This case study was developed in partnership with Jahiz, an upskilling initiative equipping UAE federal government employees with future-ready skills.

The potential of predictive AI insights to revolutionise healthcare delivery is exciting, offering the promise of more personalised, effective and cost-efficient care. AI's capabilities in disease prediction and diagnoses, like creating individualised treatment plans, is just one of its many transformative possibilities. Imagine a future where patients with chronic conditions are monitored in real-time, allowing for timely interventions and improved quality of life, and where the chances of a patient requiring hospitalisation for a preventable disease decrease substantially.

This case study delves into a project by Dubai Health in the United Arab Emirates (UAE), which explored the real-world application of these cutting-edge technologies in diabetes management. It focuses on one specific intervention and extracts key insights that public health departments worldwide can apply to similar projects.

Why is it important to find better solutions to diabetes treatment?

Dubai Health was led to pursue this project because type 2 diabetes is an escalating health concern in the Middle East and North Africa (MENA) region, with the UAE reporting some of the highest prevalence rates. However, the impact of diabetes on public health systems worldwide is not to be underestimated, with Type 2 diabetes being one of the “leading causes of death and disability worldwide”. It is estimated that, by 2030, “the global prevalence of diabetes is projected to reach 643 million”, and the cost of Type 2 diabetes is “predicted to rise to more than $1054 billion by 2045”. The costs of inaction are growing.

For patients themselves, treatment is often challenging and inconsistent. Problems like insufficient home blood glucose monitoring and less patients’ awareness about their chronic disease criticality, further exacerbate the disease, resulting in avoidable hospitalisations, for example.

For doctors, a key challenge is the high rate of patients not coming to their follow-up appointments. This results in poorer health outcomes and a greater risk of complications, as the lack of continuous care often leads to the need for more intensive and serious interventions down the line, taking a toll on already strained health systems.

How could remote patient monitoring revolutionise diabetes management and patient outcomes?

woman reading

Remote patient monitoring, when applied to diabetes management and combined with predictive analytics, could fundamentally transform how the disease is managed. By leveraging these technologies, healthcare providers can implement an early warning system that continuously tracks a patient's vital signs, such as blood glucose levels. This system would enable real-time monitoring and timely interventions. It also means that, at the point of diagnosis, patients are equipped with a health monitoring system which helps them manage their illness. In the long term, this means substantially less time spent in the hospital.

For example, if a patient's blood glucose levels start to rise, the system automatically notifies the patient, alerting them to take corrective action before their condition worsens.

This predictive approach helps prevent complications and empowers patients to manage their health more effectively on a day-to-day basis, ultimately leading to improved long-term outcomes and reducing the need for more serious interventions later on.

Artificial Intelligence interprets the data from the monitoring device and sends early warning signals to doctors and patients. These automated alerts can effectively notify healthcare providers and patients when an intervention is needed, ensuring timely and proactive care.

How was the Dubai Health pilot structured?

A group of 38 patients was equipped with Bluetooth-enabled glucose monitoring devices linked to a customised mobile app. The glucometer readings were automatically transmitted and evaluated by AI-based software. For each patient, customized normal, borderline, and abnormal ranges were established. Normal readings required no feedback, helping to prevent alarm fatigue. If the readings fell within the borderline range, the AI software automatically sent a response to the patient, suggesting corrective actions to prevent further issues. For clearly abnormal levels, the AI software alerted a nurse, who then contacted the patient to recommend corrective actions based on the doctor’s advice, if necessary. Throughout the monitoring process, patients received continuous support through reminders and personalized advice tailored to the data from their devices. All 38 patients attended a follow-up visit at the three-month mark to assess any changes in their condition and overall health.

The use of predictive AI within this pilot meant that each patient received medical advice far quicker than would be possible were an appointment required to make each medical decision, and management team was able to prioritize its time and resources on the patients most in need of personalized, dedicated medical care.

__How did the pilot go? __

Overall, the 38 patients experienced significant health improvements. They readily engaged with the telemonitoring system, leading to notable enhancements in key diabetes markers, including reductions in weight, blood pressure and overall blood glucose levels. These positive outcomes highlight the potential of telemonitoring to genuinely enhance health outcomes for patients with diabetes.

At scale, the benefits could be transformative. The use of predictive AI insights in the pilot meant that each patient received personalised medical advice much quicker than would be possible were an appointment required. Because they received care and advice at home immediately after their diagnosis, they were not lost to follow-up. Doctors themselves were able to prioritise their time and resources on the patients most in need of personalised, dedicated medical care. Ultimately, it means fewer hospital beds being used for cases that are preventable.

__So, what’s next for the Dubai Health Team? __

First and foremost, this wasn’t the only pilot study conducted. The Dubai Health Team also explored the use of remote patient monitoring to reduce hospital readmission rates in patients with congestive heart failure and to lower the incidence of complications in prenatal care. Additionally, Dubai Health is now looking forward to collaborating with other industry leaders to develop customised predictive models tailored to individual patients' vitals and symptoms. AI is also being leveraged to establish smart clinics and assist in the diagnostic process for many diseases, from glaucoma to lung cancer, further enhancing the precision and effectiveness of healthcare delivery.

What key considerations should public servants or policymakers understand before implementing a similar project in their country?

This case study underscores the potential of AI-enhanced telemonitoring as a scalable solution to address public health challenges. It offers valuable insights for public servants and healthcare providers globally who aim to improve chronic disease management through innovative technology.

However, the success of such initiatives relies heavily on real-time data collection, which necessitates robust patient data governance models to ensure privacy and security.

Significant upfront investment in data safety and protection is essential for these projects to succeed, but the long-term benefits make it a worthwhile endeavour.

It's also important to note that all the interventions discussed in this case study are AI-enabled, designed to enhance the efficiency and effectiveness of doctors' work without replacing their expertise. AI is a powerful tool that supports, rather than displaces, the critical role of healthcare professionals.

This case study is based on the following journal article and conversations with the Dubai Health Team.