With nearly half (46%) of the estimated 415 million people living with diabetes globally left undiagnosed, diabetes remains an immense burden on society. Developing solutions to help sufferers prevent and manage diabetes should, therefore, continue to be a key priority for medical professionals around the world. This should not just be limited to improving medication; educating the public on ways to prevent and manage the disease also remains crucial.

Artificial Intelligence (AI) and machine learning in medicine have rapidly developed over the past several years, and AI-based tools can now be transformative in all aspects of diabetes care, from obtaining improved data for clinicians, to improving patients’ day to day quality of life. One day, we hope, it could even have the power to help reverse Type 2 and gestational diabetes.

AI models support those at risk, patients living with diabetes, and clinicians

AI has the potential to revolutionise how we approach diabetes – starting with individuals at risk of contracting it. For example, there are now machine learning tools which can analyse everything about a pre-diabetic individual’s lifestyle and habits, from physical and mental health to social media activity, to model their level of risk of diabetes and help them reverse any early effects.

For those already suffering from the disease, AI can utilise personal data to help patients to adapt their lifestyle. Helping them to choose appropriate food and drink on a daily basis, and making recommendations on physical exercise, for instance, can all aid people in managing the disease. While some of this information can be made available to patients through traditional methods, the ability of AI to process vast amounts of data means that patients can get tailored recommendations much quicker.

It is also not just patients that can benefit from AI tools for treating diabetes; such tools can significantly improve the accuracy of the data collected by the clinicians working with them. At the Qatar Computing Research Institute (QCRI) – part of Qatar Foundation’s Hamad Bin Khalifa University – this is something we have been working on, to help clinicians improve their support of the local population, where rates of diabetes are high.

We recently developed the System for Integrated Health Analytics (SIHA), a web-based cloud computing platform which allows clinicians to efficiently monitor and analyse patients with diabetes and other chronic diseases, and gain deeper insights using visual and predictive analytics. The platform also enables clinicians to collaborate with researchers for additional insights, ensuring they can stay on top of the latest studies. It can even be used to perform the equivalent of clinical trials on a group of patients, analysing their overall activity, sleep and nutrition data from a single dashboard. Tools like this increase precision and simplify the process for clinicians, allowing them to make more accurate assessments and recommendations.

Helping patients achieve a better quality of life

New AI innovations are, of course, underpinned by the need to improve the lives of those living with diabetes. One of the main challenges that AI can address is the diverse experiences of those living with diabetes. Every patient experiences different circumstances and has a different lifestyle. From the personal data inputted, machine learning tools can ‘learn’ from patients’ variations to make predictions and recommendations that can help them reach their goals.

By shifting the diabetes care model for patients from regular visits to their clinician to inputting data into AI tools, it also allows for a more complete, constant view of each patient’s glycaemic status. Using smart technology, patients can get a continuous feedback loop of data on their blood glucose levels, and learn to organise their day around what works for them. From planning meals to physical activity, they can start managing their diabetes in advance, removing the stress and anxiety of uncertainty.

Understanding how Ramadan impacts diabetes

Continued research into the impacts of AI on diabetes will be integral to the fight for the reversal and cure of Type 2 diabetes. At QCRI, we are currently conducting a large-scale project using AI which models the impact of different lifestyles on diabetes. The team have developed a machine-based algorithm that takes into account clinical and demographic data as well as data on physical activity and glucose variability. This will allow clinicians to predict hyperglycaemic and hypoglycaemic events in diabetic individuals.

As part of this project, we looked at the impact of fasting for diabetic patients to help those who fast for Ramadan to predict and manage glycaemic events. Working in collaboration with physicians from Hamad Medical Corporation (HMC), more than 80 patients with type 2 diabetes received a Fitbit to measure activity and sleep, as well as a wearable device that continuously monitored their glucose readings before and during Ramadan. We used this to develop a tailored AI algorithm which proved to have a high predictive performance for hyperglycaemic excursions.

The success of the trial showcases AI’s potential to save lives by predicting low blood sugar events during fasting, and providing clinicians with key insights into risk factors that cause these events, such as a patient’s treatment regimen. The same AI models are now being generalised to be implemented in a large-scale national trial led by HMC to reverse Type 2 and gestational diabetes.

Advancing research

We and other research institutes around the world are working to improve what can be offered through AI for the fight against diabetes. Our project around Ramadan, for example, raises the prospect of disease management protocols that take into account local cultures and traditions, particularly when it comes to the observance of religious practices. We hope to see more studies in this field.

Alongside this research, a top priority needs to be educating individuals on the causes of diabetes. If we can catch problem areas in diets or lifestyles, we set ourselves up for greater success in both preventing and treating the disease.

Dr. Faisal Farooq, PhD, is the Head of the Center for Digital Health and Precision Medicine within the Qatar Computing Research Institute (QCRI), part of Qatar Foundation’s Hamad Bin Khalifa University. He has more than 15 years of experience in artificial intelligence research in the health sector, managing large teams with multi-million-dollar R&D portfolios. His work has included leading research initiatives with major academic medical centres in the USA, such as the Mt. Sinai Medical Center, Memorial Sloan Kettering and MedCentral Health. He has also worked with large pharmaceutical companies such as Bristol Myers, Pfizer and Merck. Prior to joining QCRI, he led IBM’s 100+ member team delivering AI products to the healthcare market.

 An expert in machine learning, Dr. Farooq’s research focuses have included leveraging AI for drug discovery, and to improve quality of care, predictive analytics, “just in time” intervention and our ability to counter disease progression. He currently holds more than 20 patents and has authored more than 50 scientific articles.