According to the report of the McKinsey Global Institute, “Notes from the frontier of the IA”, from April 2018, the impact of artificial intelligence on the value of companies by sector will affect tourism (128%), followed by transport and logistics. (89%); retail trade (87%); automotive and assembly, high technology, gas and oil , chemical industry , media and entertainment, raw materials, agriculture, and banking (50%); health, public administration and telecommunications (44%); pharmaceutical products, insurance, semiconductors, the aerospace and defense industry (30%) . These statistics give us an idea of the depth that Artificial Intelligence is immersed in our society. In my case , I will only refer to the influence of AI in the medical field.
The big companies in Silicon Valley have opted to include artificial intelligence in developments that are related to the field of health. Advances in terms of artificial intelligence can even translate into economic advantages. According to a report by the firm Frost & Sullivan, by helping to diagnose and detect diseases prematurely, artificial intelligence will reduce health spending. Also , these services are available to people since they can access them from their phones or any other smart object like the clock. Which leads us to ask ourselves : could tomorrow’s medicine be a computer program? People can already, in some cases, know what disease they have without going to the doctor, since in an application installed on their cell phone it is possible to find the answer. In addition , you have the possibility of keeping track of your health status and the application alerts you when something is abnormal.
Although artificial intelligence applied to medicine has allowed research to go more quickly, data still needs to be integrated into the health care service where doctors can know what is available or how to use the new tools.
But what is Artificial Intelligence?
Artificial intelligence is considered a branch of computing and relates to a natural phenomenon to an artificial analogy through computer programs. Artificial intelligence can be taken as a science if it focuses on the elaboration of programs based on comparisons with the efficiency of man, contributing to a greater understanding of human knowledge. If on the other hand it is taken as engineering, based on a desirable input-output relationship to synthesize a computer program. ‘the result is a highly efficient program that works as a powerful tool for those who use it.’ Through artificial intelligence, expert systems have been developed that can imitate man’s mental capacity and relate syntax rules of spoken and written language based on experience, and then make judgments about a problem whose solution is achieved with better judgments and more quickly than the human being. In medicine it is very useful when 85% of diagnostic cases are correct. ‘
Artificial intelligence and Medicine
The acceptance of health-oriented applications and devices have grown exponentially in recent years, and specialists believe that this trend is due to the use of smartphones. Lumoid.com conducted a study that shows this trend. The results of this research indicate that 72% of adolescent women buy a Smart Watch or Smart Band because they can obtain information related to the quality and quantity of sleep and have access to tools to monitor their physical activity. The collection of information is the biggest problem in the use of machine learning. For example, when you want to create an algorithm that predicts a disease you need many medical records, most of which you cannot access. There are hundreds of startups dedicated to health care using artificial intelligence. In the following paragraphs I will mention some medical areas in which artificial intelligence is used.
Patient Information and risk analysis
The patient’s medical information is handled digitally , which allows them to store all their clinical history and have enough data to make predictions to prevent future diseases. Researcher Narges Sharif-Razavian using AI has been able to detect diseases in their initial stage, taking information from the multiple results of laboratory tests; the results of his research are very interesting especially for the early detection of diabetes 2. Another example of how this technology is being applied is the ‘Cardiogram’ project, an application that uses the Apple watch sensor to measure heart rate in real time and, by means of algorithms, it detects when the heart rhythm is not normal and warns the person to be alert. According to them, the results were better than the average diagnostic rate achieved by doctors. Other types of more complex diseases can also be detected by means of artificial intelligence. A team from the University of New York designed algorithms that allow to detect accurately and in time different diseases, cardiac insufficiencies. C on artificial intelligence documents such as medical guides are improved. ‘ A medical guide is a document in which the possible diagnoses are recorded given a list of symptoms. ‘ These guides, when fed by learning algorithms, would allow doctors to give a successful diagnosis of a certain disease.
Diagnosis through image analysis
Normally medical diagnoses are performed through the visual review of a patient or laboratory results such as radiographs or resonances . With AI, researchers using many images of some disease can detect with great precision the presence of that disease. Molescope , un Android application and IOS helps benign moles detection and early detection of skin cancer. Researchers at the University of Bari (Italy) using a magnetic resonance imaging system of 38 Alzheimer’s patients and 29 healthy subjects managed to detect with 84% probability the process of cognitive deterioration that causes the disease. This tool was able to predict with 82-90% reliability that patients will suffer from Alzheimer’s 10 years before the first symptoms.
Monitoring and management of lifestyle
M any of our personal devices have a myriad of tools to manage our lifestyle. Among them we have tools for the analysis of sleep quality in devices such as Smart Bands. ‘ The quality of sleep is usually calculated by recording the time you spend during the deep sleep phase, that is, detecting how much you move while you sleep, these devices also have the option of setting alarms capable of rising comfortably without drastically removing you from deep sleep. ‘Also, other tools that allow us to have a control diet or exercise s and predict the loss of fat tissue.
In the case of nutrition there several tools , VITL (https://vitl.com/), a text assistant (chatbot) which based on a series of questions gives recommendations based on dietary supplements or diets. Other tools like nuritas.com have accurate systems to identify your health and hence starting their learning algorithms recommend the best way to improve your condition.
Surgery and emergency rooms
Devices for monitoring surgeries at the time of an emergency include learning software, is the case of companies such as Gauss Surgical which develops solutions that give an estimate of blood loss, helping to optimize transfusion decisions, recognize the state of hemorrhage and improve patient outcomes. Other devices are the pain meters used in a surgery. The company MEDASENSE built a device that with the help of artificial intelligence could measure the amount of pain a patient suffers.
Hospital administration and patient safety
The administration of a hospital is crucial , poor administration can cost a patient inattention or incorrect delivery of drugs and this can result in the loss of a life. Companies like Mdanalytics carry out market analyzes on new health care techniques, including the latest research on drugs, medicines, government health agencies and medical associations. Some other companies have developed management applications which consider all the best practices to improve health care. These applications work from the capture of patient information, visualization of reports and the schedule of actions taken. Systems of this type mostly work in the cloud and specific services such as Amazon Web Services due to the amount of information to store and the need for rapid processing of this .
Imagine that you have a rare disease, making a consultation with a specialist is extremely difficult in some cases and obviously expensive at the same time. Many companies have been given the task of solving this problem, currently there are groups of specialists which can be addressed through specialized online video chats which give the doctor access to all patient information. The use of expert systems on chatbots has also been adopted, that is, a chat that has the possibility of answering technical questions in this case, giving diagnoses.
American scientists have discovered an early warning system to detect depression through the photos the patient posts on social networks. The new research, published in the journal ‘EPJ Data Science’, shows that computers, by applying machine learning, can successfully detect depressed people from clues in their Instagram photos. The detection rate of computers is 70% more reliable than the 42% success rate of general practitioners who diagnose depression in person. In addition to these studies, automatic learning systems have also been developed for the detection of stress and anxiety. Within this area of mental health, the idea of using virtual assistants has also been implemented, for this case the knowledge that is intended to simulate is that of a psychologist giving online therapies.
One of the largest markets focused on health is that of the wearers, in the first 5 months of 2017, Kaspersky Lab researchers detected 7,242 malware samples in IoT devices, 74% more than the total number of samples in the lapse from 2013 to 2016.
The rise of the Internet of Things (IoT) has made hackers attack more and more all the variety of devices connected to the Internet. The global health industry plans to invest around $ 410 million dollars in IoT devices in 2022. Along with this trend has seen the emergence of some procedures such as Medjack, through which attackers seek to compromise the devices that connect to medical devices. E n August this year, one of the leading manufacturers of Pacer issued an alert calling for firmware update about 465,000 patients after discovering a vulnerability that gives the attacker the ability to perform different attack vectors that would impact direct on the health of patients. Currently working on security solutions such as the use of blockchain and communication through more secure protocols to increase confidence in the use of these medical devices.
There is no doubt that artificial intelligence is transforming health care. Undoubtedly artificial intelligence has significantly improved health care, from everyday applications that help nutritional control to specialized devices for the safer performance of surgeries and applications that provide expert diagnosis. The use of these tools will completely change the way in which patients are treated, directly impacting on better care, follow-up and more precise diagnoses.