The healthcare industry is changing, driven by medical advances that arguably are being made faster than any period throughout human history. Healthcare practices that have been the convention for centuries are being abandoned as new medical communication methods and healthcare services are adopted. These advances help to save time and to deliver services that would otherwise be unattainable. In some cases, they can grant people several extra years of life.
Despite these considerable advances it is not unreasonable to expect the most revolutionary changes to happen over the next 3–4 years, and some of the services in question are already being implemented. For example, there are clinics that offer online medical consultations and e-prescriptions. In Sweden, Hurley Clinic conducted a pilot project for six months, implementing internet-services in their ambulatory practice. Around 133,000 patients took part in the experiment, which demonstrated that 60% of ambulatory appointments could be performed by doctors remotely.
In future, experts say that online consultations will replace up to 50 percent of all ambulatory doctor visits. Meanwhile, in-person visits will only be required in certain cases. This will help reduce patient overload and get rid of queues in public hospitals. Doctors will gain access to medical records using chat bots, which will enable them to better prepare for visits and to carry out consultations faster.
Chat bots will also improve patient quality of life, especially for people with limited access to healthcare facilities, including those living in remote regions. Simple chat bots will help to find specialists in one’s location and to book an appointment; they will remind people to take their medication and monitor the person’s health. More advanced chat bots will be smart enough to ask relevant questions about symptoms and notify a human nurse or make an appointment with a specialist when necessary. There are some specialized medical chat bots that can check the compatibility of two concurrently prescribed medicines, or for people who have to constantly monitor their health. For example, those who suffer from diabetes. Still other chat bots can notify insurers or take care of the patient’s bills.
E-prescriptions are also gradually becoming commonplace. To be able to use them, a clinic needs three things: access to a patient’s medical records, an agreement with their insurer and an agreement with a pharmacy (or a pharmacy chain). Specialized decision support systems that check the compatibility of a prescribed medication with other drugs taken by the patient would also help. When a doctor makes an electronic prescription, it is also sent to the selected pharmacy that will provide the necessary medication. E-prescriptions also often contain information about alternatives that are appropriate for a particular patient.
Telemedicine and e-prescriptions are only the beginning. What will come next? A couple of years after the rise of telemedicine, we will probably witness widespread adoption of artificial intelligence in healthcare. Neural networks will assist clinical diagnosis: for example, they will help with non-invasive diagnosis of different tumors, from melanoma and other types of skin cancer to brain tumors. Using millions (if not billions) of medical images, researchers from different countries are training neural networks to recognize malignant tumors. AI could soon be used to take a first look at one’s X-ray and MRI scans in order to decide whether a human diagnostician needs to see them. Does this mean that the first generation of neural network models will replace human staff? The answer is yes and no. The less qualified jobs will disappear as they will be replaced by AI, but the demand for specialists who can work with new technologies will increase.
The second generation of AI in medicine will be able to perform medical image markup and write reports to speed up diagnosis. Another promising research field is predictive analytics. Based on personal health records of a particular patient, AI solutions will help to assess a person’s risk of developing certain diseases, including cancer, and predict their frequency. All of these developments will emerge during the next 5 – 10 years.
Over the next decade, we will see widespread adoption of AI-powered clinical decision support systems (CDSS) that will be integrated with clinics’ IT systems. CDSS applies machine learning to clinical data in order to help nurses and physicians to make optimal decisions, avoid side effects and do many other things.
Widespread implementation of CDSS will have a strong impact on health care specialists (physicians, pediatricians) – low-skilled doctors will lose their jobs, while high-level professionals will gain yet more value based on their capacity to deal with matters outside the scope of AI. In the next 10 years we will see huge investments in this sector as well as breakthrough opportunities for specialists that are able to work with medical AI.
All of these innovations are dependent on Big Data and the accumulation of all sorts of medical information. Wearable gadgets that monitor heart-rate and physical activity and take blood pressure, will empowers doctors with more information about their patients even if they’re meeting them for the first time. With real time monitoring, patients can be alerted that they need to take medication or do more or less exercise. The latter option may give peace of mind to those who have health issues, but nevertheless like to exercise.
Predictive analytics and medical AI can also be applied to hygiene issues in order to improve public health. Having detailed information about the number of visitors to a particular restaurant, or how many teachers and students attend a certain high school, sanitary inspectors can use AI solutions to create risk-profiles for those facilities. In poorer countries, such information can be a question of life and death. If there is an outbreak of an epidemic, such as influenza, Ebola or the Zika virus, information about school attendance can help doctors to identify the most affected areas. Meanwhile, disease mapping will provide them with further knowledge to predict the place of the next outbreak.
