By 2035, it is expected that two-thirds of UK adults over the age of 65 will have more than one health problem.  This is called "multi-morbidity." Seventeen percent of people would have four or more illnesses, which is twice as many as in 2015. One-third of these people would have a mental illness, such as dementia or depression.

Life expectancy has gone up by about three years for both men and women. This means that people will spend more time with more than one illness.

The problem is there are too many people and not enough doctors. Other industries solve this problem by using structured "algorithms" to make product development more scalable, efficient, and consistent in quality. Can the same thing be done in healthcare to make it more algorithmic, scalable, and cost-effective while improving patient outcomes and making it less variable?

Nothing can replace a clinician's judgement when it comes to patient care, but as care continues to shift from treating diseases after they happen to proactive and preventive care, more organisations are looking to advanced technologies like artificial intelligence and machine learning to help them give the right care to the right patient at the right time.

The application of AI has already been institutionalised in other industries.

AI has already demonstrated it can raise efficiency and growth in a variety of industries, such as manufacturing, retail, and e-commerce, among others, by improving the decision-making process by analysing vast volumes of data. Artificial intelligence has the potential to enhance the integrated care approach necessary for the management of chronic diseases. Recent progress in image processing, deep learning, neural networks, and natural language processing (NLP) has opened new, previously inconceivable possibilities.

Recent research conducted by scientists at Google resulted in the development of a novel AI algorithm with the ability to forecast the presence of heart disease in patients by examining scans of their retinas. The software used by the company is capable of making reliable inferences about data, such as an individual's age, blood pressure and smoking status, among other things. This can be used to predict the patient's risk of suffering a significant cardiac event, such as a heart attack, with approximately the same level of accuracy as the approaches that are now considered to be the gold standard. This shows there are many possibilities and we have only scratched the surface of what can be done with AI.

AI's role in treating chronic diseases

In 2017, Paschalidis and his team in Australia applied machine learning to electronic health records (EHRs) to predict hospitalizations due to diabetes and heart disease. Their findings were reported in the Harvard Business Review. An estimated 82% accuracy was achieved in predicting hospitalizations one year in advance using this strategy. Now, Paschalidis and his team will utilise EHRs and real-time health data, which includes information from wearables, implanted devices, and home-based networked diagnostic equipment, to generate even more in-depth predictive capabilities.

This demonstrates the feasibility of creating machine learning algorithms to predict which patients may develop cardiovascular disease or diabetes. These algorithms can be used by healthcare practitioners to identify high-risk patients at an early stage and offer them individualised care.

Not only can AI be immensely useful across the board in Chronic Disease Management (CDM), from diagnosis to treatment to management, but it can also help prevent diseases before they become chronic.

The first step in the CDM process is making a diagnosis, which might be prompted by the patient's reporting symptoms or by routine checkups. After that, a doctor will formulate an approach to treating the patient. After a strategy has been established, patients should stick to it and often check in on its efficacy. When the clinician conducts follow-up evaluations, they can decide whether or not to make any necessary changes to the plan.

Diagnosis

According to recent studies, AI is just as capable of making medical diagnoses as people are by making use of patient demographics and health history along with medical imaging, ECG data, and specific illness detection.  Utilising past medical data, image analysis, and ML models can identify patterns to identify diseases. For these machine learning (ML) models to learn about similar groups of subjects, associations between subject features and outcomes of interest, and other things, they need to be "trained" using data generated from clinical activities like screening, diagnosis, treatment assignment and so forth.

There are a variety of AI algorithms that have been employed for disease diagnosis, as detailed in an article by Jiang on the National Institute of Healthcare. Jiang et al. discuss a variety of methods, including support vector machines, neural networks, and decision trees.

Even though more research needs to be done to back up the research, this is a big step forward in being able to quickly and cheaply diagnose chronic diseases on a large scale.

Treatment

When it comes to cancer treatment, the NIH reports that 99 percent of suggestions from IBM’s Watson align with what doctors would choose.

The technology can help doctors by modelling medications and suggesting treatments using artificial intelligence and machine learning algorithms trained on healthcare data.  Practice guidelines, meta-analyses, patient characteristics, and clinical trial data are analysed by these models to recommend doses and treatment approaches.

Artificial intelligence can also aid in the development of individualised treatment plans for patients, rather than a blanket approach based on illness stage rather than particular symptoms. Costs can be reduced and the quality of care can be increased because of tailored treatment that allows clinicians to step in before a patient's condition worsens.

Management

In the event of chronic disease, care management is typically a lifelong process that the patient manages. A virtual nurse assistant powered by artificial intelligence (AI) can support patients in ensuring medicine adherence and monitoring by recording, reminding, and connecting to biometric devices to collect crucial data.

Sensely, a medtech startup, recently debuted AI-powered nurse avatars. These avatars, available in over 30 languages, cover a wide range of content modules, including system assessment, health information, wellness, and chronic care. Patients receive a reminder each morning to complete a check-in process, with Sensely's avatar "Molly" instructing them to record weight and blood pressure data. Sensely then produces a risk assessment for each patient and sends it to clinicians in real time, prompting the necessary actions.

These types of virtual assistants are extremely useful for patients who are self-managing their chronic care.

Measuring and controlling chronic pain is another key issue at this stage. Patients may have chronic pain that is debilitating and difficult to diagnose clinically. AI and facial recognition algorithms can help detect and monitor chronic pain in patients by monitoring facial muscle movements in people who can't self-report discomfort and providing a quantifiable score of how much pain they're feeling. This information is crucial for clinicians to make treatment plan adjustments.

Prevention

Aside from assisting patients who have been diagnosed with chronic diseases, AI can detect disease indicators in people before they become chronic. AI-powered machine learning models can assist doctors in identifying patients at risk of heart disease, hypertension, and pre-diabetes allowing for early intervention and preventative treatment methods.

According to a new Google study, a retinal examination using artificial intelligence may reveal a person's risk of a heart attack or stroke in the next 5 years.

AI can also help prevent hospitalizations for existing chronic patients who are undergoing therapy. Continuous monitoring of patient vitals and treatment adherence by an AI-based system can detect deteriorating situations before they require hospitalisation.

Conclusion

AI, in conjunction with sensor technology, has the potential to unleash the potential of data, offering actionable insights to guide clinical decisions to diagnose, treat, and manage chronic illnesses, thereby providing patients with inexpensive and scalable care at the correct time.

We have only scratched the surface of the technology’s potential applications in AI and chronic disease management. Worldwide, countries and health-care organisations are making progress in harnessing health information technology to improve outcomes and access, setting the framework for the future of healthcare.

Our most recent report, Artificial Intelligence and Aging: Machine Learning for Human Health and Longevity, aims to answer fundamental questions about AI and aging, lay out the next concrete steps for adoption, and make policy suggestions to the UK Government. Finally, it brings together a series of carefully selected brief case studies from around the world to:

• Identify successful uses of AI and ageing in advancing public health goals, with a focus on achievements that would not have been possible with older technologies.
• Investigate alternative methods for tackling common problems and develop applicable conclusions.

 Download our free report to learn more: Artificial intelligence and ageing