The healthcare AI stream at the AI Summit in London during June provided a fascinating look at how AI is being used to solve real healthcare problems. Consider some of the meta challenges of healthcare: overwhelming amounts of poorly connected data; the unpredictable nature of sickness and wellness; longer life expectancy; and costs. Now consider some of the strengths of AI: disparate data analysis and prediction and the automation of mundane, redundant tasks. There is perhaps no other field where AI could have such a significant positive impact on outcomes.
Two cases discussed at the AI Summit were particularly enlightening: Cleveland Clinic’s work with AI-driven prediction and Hive/Carers UK’s launch of evolved senior caregiving.
Cleveland Clinic Attacks Burnout
Not-for-profit Cleveland Clinic (CC) handles 5.4 million patient visits annually. Dr. Brian Donley, CEO of Cleveland Clinic London, described how the organization arrived at AI as a potential solution to one of the industry’s biggest challenges: worker burnout. Donley said 54% of healthcare workers suffer from burnout while the average across all other industries is 23%. Doctors in particular are under even more stress today. The dual-edged sword of technology can help in so many ways, but it also increases the amount of information doctors must produce (clinical documentation) and ingest (43,000 medical journals, 810,000 professional articles, patient-entered data, trials, etc.). So CC is looking to become more efficient and is using AI for patient-focused analysis and prediction.
Donley said CC has harnessed nearly 1 billion points of data, which it is using for predictive readmission and length of stay and for predictive personalized survival modes. The organization found that a significant number of patients are readmitted after being released from the hospital. Analyzing clinical, demographic, and comorbidities data, CC was able to understand patterns and build predictive models for treatment and length of stay for specific illnesses. The results have included a significant reduction in readmissions and length of stay. Not only does this mean a better outcome for the patient, it also enables CC to staff smarter. Given the profiles of current patients, the organization can predict the type of specialist and generalist staff needed and for how long.
Hive and Carers UK Enable Smarter Senior Care and Independent Living
Aimee Clark, Strategic Partnerships manager at Hive, and Madeleine Starr, director of Business Development and Innovation at Carers UK, gave a presentation called “The Changing Face of Care.” Nonprofit Carers UK shared data from its State of Caring Report 2019, which painted an interesting picture of an evolving trend: the pressure on caregivers for seniors and children. Carers UK estimates there are 6.5 million caregivers in the U.K. – 3 million combine care with work and 1.5 million care for a senior and a child (known as “sandwich carers”); 72% have mental health issues as a result. In addition, the survey found that 50% of those over age 65 don’t talk about the support they need, and 50% of those over 40-60 years old haven’t talked to their parent about how they can help.
A Big Issue
(Source: Carers UK)
Working with Carers UK and multiple partners, workplace productivity company Hive ran trials around a monitoring system designed to help seniors live independently. The solution includes a mobile app for the caregiver that remotely monitors activity in the senior’s home. Sensors in the home can be set to report activity for various things such as the TV, kettle, toaster, cupboard, door, living room, kitchen, etc. Reports can show 24-hour and weekly patterns of activity. As the system learns patterns, caregivers can set up a unique home profile for when their loved one is typically sleeping, their wake-up window, when they have consistent activity every day, and specific appliances used in this time period. Caregivers can also set up custom notifications, such as no activity. An alert for such a notification might say “No activity detected since yesterday at 9:03 p.m.”
Monitoring to Learn
The results indicate that the stress of caregiving can be reduced via this system, as caregivers can constantly monitor their charges’ situations without physically being present. Clark shared a few quotes that were illustrative of caregiver sentiment:
- “She has a physio for her hip. I got a notification to say she hadn’t gotten up yet and the physio was coming in half an hour. I think her sugar levels were off, she wasn’t feeling that great.”
- “I got a notification to say the door had been left open, which seemed unusual. I gave mum a call and she told me she’d been in the garden that morning. Before we knew it, we were talking about her hydrangeas.”
Clark admitted the challenge for such a solution includes gaining the permission of seniors to place monitors and placating resistance to a sense of 24/7 monitoring.
It is encouraging that two such practical use cases for healthcare AI were so prominent at the 1-day conference. This is an indication that companies are thinking about practical challenges and leveraging AI to solve them, rather than trying to become AI experts and then finding a problem to solve with the technology.