In the Centre for the Fourth Industrial Revolution's Top Emerging Technologies of 2023 report, the organization nominated AI-facilitated healthcare as the 10th top global trend. After stressing how the COVID pandemic magnified the shortages of healthcare systems around the globe, the report suggested that perhaps one of the most important goals of AI systems is that of improving healthcare. But how, exactly, can artificial intelligence help if not transform healthcare?
Providing Diagnostic Assistance
By leveraging machine learning algorithms to analyze medical data and images, whether they are CT scans, X-rays, retinal images, and more, Artificial Intelligence can play a crucial role in interpreting medical images and data.
Although AI can't replace the expertise of radiologists, it can certainly supplement it. For instance, trained on vast data sets of diseases and anomalies, AI excels at recognizing both patterns and abnormalities in medical images. And it often does so with more efficiency, speed, and consistency than do its human counterparts. AI systems also maintain a consistent level of performance; that is, they are not affected by fatigue or distraction, which can affect human radiologists. It can also help with quality control by flagging potential errors and inconsistencies in medical images that may lead to misdiagnoses.
As a result, AI may provide more reliable results throughout the day and across different cases. This heightened accuracy could make the difference between life and death in emergency situations. AI can also integrate radiological findings with electronic health records (EHRs) and other clinical data. Thus, it helps to provide a more comprehensive, holistic analysis.
"Errors and discrepancies in radiology practice are uncomfortably common, with an estimated day-to-day rate of 3–5% of studies reported, and much higher rates reported in many targeted studies."
Aiding in Drug Discovery and Development
AI is also improving the efficiency of drug discovery and development by using advanced computational techniques to analyze large datasets, predict molecular interactions, and streamline various stages of the drug development pipeline.
Algorithms can assess large data sets (genomic, proteomic, and clinical) not only to identify proteins and genes associated with diseases, but also to pinpoint targets that will respond to clinical intervention.
Machine learning can also predict the binding affinity of molecules to target proteins. This prediction allows researchers to prioritize and then test possible drug compounds. In other words, AI can help both streamline and reduce the cost of pharmaceutical research.
For instance, in 2020, DeepMind's AI system (AlphaFold), made headlines by accurately predicting 3D protein structures from amino acid sequences. According to its website, AlphaFold regularly "achieves accuracy that is competitive with experiment." Why is this accuracy important? Understanding the three-dimensional structure of proteins assists researchers in designing drugs that can interact more precisely with specific proteins, creating more effective and targeted treatments.
Artificial intelligence can also optimize the design of critical trials by analyzing patient data, summarizing biomedical literature, identifying suitable patient populations, and predicting potential challenges. In addition, AI helps with investigating drug repurposing, predict drug toxicity, and identify disease biomarkers.
Personalizing Healthcare
Personalized medicine, also known as precision medicine, involves tailoring preventive care, medical treatment, and interventions to an individual's characteristics. These might include such factors as genetics, lifestyle, and environment.
For instance, for diabetes treatment, precision healthcare uses AI to analyze an individual's lifestyle factors, such as diet, exercise habits, and stress levels.
With this information, doctors can then create personalized dietary plans and exercise regimens to manage blood sugar levels effectively. Diabetes treatment might involve using wearable devices and health trackers to monitor and adjust lifestyle interventions based on real-time data.
Take Continuous Glucose Monitoring (CGM) systems, which provide real-time data on glucose levels throughout the day. Precision medicine utilizes this data to adjust treatment plans dynamically. Thus, Artificial Intelligence allows for more accurate insulin dosing based on the person's unique glucose patterns.
In precision medicine, AI may also assess an individual's genetic makeup in conjunction with their medical records. The goal is predicting how a patient may respond to certain medications and treatments. And in nutrigenomics, AI amasses a wealth of genetic, molecular, and nutritional data to make personalized dietary recommendations. Machine intelligence, then, helps create personalized plans that also better engage patients in their healthcare journeys.
Using Predictive Analytics to Improve Patient Outcomes
Predictive analytics uses various statistical algorithms, machine learning techniques, and data mining processes to analyze historical data and make predictions.
The goal of predictive analytics is identifying patterns, trends, and relationships within data to forecast future behavior or outcomes. By leveraging the power of data, organizations can make more informed predictions and better decisions while reducing their risk.
For example, AI can assess large datasets to identify both patterns and risk factors associated with disease. In doing so, it helps healthcare providers estimate the likelihood of certain individuals developing conditions and diseases. Providers can then suggest preventative measures, such as lifestyle interventions and early screenings.
AI also assists in monitoring patients, such as in chronic disease management. By continuously analyzing patient data, such as vital signs and other health metrics, it can detect deteriorating health conditions earlier. With this data, healthcare providers can intervene more quickly and reduce the risk of complications.
Adhering to taking medication is obviously crucial to patient outcomes. AI can help predict and improve medication adherence by analyzing patient data and identifying factors that may influence a patient's ability to follow prescribed medication regimens. Predictive analytics, then, when combined with early intervention, can improve patient outcomes while creating a more proactive and efficient healthcare system.
Acting as Virtual Health Assistants
Artificial Intelligence-powered virtual assistants are also used throughout healthcare. In these roles, they provide information, answer questions, help in preliminary diagnoses, improve patient engagement, and direct people to healthcare services.
DeepScribe is one such trusted medical scribe. After extracting information from the conversation during a doctor's visit, it uses AI to create a medical note. This note "syncs directly with the provider's electronic health record system, so all the provider has to do is review the documentation and sign off at the end of the day."
Accessing Therapy Through Artificial Intelligence
The National Council for Mental Well-Being, which announced a mental health crisis in the US, found that 56% of Americans seek some form of help for their mental health. However, in that same survey, 74% of people stated that they didn't believe these services were accessible to everyone whereas 25% admitted that because of the cost, they often had to choose between getting therapy and buying necessities. According to a Forbes Health article, the cost of therapy for those without health insurance ranges between $100 and $200, depending on location.
Accessing healthcare is also impacted by a person's location and income. That is, those who live in rural areas and those with lower incomes are less likely to seek mental health. And, for many, there is also the perceived stigma of getting help. But AI can help, particularly a certain type of chatbot that acts as the first line of treatment for those with limited access to and funds for therapy.
It turns out that chatbots are great at tasks other than mundane writing chores. That is, by acting as digital therapists, chatbots may assist with cost, privacy, and accessibility. A study by Woebot found that 65% of their app's use was outside normal hours, with the highest usage rate at 5-10 p.m., when most doctors' offices and clinics are closed. Virtual mental health assistants, in contrast, are ready when people need them, whether that is for a private session at work or at home, after a person is done for the day.
Two of the most popular chatbots are Wysa and Woebot, accessible through mobile apps.
Using principles of Cognitive Behavioral Therapy (CBT), these chatbots provide mental health support. They help users manage their stress and anxiety by tracking mood, offering coping strategies and mindfulness exercises, and engaging in conversations.
In other words, these chatbots create an anonymous safe space to talk about worries and stressors, working towards deescalating them. Others include Youper, and Replika, the "AI companion that cares."
Keeping Pace With Artificial Intelligence in Healthcare
There is no doubt that Artificial Intelligence continues to transform the workplace, it will also continue to integrate (if not infiltrate) healthcare systems, inform new technologies, and guide medical interventions. Indeed, this article has just scratched the surface of AI's potential. There is still much to say, for example, about AI solutions for bolstering healthcare infrastructure and services in developing nations.
Those wanting to make a difference in data-driven healthcare, and ensure that artificial intelligence is used responsibly, securely, and ethically require specialized advanced education.
Michigan Technological University offers several online health informatics programs through its Global Campus.
Along with Online Artificial Intelligence in Healthcare and Online Public Health Informatics certificates, Michigan Tech has two other certificates and an Online Master's in Health Informatics. In these CHI programs, students also get access to HIMSS, an interdisciplinary society that unites people striving to improve the global health ecosystem.
If you'd like additional information about these programs, reach out to the designated Graduate Program Assistant, Margaret Landsberger at margaret@mtu.edu. Or request information about one or more of MTU's Health Informatics Programs.