AI usecases in Healthcare
(Second in a Series)
Discussed much more thoroughly in the last article AI in Banking, Artificial Intelligence (AI) is a powerful force for business. Does it have a place in Healthcare, too? You better believe it.
In this country, healthcare is a business, even if it is full of altruistic individuals that are just seeking to help others. We thwart disease; we repair damage; we cope with aberrations in bell-curve physiology; and most importantly, we make lives better.
But that doesn’t work very well without a solid business foundation! We still need to know where the money is coming from, where it is going, how it is invested, and what constitutes a good purchase decision for everything from a scanning electron microscope, right down to the price of cotton swabs.
Contact us if you need help implementing AI
For example, an additional MRI scanner would undoubtedly increase the daily throughput for our clients, but would it be better to hire a couple more technicians to keep the current machines running 24 hours per day? If a patient has a choice of waiting three weeks for an MRI or coming in today at 3 AM, which will they choose? People with day-to-day jobs might not have the time or be able to afford time-off during “office hours” for an appointment.
Easing the Workload
AI is the perfect tool for analyzing data, and although that problem may seem simple, there are dozens more that could be handled by Artificial Intelligence. An AI could do all the paperwork for a family physician. Insurance claims are mind-numbing in this country and take about half of a doctor’s time. If s/he sees patients for six hours, there is going to be six hours of paperwork to accompany that.
The Royal Flying Doctor Service (RFDS) was formed in Australia in 1928, by Rev. John Flynn by combining two fledgling technologies: the ability to fly with airplanes, and the ability to communicate with radios. Rev. Flynn had established the Australian Inland Mission (AIM) in 1912 to provide remote communities with medical aid.
The Outback of Australia is so vast that the AIM wasn’t up to the task. Too many people were too far away even from the hospitals dotted throughout the continent. Putting doctors in planes and equipping them with radios meant that medical assistance was usually only a couple of hours away. The RFDS flew 20,000 miles in their first year, over the course of 50 flights.
AI to the Rescue
The advent of Telehealth has made it possible to deliver healthcare to small, remote communities. This is already common in isolated communities in Canada and Australia. Doctors, Nurse Practitioners, Nurses, mental health professionals, and many others can deliver their services by video link.
In a clinical environment, the remote medic can read real-time ECGs, listen to heartbeats and lung function, and have the patient or an aide perform physical manipulations, as well as just listening to the patient. It’s not perfect, but so much better than waiting months for treatments when going to a bigger city, or waiting for when a medical professional visits.
AI, on the other hand, could fill a role with an adaptation of our modern robot surgeons that are already in use. A remote doctor could manage the robot (complete with their friendly smiling human face visible to the patient), with its highly sensitive and well-developed feedback system, to perform tasks as if s/he were actually in the room. The patient, knowing that a human was at the other end of the machine, could become accustomed to this activity.
The AI would eventually become so knowledgeable that for everyday consulting it could manage entirely on its own. At the opposite end of the spectrum, robotic surgeons would continue to gather empirical knowledge from all the best surgeons in the world. While still monitored by a human surgeon for safety, this would one day allow flawless surgery of any kind. Everything in between these two extremes would be the domain of medical professionals, freeing up more time to interact with people.
AI Enters the Patient Care Arena
In some places, notably Japan where they love their robots, big white (easy to sanitize) friendly robots now pick up patients to place them on gurneys or in bathtubs eliminating injuries for orderlies and nurses. The unit must be smart enough to calculate load and balance, but also capable of detecting pressure, and supporting (often elderly) patients without injury. Of course, they’re still guided by humans
The friendly Panda robot in the pictures deliberately doesn’t look human because that might frighten Alzheimer’s patients. But it can lift someone from the floor and place them in bed, help them to stand, help them to walk, seat them in a wheelchair, and other strenuous activities that result in hospital worker injuries. More importantly, like here, the population in Japan is aging and will need support that might be beyond the capability of the human staff, so they are aggressively looking for solutions from Artificial Intelligence.
Surgical schedules that make the best use of available space and equipment would streamline operations, allowing more to get done, in less time, at a lower cost. The same can be done in Radiology, Physiotherapy, and even the Laundry Department. There is no such thing (yet) as a perfectly organized hospital—there’s room to improve, and AI will get us closer.
The onslaught of cyber assaults has not diminished. Clinics and hospitals are among the most vulnerable because they have so many out-of-date computer programs still in use that are no longer supported by the manufacturer. Updating web browsers is free in most cases. Some hospitals have refused to say “Our employees aren’t allowed to use the computers on the internet, so we don’t need that,” and we all know how employees always obey the rules.
Medical professionals leave tablets and laptops lying around all over the hospital, and they are regularly stolen. This puts Personally Identifiable Information (PII) of patients into the hands of criminals.
AI can identify any of these devices, where it is being used, and if the typing, voice, or data entry technique matches the regular user. Even if it locks after 90 seconds, people use such simple passwords that they can be defeated in just a few minutes. AI can stop the misuse.
An AI, capable of thinking much faster than a human being, could recognize a cyber attack in milliseconds and isolate the affected system. This leaves the remainder of the Computer System up and running, without even a noticeable blip to all of the other users.
Better Analysis than Humans
MRIs, X-rays, PET, CT, and CAT scans are more efficiently and accurately interpreted by computers and AI than humans. The machine can spot tiny variances that a human might overlook, as well as patterns that are located too far apart on an image for a human to make an association.
Having an app on your phone which possesses an encyclopedic medical knowledge database will make it possible to have a consultation on the spot. The difference is that this “ChatBot” is a Medical Expert System, (such as the British Babylon, which charges £50/$66 per year for unlimited consultations) armed with your medical history, and genuine knowledge, not some convincing-but-misinformed “expert” on the internet.
Stay-at-home patients can consult as often as they like with Molly, the Nurse from Sensely. It, or she, is designed for chronic conditions, to track health, and offer support, even for palliative care. Voice recognition and deep empathy are key factors, but it also connects seamlessly with electronic monitoring equipment, so it knows what is going on at precisely the moment it is connecting with the patient.
AI Drug Development
One Pharmaceutical company had an AI comb through its database looking for existing molecules that would be effective in combating an Ebola outbreak. What could have taken months or years was accomplished in a day, producing two new drugs that could go to testing & trials. Drugs could be cheaper, tested sooner, and save millions of lives with the help of AI.
AI Problem Tracking
This is what AI is best suited for: analyzing data. IBM’s Watson computer, which you may recall beat the all-time Jeopardy champion, now lives on the internet where people can people utilize its data analysis skills for any purpose. Zorgprisma Publiek, in the Netherlands, uses public health records (which include clinic names, the condition treated, drugs prescribed, and treatment strategies) to inform clinics about misdiagnoses and improper treatment schedules so they can identify their common errors and improve their methodology.
AI Training of Medical Personnel
Familiarizing medical professionals with the capabilities of artificial intelligence can be performed by AI systems. These hands-on people wouldn’t need to be intimidated by introducing a layer of technology. Equipped with a high-level voice recognition interface, interactions would be like talking to a friend, and this could get these already intelligent people up to speed in short order.
Once that was accomplished, doctors could “consult” with the AI Expert System, to analyze problems, compare treatment strategies, and project outcomes. This introduces a whole new aspect of personalized medicine, making it even more customized to each person.
With the invention of modern medical telemetry devices, the amount of information being generated is incredible. Where no human could sort through all the incoming data, AI is ideally suited to the task. The more connected we are, the more personalized the data assessment can be.
The amount of exercise in a given period, combined with actual blood pressure, heart rate, oxygenation, and more could allow an AI to customize medical advice. It could alert at-risk individuals about problems in the early stages when they are still easy to treat and have the best chance of a successful outcome.
Not only will this provide an improved health status for the average person, but it will manifest itself as fewer expensive claims against insurance plans. As we all know, the healthier you are, the less expensive you are. AI is the future of medicine!