HUS Used Speechly to Improve Meal Tracking in Hospitals

HUS is a hospital district for population of over 2 million and budget of over 2.5 Billion, treating over 650.000 patients every year.


Need for more accurate meal tracking in hospitals

Photograph by Marcus O’Leary

HUS wanted to develop a mobile application for the treatment of gestational diabetes, a condition in which a woman without diabetes develops high blood sugar levels during pregnancy. Around 10,000 women in Finland develop this condition every year of which roughly half develop type 2 diabetes later on. The total treatment costs for these type 2 diabetes cases and their complications can reach up to EUR 28 million a year.

The mobile application would measure and store in the cloud the mother’s blood glucose levels, physical activity, nutrition, pulse and daily weight. This data would be forwarded in real time to health care personnel in order to provide guidance and support to the patients.

One of the trickiest pieces of data to obtain was the nutrition data, which would be collected through meal diary entries reported by the patients. This had been done in previous contexts either using paper forms or by typing on a mobile application. Both methods have resulted in patients complaining about the tediousness of the task, ultimately resulting in high churn. Moreover, health care personnel were also spending significant time manually transforming meal diary entries into a digital format for analysis purposes.

To solve these challenges, HUS wanted to find a partner to develop a voice-controlled mobile meal diary solution that would work across all major platforms and browsers.

Voice-enabled meal tracking app

Speechly, together with it’s integrator partner Reaktor, developed a responsive mobile web application leveraging Speechly’s Voice Interface API to enable voice UI features. The web application allows the patients to verbally explain what they have eaten, and the system captures the information and calculates the nutritional values on the fly. The application supports utterances like “two pieces of toast, butter and cheese with a cup of coffee and milk” or “for lunch I had a plate of paghetti Carbonara, a piece of bread and a glass of mineral water”. The Speechly API provided voice support for all browsers across a large variety of mobile devices, which was an important requirement to ensure that all patents could benefit from the app.

faster for users to record meals
data collected
faster for researchers to go through digital meal journals

How would you use Spoken Language Understanding?

Our specialist will contact you as soon as possible to discuss your use case in more detail.