How AI transforms healthcare?
Medicine is, probably, one of the oldest industries, however, the medical knowledge that we have accumulated over millennia is still far from complete. The human body seems to be as unknown as the farthest corners of the Universe. For centuries, shamans, healers, and doctors have been the most respected people in the society, as they were the ones who could resolve the life-or-death matters in the very literal sense.
Today, medicine and healthcare enjoy the fullest support of the cutting-edge technology. Whole institutes are working on new drugs to treat the diseases that were considered incurable. Hospitals are getting powerful equipment to use in various treatment, surgery, or follow-up care protocols. Naturally, the computer industry is also making its contribution to healthcare. One of the actively explored areas is the application of artificial intelligence in medicine and healthcare.
AI has been already used to ease the administration and document maintenance in healthcare. Database support, medical record handling, schedule management – AI has proved its superior abilities in these tasks. However, artificial intelligence is capable of much more. There are several areas in medicine and healthcare where the outstanding powers of AI can be used for our benefit.
AI and medical data mining
Each patient accumulates quite a lot of medical data over the years, and sometimes it may be hard for doctors to find what they need in a particular case. AI can help organize the information and search for what is currently relevant. When the doctors work with the patient’s data that has been pre-structured by AI, they spend much less time to make the diagnosis or prescribe the treatment plan.
AI can sort between high- and low-priority CAT scans and X-ray images, so that doctors immediately get to work with what is important for the particular case. In medicine, like nowhere else, time matters, and it is often critical to shorten the time between the examination and diagnosis and treatment beginning. Here, AI can help a lot.
Using AI to analyze the patient’s previous history can help in creating personalized treatment plans designed to ensure the maximum effect for a particular patient. Oncora Medical, the company designing AI-based solutions for radiation oncology, has built a platform analyzing the patient’s past treatment details, such as how the patient tolerated particular medicines or radiation doses, how their disease responded to the treatment, and so on. This data is a great source of knowledge allowing to adjust the protocols for each patient to achieve the maximum effect while reducing the negative impact.
AI and diagnostics
Artificial intelligence is gaining increased popularity in the area of diagnostics. Its ability to process gigabytes of data quickly and efficiently allows it to notice the trends and deviations quickly. Often, AI can notice an irregularity that a human brain may miss, thus giving the patient a chance to start the treatment earlier.
AI with a powerful image recognition feature is capable of studying the images obtained by computerized tomography or X-ray and detecting even minor changes or deviations from the norm. Moreover, in such medicine areas as oncology the treatment can be sometimes almost as risky as the disease itself, thus it is important to correctly define the tumor as malignant or benign. A correct diagnosis means appropriate treatment, while an incorrect one may lead to losing the precious time or subjecting the patient to an unnecessary traumatic treatment.
Using AI in diagnosing illnesses not only helps to save time but also increases the accuracy of the diagnosis. The share of errors made by AI is less than that made by human diagnostics experts, as AI is not prone to fatigue or distraction and can notice the smallest irregularities.
Many companies and research institutions are working on implementing AI for diagnostics. Infervision, a Chinese company claiming to provide “an extra pair of eyes for radiologists”, applies machine learning to teach its AI-based platform to identify even the earliest stages of cancer by the images obtained through computer tomography. The platform has shown its efficiency in detecting the smallest cancer lesions that the human eye is unable to catch.
AI and surgery
The time when complex surgeries can be fully entrusted to robots are still far away, however, today artificial intelligence is already being put to use in operating rooms. Mostly, the robotic surgical instruments are used in microsurgeries where the stability and precision of movements are of extreme importance.
Robots convert the movements of the surgeon’s hands into micro-movements that are free of any natural tremor. An AI-driven surgical instrument can perform movements that are physically impossible for human fingers. Micro-movements mean smaller and more precise cuts and less suturing, which, in its turn, result in easier and shorter postoperative recovery.
Right now, we cannot say that the currently available solutions using robots for surgery are pure AI, as the robots do not operate on their own. The operation is planned and conducted by a doctor who uses the robot as an extension of their hands.
There are already several examples of successful surgeries performed by robots. In 2017, a robot built by Microsure was used in an operation on small blood vessels – as small as 0,3 mm – which went well and left the patient recovering afterwards. In this case, the surgeon coordinated the robot while the AI algorithms were applied to reduce the doctor’s hand tremor and achieve the maximum precision.
The experts voice their optimism about the further development of AI applications for surgery and hope that the quality of operations and the patient’s post-op recovery will become higher.
AI and drug development
The pharma industry is now suffering the consequences of the Eroom’s Law stating that the time and costs required for the drug discovery increase despite the technology progress in this area. The huge costs are the result of the trials of failed drug candidates. The share of failures can be as high as 9 out of 10, while they consume tremendous resources.
This problem is a subject of concern for many pharmaceutical companies who are already investing in AI solutions to assist in quicker and more cost-effective drug development. The report by Global Market Insights states that by 2014, the total AI healthcare market is going to reach $10 billion, out of which about 35%, or $4 billion, will be the investments in drug discovery.
To a large extent, drug development is about analysis and processing of data. The volume of pharmaceutical data is getting bigger with each passing year, as now there are already billions of compounds that have drug-like features. You can imagine the amount of data associated with them – the targeted diseases, possible outcomes, side effects, toxicity levels, and so on.
AI can be successfully used to process this data to identify patterns and certain key features. Moreover, artificial intelligence can be taught to predict the possible drug efficacy or toxicity. This can reduce the time and costs required to discover the drug greatly.
In 2017, the ATOM (Accelerating Therapeutics for Opportunities in Medicine) consortium was founded by GlaxoSmithKline pharma company, the Department of Energy’s Lawrence Livermore National Laboratory, the National Cancer Institute’s Frederick National Laboratory for Cancer Research, and the University of California, San Francisco. Among other technologies, the consortium uses AI to develop a fast and effective drug discovery process that will allow reducing the time needed to prepare the drug for clinical trials from 6 years to just 12 months.
GSK is going to provide the data it accumulated from drug trials and researches to create the base for AI-driven predictions. Together with data science, supercomputer engineering and other technologies, AI will be used to form a drug discovery platform revolutionizing the traditional drug development processes.
AI and preventive care
Everybody knows that a disease is easier to prevent than to treat. Unfortunately, too many patients apply for treatment when the disease has already progressed too far. This is why there is so much talk about preventive care – the healthy lifestyle, vaccinations, regular checkups. Proper preventive care can help to catch the disease at early stages giving good chances of a complete recovery or lightest therapy. At the same time, patients with chronic conditions, such as diabetes, need regular monitoring to detect the symptoms of a deterioration that requires urgent treatment.
In the area of preventive care, artificial intelligence shows great results when used together with the Internet-of-Things technology. Various wearable devices that can collect, transmit and analyze the patient’s data are used to detect the onset of quite a number of illnesses. With an AI-powered wearable device, the patient does not need to visit their doctor too often, however, their condition is monitored at all times.
Another way AI can be used for disease prevention are self-diagnostics applications. By feeding in the symptoms, the patient can get a tentative diagnosis and the basic recommendations. Of course, such apps should never be used instead of a doctor’s consultation, as they only outline the probable cause of the symptom.
These apps are built as chatbots with the natural language processing feature. They can recognize the plain-text description of the patient’s condition and suggest the diagnosis. One of the best-known diagnostics apps is Your.MD that uses AI to analyze the symptoms and provide the diagnosis.
Are we ready to be treated by artificial intelligence?
This is a big question, as no matter how advanced the technology gets, it should be accepted by the public. People tend to be very particular when it concerns their bodies and health. However, despite the expectations, the public response to AI in healthcare is rather positive and welcoming.
A recent survey showed that 47% of the respondents admitted to be comfortable with AI assisting their doctors. However, those uncomfortable with AI said that their discomfort was due to the “lack of human interaction” during their treatment. It seems that people consider empathy and understanding to be as important as the accuracy of diagnostics and the effectiveness of treatment.
At the same time, the same survey proved that only 35% of the participants felt assured about the security of their data provided to AI-based tools. The security concerns may be a show-stopper for a wider AI adoption in healthcare.
In any case, with the amount of medical data that is currently available – hundreds of diseases and conditions and billions of chemical compounds to be used as drugs – the human brain may become unable to process it all. This is where artificial intelligence will lend its shoulder and offer its powerful capacities to produce a correct diagnosis or treatment protocol.
Despite the rapid development of the technology, we should not fear that AI is going to replace human doctors. On the contrary, artificial intelligence is used to assist doctors where the human memory cannot process the huge amounts of data or the human hands cannot perform micro-movements.
No matter how advanced AI gets, it will never be able to express human feelings and emotions, and emotions can have as much effect on the patient as the medicines. We need to feel the care and empathy, and we cannot get them from a machine.
We are sure that artificial intelligence is going to change the healthcare industry, and quite soon. We should embrace this change, as it will improve the way healthcare is provided and let us live healthier lives.
Stay healthy and subscribe to our blog for more insights from the world of advanced technology.