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?
1. Providing Diagnostic Assistance
“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.”
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 its human counterparts. AI systems also maintain a consistent level of performance regardless of factors like fatigue or distraction, which can affect human radiologists. Therefore, 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 integrate radiological findings with electronic health records (EHRs) and other clinical data. Thus, it helps to provide a more comprehensive, holistic analysis. It can also help with quality control by flagging potential errors and inconsistencies in medical images that lead to misdiagnoses.
2. Aiding in Drug Discovery and Development
AI is also transforming drug discovery and development by leveraging 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 3D structure of proteins assists researchers in designing drugs that can interact more precisely with specific proteins. The outcome is 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 investigate drug repurposing, predict drug toxicity, and identify disease biomarkers.
3. 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.
Artificial Intelligence steps in by analyzing large datasets, making predictions and insights that enable more personalized, targeted, and effective healthcare.
For instance, for diabetes treatment, precision healthcare considers an individual’s lifestyle factors, such as diet, exercise habits, and stress levels. Then, doctors use AI to 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. In other words, machine intelligence helps create personalized plans that also better engage patients in their healthcare journeys.
4. 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.
PA, then, leverages the power of data so that organizations make informed predictions and better decisions while reducing their risk.
For instance, AI can assess large datasets to identify both patterns and risk factors associated with disease. Thus, it helps healthcare providers estimate the likelihood of certain individuals developing conditions and diseases. They 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.
In other words, predictive analytics, when combined with early intervention, can improve patient outcomes while creating a more proactive and efficient healthcare system. Learn more about the growing importance of data analytics to healthcare.
5. 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 inquirers 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.”
Artificial Intelligence helpers are used in telepsychiatry and with predictive analytics to identify at-risk individuals. But it is AI’s use in chatbots for mental health support that might be one of its more important functions. Why?
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 mental health. However, the US has a surprising lack of resources. In the same survey, 74% of people stated that they didn’t believe these services were accessible to everyone whereas 41% said options are limited. And because of high cost and inadequate health insurance, 25% of those surveyed admitted that they often had to choose between getting therapy and buying necessities.
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 stigma of getting help. In a recent study, 42% of employees confessed to hiding their anxiety from their employer. They were worried about being judged, demoted, or, at worst, fired.
Accessing Therapy Through Artificial Intelligence
But chatbots, who turn out to be good for things other than writing your term papers, could help in this mental health crisis.
That is, chatbots are often the first line of treatment for those with limited access to and funds for therapy. According to a Forbes Health article, the cost of therapy for those without insurance ranges between $100 and $200, depending on your location. And 26 million Americans (7.9% of the population) don’t have health insurance.
Chatbots, then, assist with therapy cost, accessibility, and patient privacy. These apps are available when people are at work or done for the day: ready when they need them. A study by Woebot, in fact, found that 65% of their app’s use was outside normal hours, with the highest usage rate between 5 and 10 PM. Although there are health hotlines and hospitals, most doctor’s offices and clinics close at 5 PM or earlier. So people who can’t leave their jobs to get help are often left stranded. (And, for employers, it turns out that these apps also reduce work downtime, too.)
Two of the most popular AI-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 will continue to integrate (if not infiltrate) healthcare systems, inform new technologies, and guide medical interventions. Indeed, this blog 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.
Besides its online certificate and MS in Applied Statistics, Michigan Technological University offers several online health informatics programs through its Global Campus. Along with Foundations in Health Informatics and an Online Public Health Informatics certificate, the Michigan Tech Global Campus has two other certificates and a graduate degree. 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 of MTU’s Health Informatics Programs.
