Gold
Creative Use of Data
Data Collection and Research
| Entrant: | Klick Health, Toronto |
| Brand: | KVI Brave Fund |
| Title: | "Voice 2 Diabetes" |
| Corporate Name of Client: | KVI Brave Fund Inc. |
| Agency: | Klick Health, Toronto |
| Chief Creative Officer: | Rich Levy |
| Executive Creative Directors: | Bernardo Romero/Tim Jones |
| Group Creative Directors: | Amy Fortunato/Andrea Bistany |
| Agency Senior Associate, Creative Producer: | Holly Phipps |
| Agency SVP, Digital Health R&D: | Yan Fossat |
| Agency Lead Scientist: | Jaycee Kaufman |
| Agency Engineering Lead: | Anirudh Thommandram |
| Agency SVP, Communications: | Sheryl Steinberg |
| Agency Director, Communications: | Marisa McWilliams |
| Agency Manager, Communications: | Amanda Fergusonf |
Description:
Distance, cost, and lack of resources prevents millions of people from accessing diagnostic blood tests. However, over a billion people living in India have mobile phones, which now can be used to save their lives.
Voice to Diabetes is an innovative tool that leverages AI to diagnose type 2 diabetes based on the analysis of a person's voice.
Traditionally, Type 2 diabetes diagnosis involves invasive blood work, but this study introduces a non-invasive and efficient alternative.
Voice to Diabetes harnesses AI and machine learning techniques to detect subtle vocal changes that are imperceptible to the human ear.
By incorporating features like pitch and intensity changes, the tool successfully distinguishes between individuals with and without Type 2 diabetes.
The study, published in a Mayo Clinic journal, involved 267 participants in India, with 192 without diabetes and 75 with a prior diagnosis.
Using a smartphone app, participants recorded short voice clips multiple times daily for two weeks. We analyzed 18,465 recordings, identifying 14 different acoustic characteristics that led to the creation of a highly accurate diabetes identification tool.
For women, changes in pitch and deviation from the average pitch were crucial, while for men, the diagnosis relied on variations in voice strength or amplitude.
With an accuracy of up to 89% for women and 86% for men, this innovative approach could pave the way for an accurate, low-cost, and accessible test for diabetes.