Diabetes: How startups are leveraging data driven decision-making for care management

Data driven decision making is improving diabetes care

Data driven decision making is improving diabetes care

Diabetes is perhaps one of the most common ailments we are aware of. Some estimates place the global figures of patients diagnosed with diabetes in excess of 340 million. WHO predicts the disease will be the seventh leading cause of death by 2030. But amidst these alarming numbers, a lesser focused fact is that diabetes is an ailment that can be managed and controlled.

In layman terms, diabetes is classified into two categories, Type 1 and Type 2. Type 1 is rarer, approx. less than 10 percent of diabetics suffer from this ailment. In it, cells producing insulin are attacked by the body’s immune system. Lack of insulin lowers cellular ability to absorb glucose, which is needed for energy production. In Type 2 diabetes, the body only exhibits delayed response in insulin utilization or shows resistance to insulin.

In the recent decades, giant strides have been made in technological advancements in diabetes management. Testing methods, glucose monitoring and insulin pumping systems have improved. As with other industries, researchers have also focused on potential methods to leverage analytics and machine learning algorithms for better diabetes management. In the last few years, a lot of such research has spiralled out into the realms of the consumer market. In today’s post, I have tried to showcase some of these startups that are focusing their energy into a pursuit that is very noble and needs a lot of attention.

Australia based ManageBGL’s launched its PredictBGL mobile app earlier this year. The app collects data from connected fitness wearables, digital insulin pens, insulin pumps, blood sugar meters and continuous glucose monitors to predict the blood sugar levels of users. This is an extremely useful application, particularly in addressing dip in blood sugar levels overnight. The startup is currently being incubated at 500Startups.

TypeZero Technologies
Commercialized out of University of Virginia (UVA), TypeZero is a VA startup that has built a closed loop system proprietary technology called DiA (Diabetes Assistance system). The DiA runs an Android based algorithm that interacts with the insulin pumps and the cloud-based data servers. BigFoot Biomedical and Bionic Pancreas Project (Boston University) are other ventures working on similar concepts.

Siren Care
San Francisco based Siren Care has developed a smart textile based product Invest IVS to address the problem of diabetic foot ulcers which patients aren’t able to detect due to reduction in sensory perceptions. Invest IVS is a sensor embedded sock that tracks temperature (marker for ulcer formation), combined with a smart wearable anklet tracking motion that connects to a smartphone via BLE. Abnormal temperature behavior of the foot, if detected, is alarmed to the user so that she may take an appropriate corrective measure.

Another player in this domain is Calgary (Canada) based Orpyx Medical Technologies that has developed a shoe based detection system called SurroSense Rx.

New York based GlucoIQ helps physicians monitor diabetic patients in real-time by collecting information from wireless glucometers. It also helps physicians identify patients with abnormal readings and conduct proactive outreach. The product dashboard also allows for physicians to communicate with patients.

InSpark Technologies
Charlottesville based InSpark Technologies is building software tools that use proprietary algorithms for diabetes management. Its gamut of solutions include detecting undesirable patterns occurring in upcoming problem areas of the day, patterns of glucose variability, risk of hypoglycaemia, estimation of HbA1c, and insulin sensitivity on monitoring devices or on connected mobile phones.

Suggestic claims to be a lifestyle intervention based Type 2 diabetes reversal program. Focusing on the concept of precision medicine (targeted treatment based on an individual’s gene, environment and lifestyle), it employs machine learning technology to extract and validate causal patterns between each person’s genes, metabolism, food and activities. Based on such calibrations, it develops a recommendation system for each individual for diabetes reversal. Its diabetic nutrition and lifestyle programs is delivered via web, smart phone app, SMS, phone, tele-health and in-person visits through healthcare provider partnerships.

Another startup working in similar domain is San Diego based Tecsomed. Focusing on building solutions for children suffering with diabetes, its upcoming product suites includes wearables tracker as well as parental smartphone app for live tracking.

Trento (Italy) based DiabetesLab is building a mobile app based product to helps physically active people (athletes), who are insulin treated, with their diabetes management. Its tracker tracks physical activity, employs an algorithm to learn the training habit of the user and automates insulin level calculation using data driven approach.

Apart from these, a recent competition organized by Kaggle and California Healthcare Foundation for data scientists to use machine learning for medical diagnosis resulted in a unique development in diabetic retinopathy (a condition that affects approx. 45 percent of people with diabetes). If not treated, the condition can lead to severe vision impairment. Dr Graham, an academic from the University of Warwick, built a neural networks based algorithm that could rival a human eye in identifying signs of retinopathy from an eye scan. Hopefully, we will witness this technology too getting commercialized soon.

Anubhav, a data scientist, writes about new developments and future trends in the machine learning and data analytics domain.
He can be reached at anubhav@thinkbigdata.in
Follow him on Twitter at: https://twitter.com/think_bigdata

Leave a Reply