60% of all people with diabetes calculate their bolus insulin doses wrong and end up in hypo- (life threatening) and hyperglycaemia (harmful to organs). This is mostly because of impairments, laziness, bad user experience or difficult context (hands dirty while cooking). We help people with insulin therapy to increase their quality of life through easier interactions based on voice.

What it does

Sugarcane helps people with T1, T2 and Gestational diabetes to conveniently calculate the correct insulin dose and bolus.

How we built it

Alexa custom skill using various data sources, e.g. REST web services that pull food carbohydrate and live insulin and blood glucose levels. Calculating correct doses according to personalised user data (correction factors, size, gender, weight).

Challenges we ran into

Setting up the Alexa environment was hard and error-prone, also thinking through the various use cases and contexts.

Accomplishments that we’re proud of

Integration of live glucose sensor data, implementation of food database API and correct calculation of correct insulin dose.

What we learned

Implementing different APIs and data sources bringing together the different parts of the puzzle from our teams expertise as engineers, medical doctors, UX design. After initial brainstorming, we focused on the small daily interaction of calculating the insulin dose. This is a big and repeating pain point of diabetes patients, but because it’s done multiple times per day it will have a big impact on the user experience and prospective outcome of insulin therapy.

What’s next for Sugarcane

Implementing a continuous log and personalised dosing, based on learning about user behaviour and personal factors that influence diabetes response. Setting up a safe database and getting user consent according to German medical data law.

Built With

Share this project: