Modern technologies impact not only treatment protocols and therapy processes but the healthcare system itself.
Nowadays, artificial intelligence is being widely introduced into various areas, with its scope of application in medicine expanding year after year. In present-day clinics, AI is no longer an innovation; it is a standard technology that transforms the way medical organizations work.
AI and systems for big data analysis are implemented mainly to cut medical costs and optimize healthcare services. For instance, it has been estimated that in approximately 15% of cases, medics need to re-examine the patients' data 3 months after the previous appointment. Digitized health records stored on a cloud are much easier to navigate than paper charts kept in an archive.
One of the primary tasks that must be quickly resolved when patients are seeking medical help is to accurately assess who needs help right now and who can wait for a while. This sorting procedure, known as triage, is actively supported by AI technologies.
One such digital triage system, developed by Babylon Health, helps to analyze patients' call center requests. As a result, some patients are recommended to get immediate assistance, while others are advised to learn more about their disease and its symptoms (i.e., their cases are considered non-emergent). The system has been adopted by the UK's National Health Service and already saves costs on call center wages.
Another example is the joint program for triage optimization, implemented by GE Healthcare and the John Hopkins Hospital. The integration of AI to triage processes demonstrates impressive results, with emergency department patients getting assigned a bed 38% faster.
2. Switching to electronic documents
Systems capable of analyzing large amounts of data are also actively integrated into the healthcare systems of several countries. One of the states demonstrating the greatest progress in this area is South Korea: it initiated the digitalization of medical documents as far back as 2003. Seoul Bundang Hospital became the first "paperless" medical facility; by 2022, similar systems have been adopted by over 90% of national hospitals.
One of Bundang's most impactful developments is its BESTCare 2.0 system: during the first wave of coronavirus in March 2020, it promptly traced all contacts of infected persons, thus making it possible to avoid a total lockdown.
Another impressive large-scope technology is My HealthWay, an app that helps to manage three large databases: medical check-up data collected from national health insurance records, vaccination history, and prescription data . No matter which clinic the patient goes to, the physician can use the app to see their complete medical history; besides, the patient can also check the dates, causes, and results of their prior medical appointments. The system is expected to be fully integrated by 2023.
The most promising direction of AI and supercomputer application in medicine is diagnosing various disorders. The advantages seem evident: electronic devices have higher performance and minimize the possibility of errors by eliminating the human factor; besides, their use is economically beneficial, as robotic diagnosticians do not need salaries. On the other hand, a lack of human insights could become a drawback, as AI might fail to identify a non-typical case that could be diagnosed by an experienced professional.
Besides, the level of trust in machine-made diagnoses is lower than in diagnoses given by human doctors. Also, software costs remain relatively high.
Perhaps the most famous example of AI integration into medicine was IBM's launch of the Watson supercomputer. The machine that demonstrated outstanding results in other areas was ineffective when used for identifying and analyzing the data of large cancer care centers in 2015. The tasks Watson was expected to handle turned out to be too challenging for the supercomputer.
However, this launch became a landmark event: without IBM's story, any subsequent diagnostic and analytical systems would have probably been less successful.
Wrong diagnoses are indeed an outstanding issue. Experts at the John Hopkins University School of Medicine estimate that 10% of patients suffering from cancer, infections, or cardiovascular disorders get misdiagnosed, which leads to wrong treatment and further health complications. While in the case of infections, most errors are related to rare disorders, incorrect diagnoses happen more often when it comes to various types of cancer.
One of the most significant diagnostic solutions was designed by Zebra Ðedical, an Israel-based AI startup, and is used by medics in several countries to analyze CT scans and spot lung, liver, or breast cancer. To optimize operational costs and the workflow, Zebra offered hospitals to move all the algorithms to the cloud instead of paying for hosting their servers. In that case, only $1 will be charged for making and analyzing a scan.
4. Personalized treatment
The tendency of making treatment more personalized started to take shape several years ago; today, it seems obvious there is no generalized therapy that would fit everyone.
Artificial intelligence boosts the efficiency and accuracy of medical prescriptions, as well as optimizes surgical treatment. For instance, Accuray's CyberKnife improves the precision of surgeries: this AI-based system analyzes the patient's body specifics and the tumor location to make surgical interventions less traumatic.
Another helpful tool, designed by Vicarious Surgical, combines the advantages of virtual reality, artificial intelligence, and robotic surgery and enables it to perform the least invasive surgeries.
5. Hospital work optimization
Any medical facility is a complex mechanism, with doctors, nurses, and many other personnel members contributing to its operations. Its stable functioning is the prerequisite for fast and precise medical assistance.
AKASA is one of the most popular AI- and ML-based systems used in the US, already integrated into hundreds of hospitals all over the nation. It can be configured to meet the requirements of a specific medical facility. For example, the system helps to automate claim management, reducing the time needed to process a claim from seven minutes to one and increasing the number of processed claims. The medical personnel do not need to waste their time on paperwork and can focus instead on the patients and their problems.
Another aspect involves wielding AI capabilities for routing patients and tracking the position of each hospital staff member, which is especially relevant for major medical facilities. Many clinics already use indoor navigation apps that set an optimal route for patients to the office they need to visit. That prevents situations when patients lose their way in a hospital building, get late for an appointment, or fail to show up whatsoever - over 85% of patients have to ask staff members or other patients for directions.
Yongin Severance Hospital demonstrates a remarkable example of applying multiple innovative technologies, including artificial intelligence. It provides 5G coverage, enabling students or interns to watch real-time streams of complex surgeries in a classroom or a video hall without signal jams or image quality drops.
Thanks to the in-house monitoring system, it is possible to learn at any moment where an employee is located, which procedures a patient has undergone, and what the results of recent check-ups or tests are. This principle applies not only to the people but also to medical equipment, e.g., ultrasound scanners. The monitoring system helps to avoid extra movements (there is no need to spend time searching for something or someone) and reduces the number of human interactions, which is a critical factor during the ongoing coronavirus spread. Hospital patients can take advantage of other bonuses as well: apart from high-quality treatment, they can also enjoy a "virtual visit system" that helps to communicate with relatives and close ones even if they are unable to visit them in person.
6. Clinical trials
Another scope of AI application is clinical trials, with machines already performing tasks like finding new drugs, reprofiling existing ones, and finding patients that match the trial criteria.
For instance, scientists at Mount Sinai Medical Center performed an AI-assisted topological data analysis that studied medical records and genotype information of patients with type 2 diabetes, divided them into three subtypes, and forecasted the response of each subtype to the drug during the trials.
Another impressive technology was developed by the OPYL startup. It analyzes content generated on social media to identify users with an increased risk of neurodegenerative disorders, e.g., Alzheimer's.
Although human specialists still have the final say on those matters, the latest technologies can cover as much data as possible and ensure no relevant piece of evidence will get lost.
Rustam Gilfanov is a private investor, philanthropist, and venture partner of the LongeVC fund.