Assessing user retention in a longitudinal digital mental health data collection study to develop vocal biomarkers

Research output: Contribution to conferenceAbstractpeer-review


This abstract reports findings on user engagement from a study carried out to collect digital speech and scale data that could be used to develop novel voice biomarkers. The data is collected using an app-based tool that is used in the community. The study population were young people (16-24) recruited from the community using schools, mental health charities, Ulster University, and social media to reach the general public. The study design was longitudinal, lasting 12 weeks. Over 500 young people were recruited, with around 400 users completing the first mental health scale assessment. The young people completed three mental health scales,
alongside two voice assessments. The participants completed the Patient Healthcare Questionnaire-9 (PHQ-9), which assesses major depressive disorder (MDD), the Generalized Anxiety Disorder-7 (GAD-7) which assesses anxiety and the DSM-5 cross cutting symptom measure, which assesses a broad spectrum of mental health
conditions. The DSM-5 scale was of particular interest because it assesses a broad spectrum of mental health conditions and somatic symptoms including pain, memory, and substance misuse. For the voice assessment, the participants were asked to read Aesop's fable “The North Wind and the Sun” (which takes 1 minute). This text is commonly used in voice assessments both in voice biomarker studies and in the field of linguistics due to its range of phonemes. A second question collected 40 seconds of spontaneous voice data using the prompt “How has your day been?”
At the outset, participants provided mental health information including any diagnosed mental disorders, whether they consider themselves to have good mental health or not, and any use of relevant medications. Demographic data including age group, education or employment status, and sex and gender identity was also
collected.Initial engagement and adherence based on factors such as diagnosed disorders, mental health scale scores,age group, and education / employment status were analysed. These results have implications for app-based
assessments of mental health and studies using accessible app-based assessments in the community.
Original languageEnglish
Publication statusPublished (in print/issue) - 4 Jul 2023
Event International Digital Mental Health & Wellbeing Conference - Ulster University, Belfast, United Kingdom
Duration: 21 Jun 202323 Jun 2023
Conference number: 1


Conference International Digital Mental Health & Wellbeing Conference
Country/TerritoryUnited Kingdom
Internet address


  • Machine Learning
  • Voice Biomarkers
  • mental health


Dive into the research topics of 'Assessing user retention in a longitudinal digital mental health data collection study to develop vocal biomarkers'. Together they form a unique fingerprint.

Cite this