Combining Wearables and AI Can Bring Some Big Benefits to Healthcare

July 3, 2019 - 7 minutes read

Wearables are a low-risk method to collect high-quality patient data around the clock. When you add artificial intelligence (AI) into the mix, you open up opportunities for smarter analysis, more actionable insights, and better results.

In England, patients leaving a group of hospitals in the southeast are receiving their own wireless, connected armbands. These wearables watch vital signs like oxygen levels, pulse, respiratory rate, body temperature, and blood pressure. This new discharge protocol is being carried out under a new National Health Service pilot program.

The new program uses AI to analyze the patient’s data in real-time speed, and it’s already shown to decrease hospital readmission rates and emergency room visits. Home visits, which can be expensive, have reduced by 22%, and patients are following their treatment plans more closely at home.

To AI or Not to AI?

Harvard Business School professor and Innosight co-founder Clay Christensen says this AI MedTech program is targeting “non-consumption”. This is a business opportunity where consumers are required to take action but they don’t currently have the tools or technology to do so.

For example, prior to the armbands, hospital employees had to drive up to 1.5 hours round-trip to check on patients in-person once per week. With AI algorithms keeping an eye on worrisome data trends, however, patients and hospital employees get notified well in advance of any major health complication.

Many experts believe that AI will introduce very low-error predictions at a much cheaper cost than using humans. These successful case studies always lead to more AI implementations in healthcare, but sometimes AI isn’t appropriate for every sector of healthcare. And because AI MedTech applications can be expensive to design and develop, healthcare systems that focus on the bottom line aren’t well-suited for this type of technology.

AI in medical imaging tools is very expensive; hospitals are forecasted to spend over $2 billion per year by 2023. Currently, hospitals employ specialists who are trained to find diseases ranging from cancer to cataracts. But because this profession requires holistic knowledge of the body and medicine, AI isn’t the greatest asset for these specialists. Its ability to be useful to patients is low, and associated costs won’t decrease for this AI application either.

Decentralized Care

For clinics that want to decentralize care, AI is a perfect tool, however. For example, patients are faced with a multitude of choices throughout the day that directly affect their health: the choice to exercise, eat moderately and varied, meditate, and more. Usually, there isn’t a doctor around telling us what’s best for our health; we must make the choice for ourselves. But these choices add up over time to increase the cost of healthcare for everyone.

The World Health Organization estimates that 60% of related factors to health and quality of life correlate directly to lifestyle choices, like taking prescriptions, reducing stress, and exercising. With AI, sending the patient a reminder to exercise or eat a veggie meal doesn’t take any extra time or effort from the doctor. And if something looks off, the AI can decide whether to send a reminder to meditate or set up a doctor’s appointment.

Several universities and health insurance companies are working to increase patient monitoring from home. These programs have produced positive results, but it’ll take some time before the AI fully meets researchers’ expectations.

Experts look to a future where the U.K.’s NHS program gets implemented in locally first before open sourcing it for global use. Already, the connected devices were approved by the FDA for use in New York City’s Mount Sinai Hospital. In the U.S., patient readmissions cost hospitals more than $40 billion every year.

These programs have yielded three lessons on how to use AI to address non-consumption in patient-centric healthcare.

(1) Target an impact on critical metrics, like readmission rates and insurance payouts.

And then begin to tackle the metric with realistic goals instead of lofty ideas. AI can be iteratively optimized for continuous improvement if it can see an evolving trend in the data. Grady Hospital in Atlanta has saved $4 million through a reduction in readmission rates by using AI to find at-risk patients.

(2) Increase collaboration to reduce risks.

Finding others who are working to crack the same problem as you can mean the difference between three years or a decade in launch time. Ascension, Aetna, Optum, Humana, and other insurance companies are collaborating in using blockchain to increase their data pool with customer data. By reducing time to consolidate out-of-network claims, these insurance companies are also improving patient access to care providers.

Sometimes, doing the job without help can pose major risks. For example, the M.D. Anderson Cancer Center in Houston had a multi-million dollar AI project fail because it was incompatible with its own electronic health records system.

(3) Partner with highly-specialized professionals instead of competing against them.

Several modern-day AI applications compete with doctors; for example, radiology-focused AI algorithms often perform as well as or better than human radiologists in image-based diagnosing. But we shouldn’t let AI usurp the specialist because AI needs human eyes to check over its recommended decisions.

A Healthier Future

AI in MedTech and healthcare is changing and improving lives all around the world. And while AI isn’t the right tool for every medical application, it can yield profound results in situations where it’s needed. Better health and healthcare is undoubtedly a win-win for every stakeholder involved.

What AI applications are you most excited to see in healthcare in the near future? Let us know in the comments!

Tags: , , , , , , , , , , , , , , , , ,