The future has arrived, and there’s no denying we are in the age of artificial intelligence (AI). In just the past few years, this once rudimentary technology has seen exponential growth and an explosion of innovative applications across industries. But there is perhaps no more critical application for AI than in healthcare.
Even before the pandemic, the U.S. healthcare system struggled to meet the needs of Americans, especially those in rural and low-income communities — and the inequities have only grown more apparent. But, from expanded telehealth to enhanced diagnostics, AI has already shown its potential to help bridge this divide in healthcare. Through strategic partnerships and investment, our healthcare industry and government leaders can use AI to provide greater access to patient care across the country.
AI In Action
Roughly 80% of the nation’s counties, covering 30 million people, are classified as healthcare deserts.[1] People living in these low-resource areas such as Appalachia, South Texas or tribal communities face long distances and wait times to get care. And once they do, services are often inadequate due to a lack of technology and well-trained staff.
Enter AI. Services like telehealth can bridge the physical barriers of access and streamline care. Outfitted with AI capabilities, virtual health technologies can be used to connect patients to remote experts and aggregate their healthcare data. This data can be used to better inform treatment decisions, helping avoid unnecessary and costly trips to a doctor or hospital. AI is directly facilitating distributed healthcare, moving services closer to the patient in the U.S. and globally.
The potential of AI truly shines in its applications for imaging and diagnostics. In radiology, imaging technologists say that almost a quarter of their work is inefficient and could be automated.[2] AI applications can seamlessly integrate into clinical workflows to automate exam planning, scans, and processing. With routine tasks off their plate, imaging technologists can see more patients and focus on providing high-quality images for more accurate diagnoses in less time.
AI algorithms can be used across disciplines to help interpret chest X-rays, spot cancer in mammograms, identify brain tumors in MR images, and detect arrhythmias in ECGs. They’re also being used by the Department of Defense to assist in military healthcare. Trained with biometric data from over 11,000 patients, the Rapid Assessment of Threat Exposure (RATE) project uses a predictive AI algorithm to detect infectious diseases up to 48 hours before symptoms start, with results even predicting up to six days prior.[3] Integrated with wearables, RATE has the power to change how the military approaches readiness and access to care for soldiers in the field.
In another area of imaging, AI is making strides in maternal healthcare. Exacerbated by barriers to care, today, an American woman is more likely to die from complications of pregnancy and birth than her mother was a generation ago.[4] Amid this health crisis, AI can help democratize ultrasound imaging in care deserts by lowering the barriers associated with training, geography, and technology. In Kenya, a recent trial of AI-enabled portable ultrasound helped connect expectant mothers to critical care.[5] Kenyan midwives were trained on the technology within hours rather than weeks, and mothers gained a sense of confidence in their pregnancies. Applied to care deserts in the U.S., these types of technologies can transform access for mothers in need.
Deployed at scale, AI-enabled portable ultrasound can also extend beyond maternal care: EMS workers could triage patients at the scene or in an ambulance to uncover insights before ever reaching the hospital. Military personnel could aid soldiers in the field, sharing critical clinical information with specialists remotely. This point-of-care focused approach to ultrasound will enable access to life-saving care for people everywhere.
AI & Ethics
But as AI grows across industries, so too does concern for its negative impacts. Algorithms have shown the potential to reflect human biases in applications ranging from job recruitment to health diagnostics, the latter in particular opening up the potential for dangerous outcomes. This is compounded by long histories of distrust for healthcare institutions in many underserved communities. How can industry leaders develop AI-enabled solutions that patients can depend on when there is often little or no diversity in the collected data for training and testing the AI algorithms?
Any application of AI should be seen as supportive; in healthcare, it’s critical that algorithms are not used to replace professionals, but rather to augment their clinical decision-making. And since AI relies on data input, it’s also paramount that personal health data is both gathered from diverse populations and always protected and used responsibly with fairness, oversight, and transparency.
By establishing guiding principles of AI use such as these, the industry can adhere to a shared approach. Industry leaders from Microsoft, Philips and more are developing the Artificial Intelligence Code of Conduct (AICC) with the National Academy of Medicine. The AICC project, planned to be completed by December 2025, is shaping a national architecture for responsible healthcare AI.[6] In this way, leaders can help ensure AI uses data the right way and doesn’t further exacerbate health inequities.
The Future of AI in Healthcare
With AI, industry leaders can pave the way to distributed healthcare and revolutionize patient experience. Barriers, both physical and institutional, can be toppled through applications in telehealth, diagnostics, and treatment. But without proper safeguards and implementation, there is potential for misuse.
And without the combined efforts of industry, academia, non-profits, and government, the reach of these applications will remain limited. Public-private partnerships are critical. Recent funding for AI diagnostic tool programs from organizations like the Bill & Melinda Gates Foundation[7] and the Department of Defense[8] is promising, but the issue of access necessitates we extend the reach of these technologies as widely as possible. Further uptake in hospitals across the U.S., for example, can help AI close the gaps left open by healthcare deserts.
In the age of AI, it’s up to us all to carefully consider the implications of algorithms for the health of our world – and for their potential to transform lives.
To Learn More About Philips: https://www.usa.philips.com/healthcare/government
[5] United Nations Population Fund
[6] National Academy of Medicine
By Andrew Omidvar, MBA, Vice President Government R&D, Philips