Future of AI in Healthcare (Part 2): Preventing Diabetes

Future of AI in Healthcare

Note: This article is the second part of a series (see Part 1), in which Savvycom Team will discuss future of AI in healthcare – battling the world deadliest diseases.

According to Wikipedia, the first clinical description of this illness was noted down by Aretaeus during the 1st century CE. At that time, Diabetes was only described as:

A disease that caused an excessive amount of sweet urine which passed through the kidneys.

Not much was known about this illness. In fact, diabetes was quite rare. No one would have imagen that one day, diabetes will become the biggest epidemic in human history – affecting 415 million people worldwide by 2018.

Standing at number 4 In WHO’s TOP 10 deadliest disease worldwide, Diabetes is considered a progressive disease. As stated by the US National Library of Medicine, premature death caused by diabetes results in about 12 to 14 years of life lost. The patients and their family also incur medical costs that are 2 to 5 times higher than those without Diabetes. The annual direct health care costs of diabetes worldwide, for people in the 20–79 age groups, are estimated to be as much as 286 billion.

The world seems daunting after all of those statistics. But, with the rise of Artificial Intelligence, advanced medical protocol seems to be on the horizon.

Is there a future of AI in healthcare for Diabetes patients?

1. Diabetes: A world epidemic

According to WHO, Diabetes is defined as ‘a chronic disease that occurs either when the pancreas does not produce enough insulin or when the body cannot effectively use the insulin it produces’

So, what is insulin?
To answer that, let’s first start with another question: What is glucose?

Glucose is a medical term for the sugar created by the digestive system as it consuming and breaking down foods. This sugar is the fuel for the cells – much like gasoline to cars. Without gasoline, cars cannot run. Without glucose, cells cannot continue living.

However, an average human body contains more than 37.2 trillion cells, spread across a vast area. Therefore, turning food into glucose is not the final step. That glucose still has to get to every single cell in one’s body. And that’s where insulin came in.

Insulin acts as the pump that transfers gasoline to the car’s fuel tank. Since cells have ‘locks’ that are called insulin receptors, insulin fits into these locks like a key. When insulin opens the locks, glucose is allowed to enter the house of cells.

To be more specific, the pancreas is genetically coded to produced different amounts of insulin depending on how much glucose is in the bloodstream. When a person goes about their day, not eating much, it releases just a bit to keep things regulated. But during food consumption and processing, it generates a burst of insulin in response.

Hence, insulin is vital factors to keep glucose at balance level and let one’s body operate at its optimal level. Having too much or too little glucose for a substantial period of time and one’s body starts running into complications – including the 2 types of diabetes.

Remember the ‘lock-key’ metaphor earlier?

Future of AI in Healthcare

Type 1 vs. Type 2 Diabetes | Healthstyle

With type 1, there is no key. The patient’s body makes little to no insulin because the beta cells in the pancreas that make insulin are mistakenly destroyed by the body’s own immune system as it was fighting infection. This type is usually diagnosed in children and young adults.

Type 2 is more common. With this type, there are faults in the key itself. The beta cells in the pancreas produce insulin, but not enough to keep blood sugar levels within a normal range or the body doesn’t respond properly to insulin. Without enough insulin to direct the flows of glucose, the glucose is left in the blood. This is what happens when someone is having “high blood sugar”. 

According to WebMD, early warning symptoms of Diabetes include small incidents such as increased thirst or hunger, frequent urination, unexplained weight loss, fatigue, blurred vision, and headaches. Leave it for a longer time, the patient will be exposed for 3 times more risk of experiencing heart attacks, strokes, nerve damages, infections that might lead to lower limb amputation, blindness, and kidney failure.

As you can see, thanks to science progressions and enormous data of health records, we do have a certain level of knowledge in regard to how the disease progress. However, not much is known about the specific causes of diabetes. Scientists think that while type 1 diabetes is caused by sudden environmental factors (such as virus or infections), type 2 diabetes is created by a more collective group of triggers: obesity, lack of physical activities, genes and family history.

2. Is there a cure?

This is going to sound very similar for every disease mentioned in this series: there is currently no cure for diabetes.

Scientist and doctors in related fields have been trying for decades. With Type 1, clinical attempts focus on either replacing the damaged pancreas with a healthy one (through islet cell or pancreas transplant) or targeting the immune system in an effort to stave further damage to the pancreas. However, these efforts have experienced several shortcomings. Not only that donors are in very short supply, systematic reviews also find transplant results themselves tend to vary significantly. With regard to treatments targeting the immune system, the results remain blunt and non-specific.

With type 2, it has been noticed that the number of patient rises along the global rate of obesity and metabolic syndrome – a cluster of conditions related to blood sugar, excess fat and abnormal cholesterol level. This has fueled an increase in weight loss surgical interventions,.  However, depending on the country and insurance plans, such surgery can be costly. They’re also not risk-free with risks varying greatly depending on the person’s overall health profile and age as well as skill and experience of the surgeon.

In the meantime, tremendous interest lies in the usage of different types of stem cell to regenerate the pancreas. This has been applied for both type 1 and type 2 diabetes in recent years with mixed results and limitations. For example, later stages of diabetes’ patients are not good candidates for stem cell therapy.

3. Future of AI in Healthcare: Diagnosis and Effective Control

There is definitely hope since Diabetes can be controlled through effective medication and a healthier lifestyle. What’s vital here is the patient and doctors’ acknowledgment of the current situation. This means that early diagnosis, non-invasive test, and effective maintenance protocol are the key factors. With that being said, AI’s future in healthcare – particularly in Diabetes can be divided into three main categories:

3.1.Non-invasive early diagnosis:

  • How: According to WHO, although detection is improving, the delay from disease onset to actual diagnosis may exceed to 10 years. Contributed reasons to this issue include the subtleness of early symptoms along with the complicated process of diagnosing – which involves a range of actors following the Finnish Diabetes Risk Score. As this method requires human intervention and expertise, it may be exposed to human errors. 
  • Highlighted projects: According to Reuter, one of the most influential complications of diabetes is diabetic retinopathy (DR) – damages in the eye blood vessels and vision loss.  IDx-DR, a software produced by an Iowa-based company, utilizes AI software to self-assess the eye images taken by a retinal camera. After a series of comparison to a provided database, the software tells the doctor that the patient either has more than mild DR and should be referred to eye-care professionals or is “negative” and should be rescreened in 12 months.

    Future of AI in Healthcare

    Future of AI in Healthcare: IDx-DR example | Intro Wellness

    Result: In a clinical trial, IDx-DR was able to correctly identify the presence of more than mild diabetic retinopathy 87% of the time and identify those who did not have more than mild disease 89% of the time. It has now received the FDA’s authorization to provide screening decision without the need or assistant of a clinical.  

3.2. Non-invasive Glucose Monitoring Systems:

  • How: Once diagnosed, frequent adjustments of the insulin treatment plan are crucial for successfully achieving glucose controls goals. Not only is insulin optimization calculation is a time-consuming process, it also demands constant updating data from a board range of devices – glucose monitoring devices, insulin dose regimens, diet tracking calendar, exercise diary. Thus, traditional physicist only gets to see their patient once every few months. With an AI platform, machine learning algorithms can help automate the process of monitoring blood sugar levels and recommend adjustments in care.
  • Highlighted projects: Founded in 2014, DreaMed Advisor cloud-based analytics platform uses machine learning to recommend optimal insulin dosages to maintain balanced glucose levels. For example, data from diabetes management systems are transmitted to the cloud. The patterns derived from analysis through its event detections and learning algorithm are referenced to provide automated recommendations for insulin dosing and treatment plan – in real time. Doctors can then access the cumulative data from the cloud and learn the patient’s unique habits and needs.

    Future of AI in Healthcare

    Future of AI in Healthcare: DreaMed Advisor | DreaMed Advisor’s Youtube

    Result: The U.S. National Library of Medicine indicates that DreaMed began recruiting participants in December 2016 for an evaluation study in children and adolescents with type 1 diabetes. The result will be released in late 2018. 

3.3. Nutrition Coaching:

  • How: One of the biggest parts in taking control of this life-long illness is the patient’s diet. As one’s body experiencing internal chemical imbalance, that person needs to watch their intake in sugar, fat, protein and carb index. However, there isn’t one specific “diabetes diet”. Doctors need to work closely with their patient to customize their specific meal plan – which is, of course, demand extensive knowledge in the nutrition. With machine learning, AI can help recommend meal options based on the specific diet criteria of the user.  
  • Highlighted projects: Founded in 2014, California-based Suggestic is taking a nutrition-focused approach to helping diabetics manage their health.  The platform is built on an extensive database of over 1 million recipes and 500,000 restaurant menus. This data is used to train algorithms to recognize which food selections complement specific diets. The platform also uses an Adherence Score scale – ranging from green (optimal) to red (least optimal) to determine how well a meal option fits with a user’s diet.

    Future of AI in Healthcare

    Future of AI in Healthcare: Suggestic Interface | Suggestic

    Result: This app, even though only being made available for iOS devices, scores 4.8 stars out of over 150 reviews. They are also in partnership with The Institution for Functional medicine and Health Coach Institute.

_____
On one hand, we have yet to understand the root causes of Diabetes. But on the other hand, we have come up with various protocol to ensure that the patients can still live a long and healthy life through disease-modifying treatments and lifestyle alterations. The most important thing at this point is to extensively understand the stages of diabetes and what implications they may have on the people’s life.

Future of AI in Healthcare is bright. Hence, healthcare institutions need to catch up with this first wave of AI development, not only to remain sustainable and profitable but to ensure that they are doing their parts in the making of needed medical progression. It is not an easy task, and that’s why we are here help.

Striving to be a key advocate for the future of AI in healthcare industry worldwide, Savvycom founded a new AI Lab back in March 2018. Leading by Dr. Long Tran, our team have developed three AI applications (Facial Recognition, Object Identification, and AskFred – an AI Chatbot) and are now in the process of commercialization. These technologies are developed with the hope that it will become the foundation of our future healthcare-related products. For example, AskFred can be used as a personal assistant to Diabetes patient – providing answers to the patient questions with the micro/macro nutrition components in each meal.

If you have any request for further information regarding our service, please send us a quick email here.

 

Tags: , , ,

Future of AI in Healthcare (Part 2): Preventing Diabetes at: August 10th, 2018 by admin