread use of AI in the workplace. “There’s been so much hype about machines taking over in health care, she said, “And we cannot expect doctors to just automatically use whatever we throw at them.”

Southern posited that AI explainabilty – the extent to which humans can understand why AI came to the conclusion it did – is a core feature of helping providers to put into play the multiple diagnostic and treatment benefits that AI offers.

She added that machines cannot, and will not, replace humans because in health care “nothing can replace a warm handshake, a hug, a look – all of those subtle but vital human interactions that are a vital part of provider caregiving.”

AI Helps the Growing World Population to Identify Disease Threats

According to the United Nations’ Department of Economic and Social Affairs, the world’s population will reach 9.8 billion by 2050, and jump to 11.2 billion by 2100. This stunning projected population growth will combine with pernicious leisure travel; mass migration triggers, such as geopolitics; weather-associated global events, such as flooding and droughts; and opportunity-seeking immigration; and other global factors to present challenges to dealing with emerging disorders and controlling pandemics of known diseases.

AI has the ability – and is already being used – to automate surveillance systems and amass and sort health patterns and anomalies. Reporting from clinicians, laboratories, and public health agencies informs health policy and the response to unanticipated and potentially dangerous health threats.

AI can also be used for contact tracing during outbreaks, which allows health officials to identify people who come into contact with infected patients.

The pervasiveness of cell phones, laptops, tablets and other electronic devices also allows AI to issue alerts and recommendations to broad swaths of local and international sectors of the public. Several AI-driven systems are already in play.

The Bioportal project, initiated by the University of Arizona along with the New York State Department of Health, and the California Department of Health Services, uses real-time hospital data, free text hospital information, and social network data to affect hotspot analyses. The data produced provides, in real time, interactive geographic patterns of outbreaks and can be used to quickly sift through data to inform responses.

Researchers report that during the 2012 to 2013 influenza season, an AI surveillance system monitored Twitter-based feeds for mentions of the flu or its symptoms . The AI system mirrored CDC data on influenza with 85 percent accuracy.

Canada’s Global Public Health Intelligence Network (GPHIN) is a digital, automated surveillance system that uses AI to monitor Internet media, such as news wires and websites in nine languages to detect and report potential disease outbreaks or other health threats around the world. In 2004, the GPHIN detected the SARS outbreak two months before the World Health Organization did so.

The International Society for Infection Diseases’ ProMED Mail is an Internet-based reporting system dedicated to rapid global dissemination of information on outbreaks of infectious diseases and acute exposures to toxins that affect human health, including those in animals and plants grown for food or animal feed. ProMED uses AI to provide 24/7 up-to-date and reliable news about threats to human, animal, and plant health worldwide.

According to the service’s website, “By providing early warning of outbreaks of emerging and re-emerging diseases, public health precautions at all levels can be taken in a timely manner to prevent epidemic transmission and to save lives.”

Other massive surveillance and warning systems are in place, and many more are under development. There is no substitute for AI’s ability to collect, sort, and interpret public health data and issue appropriate warnings, recommendations and updates, in real time. The increasing use of such technology will undoubtedly save thousands and even millions of lives – and is probably already doing so.

AI Can Create a Health Care Democracy

The United States is infamous for denying millions of people access to health insurance and, effectively, health care. The financial costs in lost productivity and treating very sick patients first seen for their ailments in the emergency room — when their untreated medical issues have led to severe symptoms and outcomes — are the subject of much study and political debate.

But location can also limit access to health care. U.S. rural communities share common risks for poorer health. These challenges, including a dearth of local doctors, poverty, and geographical isolation, contribute to a lack of access to care, according to the Association of American Medical Colleges.

A 2017 report by the North Carolina Rural Health Research Program (NC RHRP) at The University of North Carolina, Chapel Hill, detailed further challenges faced by rural dwellers as compared with urban dwellers:

• A higher percentage of children living in poverty
• Fewer adults with a postsecondary education
• More uninsured residents under age 65
• Higher mortality rates of mortality
• Inability/unwillingness to take time from work for the long trek to health care facilities
• Longer (potentially dangerous) wait times for emergency medical services
• Higher rates of obesity, diabetes, mental health disorders and other chronic conditions – and early deaths

AI Can Provide More Access to Health Care for Rural Patients

Telemedicine, the remote diagnosis and treatment of patients via telecommunications technology, has already made a dent in the U.S. rural health crisis, according to a recent article in the Harvard Business Review. AI already works with some of these systems to monitor remote patients’ vital signs, trigger warning and appointment reminders, and allow patients to speak face-to-face with health providers via cell phone.

AI-driven devices such as smartphones, tablets and smartwatches similarly could keep patients consistently connected to providers, remotely update charts and outcomes, provide reminders and improve adherence to medication and treatment plans. Eventually, the disparity gaps could be narrowed, even closed entirely, creating a more equitable playing field for U.S. populations, regardless of geography.

AI-driven medicine can, whether on a local or global scale, gather and update community and regional health data in real time. This data, in the gears of machine-learning AI robots and computers, can incorporate the data with extant data and subsequently create, in a matter of seconds or minutes, evidence-based community-health interventions and disease eradication recommendations, all the while learning from the freshly added information.

AI also holds the promise to equalize medical treatment in other ways. Through millions and billions of dollars saved on lugubrious administrative tasks, misdiagnoses, wrongful deaths, disease spread, inadequate care, and unnecessary emergency rooms costs, the opportunity for universal health care in the U.S. could become a reality, and what an equalizer that might prove to be! The nation known for health care excellence and innovation would no longer have the most expensive system, with the sickest patients, in the western hemisphere.

C’mon, the Water’s Fine

With technological advances in medicine moving at the speed of sound, it’s no wonder that providers feel overwhelmed by new robotic and computational AI tech. But just as we learned to get used to and use seemingly alien and complicated devices such as mobile phones and PC’s a relative few years ago, AI in health care is rapidly becoming the norm.

Machine learning will continue to improve and extend patients’ lives in ways that are as unimaginable as AI-driven robots and computers were just a few years ago. Providers will have more time to dedicate to their patients and their research.