Project

Back to overview

Feasibility of a digital protocol to monitor and predict suicidal ideation

English title Feasibility of a digital protocol to monitor and predict suicidal ideation
Applicant Kleim Birgit
Number 183251
Funding scheme Digital Lives
Research institution Psychologisches Institut Universität Zürich
Institution of higher education University of Zurich - ZH
Main discipline Psychology
Start/End 01.12.2018 - 31.01.2021
Approved amount 223'962.00
Show all

All Disciplines (2)

Discipline
Psychology
Mental Disorders, Psychosomatic Diseases

Keywords (3)

passive mobile sensing; ecological momentary assessment; suicidality

Lay Summary (German)

Lead
Etwa fünf bis zehn Prozent der Bevölkerung entwickeln suizidale Gedanken im Laufe ihres Lebens Suizidgedanken. Wie diese suizidalen Gedanken und suizidale Handlungen vorhergesagt werden können ist bisher relativ unklar und eine genaue Vorhersage ist, auch in vulnerablen Personen, nicht sicher möglich. Die vorliegende Studie untersucht digitale Prädiktoren einer aktuellen und empirisch validierten Theorie der Suizidalität von Thomas Joiner (Joiner, 2009; Van Orden et al., 2011). Prädiktoren aus der THeorie sollen gemeinsam mit weiteren Indizes per smartphone bei vulnerablen Personen, Patienten, die aus der Akutpsychiatrie austreten, erfasst werden. Anschliessend soll mittels machine learning Modellen ein Algorithmus entwickelt werden zur Prädiktion von suizidalen Gedanken und Wiedereintritt in die Klinik.
Lay summary
Etwa fünf bis zehn Prozent der Bevölkerung entwickeln suizidale Gedanken im Laufe ihres Lebens Suizidgedanken. Wie diese suizidalen Gedanken und suizidale Handlungen vorhergesagt werden können ist bisher relativ unklar und eine genaue Vorhersage ist, auch in vulnerablen Personen, nicht sicher möglich. Die vorliegende Studie untersucht digitale Prädiktoren einer aktuellen und empirisch validierten Theorie der Suizidalität von Thomas Joiner (Joiner, 2009; Van Orden et al., 2011). Prädiktoren aus der THeorie sollen gemeinsam mit weiteren Indizes per smartphone bei vulnerablen Personen, Patienten, die aus der Akutpsychiatrie austreten, erfasst werden. Anschliessend soll mittels machine learning Modellen ein Algorithmus entwickelt werden zur Prädiktion von suizidalen Gedanken und Wiedereintritt in die Klinik.
Direct link to Lay Summary Last update: 19.12.2018

Responsible applicant and co-applicants

Employees

Project partner

Collaboration

Group / person Country
Types of collaboration
University of Glasgow Ireland (Europe)
- in-depth/constructive exchanges on approaches, methods or results
Leibnitz Institute Resilience Germany (Europe)
- in-depth/constructive exchanges on approaches, methods or results
New York University United States of America (North America)
- in-depth/constructive exchanges on approaches, methods or results
- Publication

Associated projects

Number Title Start Funding scheme
197471 Time and Ties: Dynamic modelling of temporal patterns in dyadic health behaviour change 01.08.2021 Project funding
169827 Optimising outcomes in psychotherapy for anxiety disorders (OPTIMAX) 01.09.2017 Project funding
196405 Predicting the endorsement of preventive behaviors in the context of the Corona virus pandemic: Examining temporal dynamics and the role of risk communication 01.06.2020 Special Call on Coronaviruses
192438 Developing a Taxonomy of Dyadic Behavior Change Techniques 01.05.2021 Project funding

Abstract

Digital technology now provides an unparalleled opportunity to collect data in individuals’ everyday lives. This includes information derived from passive mobile sensing (e.g., activity, screen lock events, light and sound detectors, wifi connections etc.), as well as assessments of psychological variables (behaviour, emotions, cognitions) through smartphones and other wearables. Such data may help remove barriers to monitoring and identification of significant mental health risk, such as suicidal ideation (SI). The present project aims to use psychological theory to design digital indices and test the possibility to exploit digital technology to predict SI and psychiatric hospital readmission (PHR). We investigate this in one of the most vulnerable populations, psychiatric patients post-discharge from an inpatient psychiatry stay, in collaboration with the Centre for Acute Psychiatric Disorders (CAPD) at the Psychiatric University Hospital. This period is clinically challenging and afflicted with high rates of suicidality, mood deterioration, frequent readmissions and thus also economically costly. During their inpatient stay (current mean duration: 17 days), eligible CAPD patients will be offered to take part in the study and those who agree to participate will have two apps installed on their personal phones. They will be pinged 5 times per day for 4 consecutive weeks following discharge to report on emotion, cognition and behaviour (App 1) and passive sensor data will be continually collected (App 2). It is aimed to consecutively include 150 patients over 12 months to test whether (i) implementation of the digital protocol is feasible, and (ii) SI and PHR outcome can be predicted using passive sensor information and psychological indices. It is planned to construct a predictive algorithm for SI and PHR using various learning algorithms (e.g. random forest, support vector machines, recurrent neural networks, feedforward neural networks). The data will be split into training and test data sets to assess how derived algorithms generalize to new data. The training dataset will also be split into subsets, where we will apply k-fold cross-validation. Digital information collected on the basis of psychological theory may, in the future and based on our results, help reach vulnerable individuals early and provide links to just-in-time, cost-effective interventions or establish prompt mental health service contact. The current effort may thus lead to saving lives as well as to significant economic impact by reducing inpatient treatment and days lost to inability.
-