Artificial Intelligence (AI) has been the trendy buzzword across multiple industries for quite some time to the date. It is often awarded epithets like “life-” and “game-changing” – or at least having such potential.
This is especially true when we speak of healthcare. There is plenty of research listing different market size numbers for “AI in healthcare” – the numbers are close in all the major sources and are listed around reaching over 40.2 USD billion by 2026.
The investors are eagerly hopping on the action in this sector, so are the end-customers – hospitals, doctors and other medical institutions.
Currently, the trends for healthcare in AI aim for the following:
- Improving patients well-being, outcomes, prolonging their life cycles
- Reducing the overall cost of medical services for the patients
- Acting in the interests of healthcare investors
While considering this meta, we’d like to have a look at the current state of affairs in the AI healthcare sector and form some conclusions about the shape of it in the nearest (and more distant) future.
We’ve studied several reports from the respected research and medical companies and interviewed our experts
New Trends in application of AI for Healthcare
Current emerging Artificial Intelligence applications are plenty:
Medical Imaging – AI-driven platforms power medical imaging applications that use deep learning for the analysis. In some cases, these tools lower radiation dose and enable contrast to produce 4 times faster scans, hence improve patient’s comfort and safety and the productivity of radiology workflow.
Management of Chronic Diseases – plenty of medical institutions (clinics, hospitals) use AI/ML to monitor patients current state with the help of sensors and data analysis automation. In some cases, medical facilities are poorly equipped for providing proper care 24/7 for chronics. AI is of great help while taking this patient-centred approach.
Wearables – the companies are integrating AI and the Internet of Things to better monitor patient adherence to treatment. This is especially useful for the patients in critical condition or those who require constant supervision.
Drug Discovery – there are more than 160 startups that use AI as a key differentiator in developing drugs or similar products and show traction, such as via seed funding. The number has grown 5 times in two years and that is only one source of tracking – Bench Science’s blog. That means that this number could be much bigger.
We could also list the following common applications of AI in healthcare:
- Lifestyle management
- Lifestyle monitoring
- Emergency room management
- Hospital management
- Risk analytics
- Insurance rates assessment
- Virtual assistant
Now let’s dive into more details on each of these application sectors for deeper exploration.
AI for Medical Imaging
NFL and NBA athletes get injuries here and there, and ESPN reports are full of the news about MRIs after each playing week. Earning their multi-dollar yearly salary, they could pay for it as easily as Thanos managed to erase
However, if your (or any of your close ones) insurance plan doesn’t cover it fully, you should be well aware that one MRI session would end up drying your wallet for up to $4,000, and even a simple x-Ray could build up a sum up to a grand.
No wonder, then, that a lot of focus has been put on developing cheaper AI imaging. Google DeepMind’s medical image-assistive AI can identify 50 sight-threatening eye diseases. there are AIs that serve as the markers for potential strokes, spot lung and liver lesions.
It doesn’t have to be a difficult platform available only on powerful machines – sometimes it is a low-cost app any modern phone could deal with.
The AI in medical imaging is developing in a huge tempo as it is being sponsored by the giant companies, for instance, Nvidia. The startup called Subtle Medical, the winner of Nvidia’s award, developed SubtlePET, a product that decreases the amount of time and radiotracer dose required for a PET scan (medical image technique often used for cancer diagnosis).
GE Healthcare has developed their Edison Application, a platform that provides comprehensive, actionable insights across medical imaging modalities – the same ones we’ve mentioned higher: MR, CT, X-ray and the ones alike.
That is possible due to multiple vendors by merging data from the machines and the Radiology Information System (RIS). Machine Learning is used for the analysis.
Certainly, there are doubters in the future of AI applications for medical imaging, but… We couldn’t load a page with an article that “Medical Imaging AI is the bubble that is doomed to break in 2 or 3 years. Funny enough, looks like these guys haven’t lasted long enough to see their prediction come true.
Getting More Personal: Chatbots, Wearables, Diagnoses and Predictive Care
The personal approach is everything in today’s healthcare. It is not only applicable in doctor-patient communication (we’ll talk of this later) but also long before these interactions occur.
There is such a digital tool called Buoy that allows potential patients to put their symptoms into the programm and receive a personalized analytics and care recommendation in real-time mode.
Buoy uses AI algorithms to analyze and calculate the inputted data and draw the conclusions about the customer should proceed upon it.
There is a strong opinion about whether AI is a good fit for patient care, and WSJ tech columnist Christopher Mims gives some decent tenets that it is.
In fact, the big progress in such pre-diagnostics (if we can call it that way) goes to chatbots. We won’t dive into it though.
We have covered the topic of chatbots in healthcare in one of our recent blog posts. You can read it here, and also if you are interested in a custom healthcare software development company for the realization of your ideas, don’t hesitate to contact us.
Instead, let’s talk about the importance of wearables and AI combination. When Artificial Intelligence is involved in the analysis of the patient data that comes, it helps to improve the quality of life and medical care for the patients.
The University of Waterloo (Ontario, Canada) researchers use AI on the gathered data to predict patients health failures. The smart shirt that measures heart rate, breathing and acceleration sensors makes it possible to “accurately predict health-related benchmarks during daily activities using only the smart shirt.”
This University also does a lot for preventive care. Their “lab in a box” replaces pricey and bulky microscopes with a simple device that fits in your hand.
Today a tissue sample analysis microscopes sized as a desk and cost from $250,000 to $500,000. Dr. Wong lists plenty of drawbacks:
“It’s very expensive. If anything breaks or shifts, you need a trained technician to fix it. You have to scan your specimen very slowly. Most countries won’t have these kinds of devices, because they can’t afford it. It’s just not possible”.
The device Wong produces is a hand-held one. It costs less than $1,000. In mass production, one unit would cost $200.
Not So Obvious Use Cases: Improving the Communication
We’ve listed pretty much obvious ways of AI adjustments that bring benefits to the healthcare industry. Now let’s speak of not that vivid examples.
No one argues that productive communication between the doctor and the patient is helpful to see improvements in the condition. Glyn Elwyn, professor at Dartmouth Institute for Health Policy and Clinical Practice, claims that AI is an innovative solution that could completely change the communication process in healthcare.
Most doctors have never had their communication skills formally assessed and do not know how they compare with their peers.
BMJ interviewed the specialist. He told that improvements in the conversations between physicians and patients could be made as the former would receive the report after the automatic analysis of such.
AI analysis of wording, phrasing in the dialogues helps to determine whether the two understood each other.
As for the future, AI also has the potential to analyze conversations in real-time and suggest diagnoses that they may not have been considering and offer a broad range of treatment suggestions.
Artificial Intelligence also can analyze and assess the tone and style of conversations.
There are plenty of drawbacks to this usage. Issues such as patient privacy come to mind from the top of the head. Also, the assessment of every workplace dialogue puts some additional pressure on the doctors. Also, currently AI systems are currently not fully capable of decoding the complicated dialogue that occurs in the medical setting. They are getting there though.
Instead of Conclusion: The Future of AI in Healthcare
The future looks bright as there is a distinct growth of artificial intelligence in healthcare. Despite some challenges in the upcoming years, and not being developed enough “to move mountains” right now, it still has significant benefits and collects tons of venture money.
AI startups have raised $4.3 billion across 576 deals in the last six years. Experts still believe that some good marketing is required to attract even more attention to AI in healthcare. Such marketing and technical teams need to obtain large quantities of data for AIs to learn from and also make it a point to optimize the imaging and processing systems to ensure the validity of these AIs.
We’ve delved into tens to hundreds of sources to gather up the material to this blog post, and the doubt that AI technology might be useless in such industry as healthcare was met in none of those sources. We dare you to find this.