Can #AI help end #diabetes?
Credit: Photo by Towfiqu Barbhiuya on unsplash.com

Can #AI help end #diabetes?

brAIn gAIn: Episode 003

Last Friday (June 24th), one of the world’s most prestigious medical journals, The Lancet, published an editorial entitled – Diabetes: a defining disease of the 21st century that coincided with the American Diabetes Association 83rd Scientific Session. Some the editorial’s jaw-dropping facts include:

  • ?1.31 billion people living with diabetes by 2050 worldwide;
  • Almost 75% of adults with diabetes in 2045 will be living in low-income and middle-income countries; and
  • Environmental factors, lifestyle choices, and the social determinants of health are key drivers of type 2 diabetes – that can be prevented and can be put into remission – which accounts for 90% of global diabetes prevalence.

Back in 2016, the World Health Organization (WHO), issued the first Global Report on Diabetes which stimulated many countries to develop comprehensive frameworks and action plans. The global fight against diabetes has many allies and partners and is, in part, driven by the Global Diabetes Compact.

But over a century after the discovery of insulin in Canada, the global health challenge of diabetes is daunting. This non-communicable disease has reached epidemic proportions, and along with cardiovascular disease and cancer, is responsible for 70% of all worldwide deaths annually.


Diabetes 101

According to Diabetes Canada (yes, pandering to my home team crowd ??), “diabetes is a disease in which your body either can't produce insulin or can't properly use the insulin it produces. Insulin is a hormone produced by your pancreas. Insulin's role is to regulate the amount of glucose (sugar) in the blood. Blood sugar must be carefully regulated to ensure that the body functions properly. Too much blood sugar can cause damage to organs, blood vessels, and nerves. Your body also needs insulin in order to use sugar for energy.”

There are different types of diabetes – type 1, type 2, gestational, and pre-diabetes – and the disease is sometimes seen as a gateway or revolving door into other serious health issues that can occur due to the complications of diabetes. These complications can include cardiovascular conditions (heart disease & stroke) kidney disease, diabetic retinopathy, anxiety, nerve damage, and lower limb amputation. And research into comorbidities associated with diabetes is an active area of medical and scientific inquiry.

Living with diabetes can be an intense experience and achieving “time in ?range” can be difficult for individuals even with access to the latest medications, technologies (insulin pumps and continuous glucose monitors [CGM]), and knowledgeable healthcare providers.

Health outcomes can be severely compromised for those with limited or no access to diabetes care (education medicine, devices, healthcare providers). This is a global issue of #healthequity that is also compounded by diabetes stigma. ?

?

#AIinHealthcare – Advances in Diabetes

The good news in the world of diabetes is that public and private institutions are devoting billions of dollars across the spectrum of diabetes research each year, which includes the deployment of #AI to find better ways to prevent, detect, and treat diabetes and allow people living with diabetes greater control over this disease. This includes …


Prediction

Researchers in the United States are using #MachineLearning to combine the results of food intake surveys and blood tests to predict the risks of cardiometabolic conditions like type 2 diabetes. In India, electrocardiogram (ECG) data was amassed and used to develop a predictive #algorithm that “was able to classify individuals as having?prediabetes,?type 2 diabetes?or no diabetes with higher than 95% accuracy after correcting for potential?confounding factors?such as age, gender and presence of other?metabolic disorders.” And Klick Applied Sciences says it can use just twelve hours of CGM data to determine if someone has prediabetes or diabetes. ?


Management

For people with diabetes, #AIinHealthcare is also improving their lives. For example, from Europe, the development of algorithms that facilitate real-time data – medical history, exercise, and recent meals – analysis and improve “time in range” by connecting an insulin pump with a CGM device for precise dosing. In Illinois, research funding has been allocated to two projects that include prediction of medication non-adherence in people with type 2 diabetes to inform physician strategies to help their patients and “language modeling-based?knowledge graphs?to identify high-risk diabetic patients and prevent?diabetic ketoacidosis (DKA).”

Meanwhile, Seattle-based Know Labs is developing a non-invasive blood glucose monitoring device that uses bio-RFID monitoring to help people manage their diabetes. Yes, we have entered the realm of tricorders from Star Trek where yesterday’s #scifi is becoming today’s science. While we’re on the subject, if you want to go down a tricorder rabbit hole, click here.


Complications

As noted earlier, the complications from diabetes can be extensive. Thankfully, this is an area where #AI is also making life better for people living with diabetes. In Connecticut, the smart folks at Yale University built a #MachineLearning-based tool “to personalize recommendations for pursuing intensive or standard blood pressure treatment goals among individuals with and without diabetes” and “is designed to facilitate?shared decision-making?between providers and patients with hypertension through a?data-driven?approach.”

Moving farther west, a team at the University of Missouri-Cambridge, through an #AI-driven study “discovered notable differences between those with familial and sporadic?type 1 diabetes, with familial cases having many more?comorbidities?than patients with no family history of the condition.” This has implications for personalizing individual patient care plans and disease monitoring as those with type 1 diabetes age.

In addition, a study published in the Journal of the American Heart Association, researchers used #DeepLearning to analyze data from electronic health records to determine the reasons why people with type 1 diabetes weren’t taking statins. This information is helpful to inform patient-level and system-level interventions.

On the vision front, a study out of Thailand found that “a?deep learning?algorithm was able to detect?diabetic retinopathy?in patients with diabetes on par with community specialists.” The use of #AI tools such as these a collaborative tool for clinicians can only benefit patient care.


Next …

The challenges for developers of #AIinHealthcare applications are endless. However, let’s put a few on the table for them, such as:

I’m still not sure if #AI can help end #diabetes, but it can, AND is, making a difference for people who live with diabetes. Thanks for reading this week’s edition of brAIn gAIn. Comments and suggestions are always treated as a gift.

= = = = =


brAIn gAIn 003 -- Resources


brAIn gAIn 003 -- Reading


brAIn gAIn 004 – Reveal

AIding Seniors Care

?

Are you craving daily #AIinHealthcare content?
Follow @AI_4_Healthcare on Twitter.
Be well!


要查看或添加评论,请登录

社区洞察

其他会员也浏览了