Artificial Intelligence: the Intelligent Choice in Medicine
Artificial intelligence, or simulated machine intelligence, is a rapidly growing sector of the medical field. There are numerous uses for robots and AI in healthcare, from reading test results and analyzing scans, to performing simple surgeries and even diagnosing ailments. There are endless possibilities to implement this technology, and the benefits will extend beyond the possible detriments that some professionals and the public are worried about. The introduction and immersion of artificial intelligence into the medical field would impact millions of people and make healthcare accessible to people in places where it was nearly impossible before. The problem with healthcare these days is its dependent on your income. Those without health insurance are more reluctant to go to the doctor because they would have to pay out of pocket, thus increasing their chances of contracting something fatal without their knowledge, or leaving something untreated that grows into something more serious. With the introduction of AI, care can become more accessible and less expensive to those with lower incomes. Although despite its potential to introduce new insights, and streamline the way healthcare is provided between doctors and patients, it comes with concerns such as privacy issues, ethical problems, and even medical practice errors.
Others opposed to the introduction of AI to medicine claim that it will take away jobs from current practicing doctors and medical professionals. However, there is actually a physician shortage across the country, particularly in more rural areas, and even more so across the rest of the world. By creating machines and robots capable of providing the same quality of care as human medical professionals, we can make healthcare accessible to everyone no matter where they live. Another benefit to having AI machines is that they don’t fatigue during 36 hour shifts, and will be able to meet the demand of high case loads and increasingly complex patients as people get older. Still, many people are skeptical about removing a key piece, the interpersonal communication and interaction between doctor and patient that would be lost with a machine. Despite all of these concerns, artificial intelligence would bring more benefits than deficits to healthcare, and if we can do it safely, it is the right choice in medicine; moreover, with the way technology is progressing to become such a large part of society and our lives, this may be an inevitability.
The main problem with medical devices and machines on the market more recently is the approval process and how they are reaching the market in the first place. There is a loophole called the 510(K) premarket notification. Essentially what this allows manufacturers to do is bypass rigorous trials and human testing analysis and demonstrate safety, functionality and effectiveness by proving it’s “substantially equivalent” to another device that is already on the market (Bleeding Edge). According to the FDA website, substantial equivalence only occurs when a device “has the same intended use as the predicate and has the same technological characteristics as the predicate” (Center for Devices and Radiological Health). This premarket notification process allows devices to be placed on the market and used in real medical procedures around the globe, sometimes having never been tested on a human or even an animal before.
This means that devices that create complications after they are implanted into patients are passed to market without ever being put to the test. The Netflix documentary Bleeding Edge details 3 devices that got passed through the premarket notification process and ended up having detrimental health effects on their patients, mainly focusing on Essure, a form of permanent birth control for women. It also talks about faulty surgical mesh, and hip implants using cobalt. The documentary uncovers how data reported on the Essure device was incomplete, and even though Bayer, the company who created the Essure device, insisted that 99% of the women that participated in their 5 year clinical study were comfortable with the device during follow ups, the numerous lawsuits from women over painful side effects and incomplete sterilization say otherwise. One prominent reason people are opposed to innovation on the medical front and the introduction of artificial intelligence is because there are devices and machines on the market currently that have had numerous cases of malfunction or harmful side effects. Despite this flaw, the majority of the decisions the FDA makes during the approval process are valid and the procedures, devices, and machines approved for market use benefit more people than they harm. Now although there is a loophole in the approval process for new devices and machines, AI can still be utilized effectively and be successfully integrated into the health care field.
There are two branches of AI application in medicine: physical and virtual. The virtual branch of artificial intelligence is helping boost discoveries in genetics and molecular medicine based on machine learning algorithms. AI allows a medical network to learn from its experiences and strive for process improvements. One such technique is the MAS approach:
[…] the MAS approach proposes to capture the dynamics of individual patients, including their responses to received medications as well as their behavioral interactions within a larger societal ecosystem. [It] allows process mapping, facilitates control, and better supports changes to the system with a demonstrated increase in response to medication, decrease of costs and more efficacious interventions. Its implementation has allowed health systems managers to analyze the dynamics of system performance across changes in social, medical and criminal justice components (Hamet).
Processes such as the MAS approach have allowed for better identification of patients with a history of hereditary disease in their family or a heightened risk of chronic illness. Artificial intelligence can also be used to synthesize information from all over the world. Patients in Europe and Asia suffering from the same ailments as patients in North America or Africa can benefit from their doctors being able to collaborate and share findings across the globe in seconds. The other branch of AI in medicine, the physical branch, includes bots that actually assist either patients or surgeons in their every day tasks. The most prominent form of “care bots” (Hamet) as they are called is helping the elderly or patients with limited mobility; robots can help them get dressed, or remember to take their medicine among other things. Robots are also present on the surgery floor either as assistants to actual doctors or sometimes as solo performers on smaller, more simple procedures.
Another benefit that artificial intelligence machines would provide in a hospital setting is ease of data synthesis between physicians for a single patient. Instead of having to manually combine charts from other doctors or even other institutions, doctors can focus on actually caring for the patient while a robot designed to do that specific job can complete the data transfer in seconds. It would save doctors time, energy, and might even start to combat the burnout problem among physicians that causes them to retire or leave the field at a young age. Time is so valuable, and being a doctor is a very time consuming profession, with the use of artificial intelligence to do charts and analysis, it can give time back so they still can enjoy doing the activities they love. Happiness in physicians and care providers is linked to better success and care for patients. According to HealthITAnalytics, 61% of consumers surveyed would be willing to engage with an AI helper to go over financial transactions, schedule appointments, or learn more about insurance coverage options. 57% would be willing to engage with an AI health and wellness coach to help manage wellness or chronic illness. Artificial intelligence has many uses, and the public is open to the possibilities that it holds. Another benefit AI integration holds is a great reduction in cost of care coming directly out of a patient’s pocket. By making care more accessible to remote regions, and just patients all over the care spectrum in general, the number of times they have to take distant trips to specific hospitals or specialists is reduced, and therefore reducing the cost of overall care.
Artificial intelligence machines can also reduce the error in disease diagnosis. According to the US Institute of Medicine, one in ten medical diagnoses are wrong and in primary care one in twenty are diagnosed incorrectly. These misdiagnoses lead to 80,000 preventable deaths annually in the US alone. Stanford conducted a study with an AI machine in which it analyzed 100,000 pictures and learned what each of the skin problems were and what symptoms exhibited what ailments, and then later analyzed 14,000 new pictures: the AI machine was able to diagnose melanoma and other skin ailments more often and more accurately than long term dermatologists (Sukel). Skeptics claim that results will be skewed, because diseases are always evolving and sometimes, old symptoms won’t work for recognition of new ailments. But there is a simple solution to this, human-supervised and unsupervised learning for the artificial intelligence machines will be monitored to fine tune the algorithms and make them as accurate as possible. AI machines will be able to associate presentations of diseases not yet known by human doctors or pathologists, and those results will be reflected in the results on those machines, and will allow for doctors to be able to learn along with the AI to use this new found data in a predictive way (Sukel).
Much of the rhetoric surrounding the discussion of artificial intelligence in medicine is focused on the lack of face-to-face interaction with the use of AI. Patients wouldn’t be able to ask questions or learn more about the condition they were diagnosed with, and that is why complete erasure of human doctors is in the very far future, if it ever happens at all. One machine that has sparked much conversation is the da Vinci Surgery System. The da Vinci surgery website describes the machine as an innovative and minimally invasive surgery technique. It details how it is used, and why it is better than traditional surgery methods. It mentions the da Vinci technique as being less invasive; it has the technology to make doctors’ movements more precise and accurate inside the patient, and allows them to be 100% in control of the system, and what is going on inside your body, all remotely. The da Vinci surgery website is relatively simple and lacks specific information about procedures, which is understandable because it would vary from patient to patient. However, even though they promote this new method as being superior and the right choice, the longest most in depth page on their website is the one detailing the risks associated with surgery.
Although they break down that information into 4 sections: risks associated with surgery in general, risks associated with minimally invasive surgical procedures, risks associated with the da Vinci Surgery System, and risks associated with representative specific surgical procedures. By mentioning the general surgery risks and minimally invasive surgery risks first, they are providing a foundation of doubt for the audience to question traditional techniques, which allows them to be more open minded to new techniques even if they carry more risks. The Bleeding Edge documentary from Netflix talks about risks associated with the da Vinci surgery system, including device malfunction, and tears, cuts, and burns from the limbs of the machine touching the wrong area of the body. However, there are always complications and a buffer period when a new technique is introduced, and reported cases of negative side effects are rare. Doctors need time to get trained and learn the ins and outs of the new technique, which is why when the da Vinci was introduced there were complications in some of the early procedures. The way the general public talks about AI in medicine, and especially the way they talk about the da Vinci system specifically, is generally in a bad light.
Older generations are more skeptical of new procedures because they know what works. Even younger people are concerned about the advances, but the fact that the da Vinci website lists possible complications that could be associated with a da Vinci procedure, they are improving their ethos, and allowing them to seem more reliable rather than just blindly promoting their product. What I think gets lost in translation between the companies and supporters of artificial intelligence and their audiences is that there will still be doctors present, at least for the distant future. There won’t be a complete elimination of human-patient interaction, but processes that can be streamlined, and even cheapened with the help of AI would be much more efficient if the technology was implemented. There seems to be a lack of depth in most easily accessible information about these new products, so people who don’t want to spend hours researching every aspect of each new idea just get a general overview, leaving out a lot of important details.
Another thing I noticed about the rhetoric surrounding AI and medicine is that most of the news stories or headlines you see regarding this topic tends to deal with negative aspects like botched cases or fatal accidents. The media grabs hold of one tragic story or accident and promotes it without getting all the background information. Also, if nothing goes wrong and everything about the device or new procedure works, then a lack of information is a good thing. But lack of information also leads to snap judgments, and snap judgments are another reason people are opposed to new innovation. If a person has a preconceived notion about something being bad, they are less likely to try it because of what they heard or read about it.
The introduction of artificial intelligence would greatly benefit medicine. Not only would it allow for more doctors to be able to try and save a patient by making charts and information accessible all over the globe, but it could take some of the load of work off of doctors so they are more well rested and able to perform at peak level as often as possible. Robots can physically help patients with rehabilitation or everyday tasks, and even help doctors in the operating room to provide precise movements or a camera view a surgeon wouldn’t normally be able to see. Some people are opposed to the idea, because there have been faulty devices passed to market before. However, with constant machine learning, this technology will be able to actively learn from its mistakes in order to prevent it from happening again, just like a human doctor could. Technology is becoming such an increasing part of our everyday lives, so it will inevitably be introduced into the medical field on a larger scale. If we begin to do it now, the machines will have a longer amount of time to gather information to provide the best results. AI is the future of medicine, and with safe introduction and close monitoring, it will become the new norm.