Improved Diagnostic Accuracy of Psychiatric Disorders using Smartphone Sensor Data, Text and AI

Important results from the project

WordDiagnostics is developing a decision-support for the diagnosis of mental illness, based on artificial intelligence. The goal is to provide more accurate diagnoses early in the care process, which with proper care reduces the damage caused by the condition. The web-based version has been developed with Lund University as a party (no. 2018-02007). The purpose of this application was to develop a prototype of an app that collects and analyzes word data, speech data and records the patient´s movements, which generates data that improves the diagnosis of mental illness.

Expected long term effects

The overall assessment is that the project has proceeded as planned, even though the third part of the project (clinical validation) is somewhat delayed. The reasons for this are mainly that we await a legal report in relation to GDPR, while awaiting the situation for the clinicians to be stabilized in the wake of Corona. The implementation of sensory data is also delayed. The new guidelines from the App Store meant that a previously unidentified obstacle required us to adapt the product.

Approach and implementation

We had high ambitions for this project, which was carried out during a troubled period where the change in relation to COVID-19 forced us all to find new routines. The project design also included that we had access to key competences within a relatively narrow time span and that several processes would run in parallel. The role of coordinator has entailed several lessons, mainly regarding the importance of being flexible and thinking and planning a step ahead". Overall, the project has adapted to the changing circumstance.

External links

www.worddiagnostics.com