In theory, AI can do a lot: diagnose, treat, perform surgery, and discover new drugs. It is supposed to be a panacea for staff shortages and a tool to personalize treatment; it should democratize access to medical care and reduce the prevalence of non-communicable diseases.
Sound too good to be true? Unfortunately, between technology and its practical application, there is a massive gap in acceptance, trust, legislation, psychological factors driving people’s behavior, resistance to change, and knowledge. But nothing is such a roadblock as a health system that pays for the quantity of services provided, not their quality.
This causes even state-of-the-art technologies to have an uphill battle. Putting the obstacles to innovation aside, let’s examine where AI could help.
Problem: disease-oriented healthcare, neglected prevention, the rise of lifestyle- diseases, and low patient compliance.
Solution: AI-driven digital health assistants.
Imagine one mobile app with multifunctional applications for health. A chatbot that gives us health guidance based on information from electronic medical records and reminds us to work out, eat well or take a pill. An app that tracks your health to see if you’re okay based on data collected from wearable devices or symptom checkers. On top of that, it will schedule a doctor’s appointment, not forgetting vaccinations and regular health checks. All in one app thanks to artificial intelligence algorithms that analyze our data and help us make the best decisions. And all the actions orchestrated by AI and smoothly embedded in everyday life.
Problem: medical errors, different standards of care, biased decisions.
Solution: AI-supported clinical decision support systems.
Everyone should have access to treatment according to current medical knowledge and the best doctors. Thanks to AI, this is possible because clinical decision support systems can compare the results of treatment of patients from all over the world, select the most effective treatments tailored to the individual attributes of the patient, and suggest procedures following the recent scientific guidelines. Such standardization would allow patients to be treated similarly, regardless of location or social status. AI would learn from those best – specialists and clinics- and make the knowledge gained available to every doctor. Doctors’ decisions would be based on the latest clinical evidence collated with data in the patient’s electronic record.
Problem: medical staff shortages, doctors/nurses burnout.
Solution: electronic medical records fed by data captured by algorithms
Doctors know that access to patient data is crucial to evidence-based clinical decision-making, but they don’t like to enter data into the EMRs. No wonder, since it takes up to a third of their time spent with a patient, according to various studies. Voice recognition systems coupled with generative AI can convert voice to text and then complete the data in the EDM using standardized medical dictionaries. Some data can come from wearables, sensors placed in the doctor’s office, and cameras recognizing a patient’s emotional state (the concept of the doctor’s office as a data lab). AI can summarize critical information by displaying it on the screen as a clear summary or alerting the doctor in case of abnormal results. No typing or manual review of data, no clicking.
Problem: preventable progress of chronic diseases, readmissions.
Solution: remote monitoring with automated signal signs analysis.
Between follow-up visits usually planned every few months, chronic patients are beyond care. It can be changed by introducing continuous health monitoring for every chronic patient using simple wearable devices and telecare platforms (widely accessible within universal health coverage). AI would analyze the data transmitted to the patient’s app – the doctor is informed if health parameters worsen. In addition, the patient would receive real-time guidance according to the development/regression of the disease. Furthermore, the captured data could be re-used for scientific purposes. Remote monitoring also allows more patients to be treated at home and reduces hospital admissions (and high care costs).
Problem: rapid increase in the prevalence of mental illness and limited access to support.
Solution: companion avatars and digital mental health programs.
There is no longer a medical specialty with no shortage of doctors. But psychiatry faces the biggest staffing crisis – the WHO warns of a mental illness epidemic with overwhelming health and economic costs. Combining generative AI systems with medical data/expertise and wearables could offer the chance to develop a new generation of chatbots – empathetic, responsive to a patient’s emotional state, and able to carry on a natural conversation. Such systems could bridge the gap between the first symptoms and consultation with a psychiatrist, helping millions of patients. Although AI-based therapists are controversial, they are better than having no access to care. Developing such systems requires the close cooperation of AI engineers, psychologists and psychiatrists, and ethics and data protection experts.
Roy Amara, an American scientist, futurist and President of the Institute of the Future, said: “We tend to overestimate the impact of technology in the short run and underestimate the effect in the long run.”
While we are fascinated by the capabilities of ChatGPT, AI algorithms are slowly reimaging inefficient care. For this to happen, it will first take care to standardize data, balance data privacy with the benefits of access to data, and create trusted AI algorithms. But the biggest challenge may be to change the current health system, which lags far behind the opportunities offered by artificial intelligence and new technologies.
Problem: high administrative costs, long waiting times, inefficient processes, no-shows
Solution: smart medical offices and hospitals
Self-driving cars, self-service stores, and now also autonomous doctor’s offices? While there is a shortage of money for cutting-edge therapies and medicines, administrative costs consume from a few to even tens of percent of healthcare budgets. In the U.S., for example, it’s about 25 percent of the $4,3 trillion (in 2021) spent on health. Besides, studies suggest that 30 percent of healthcare spending is wasted. Using AI, a patient journey can be redesigned. For example, the waiting time for an appointment in a medical facility could be used for 360-grad screening and data collection. So that the doctor already has all the data when the patient enters the office.