A Wearable AI Feedback Tool for Pediatric OCD

Mental Health Services in the Capital Region, Denmark

Status

Completed

Conditions

Obsessive-Compulsive Disorder

Treatments

Device: wearable biosensor

Behavioral: Treatment as usual (TAU)

Behavioral: Exposure and response prevention (ERP)

Study type

Observational

Funder types

Other

Identifiers

NCT05064527

H-18010607- 79689

Details and patient eligibility

About

To test the feasibility of implementing digitally enhanced psychotherapy and research in a community child and adolescent mental health center including the acceptability of the digital technology to patients, parents and therapists.

To use passively collected physiological data and actively collected clinical and biochemical data from the patient and parents to detect and predict episodes of obsessive-compulsive disorder (OCD) -related episodes in children and accommodating behaviour in parents.

Full description

Background: Psychiatric and specifically mechanistic research have stagnated mainly due to the time, labour and bias inherent in human-based technologies that dominate the field. To advance translational and precision psychiatry, researchers within psychiatry must forge long-term collaborations with researchers and developers within technology.

Objectives: To improve assessment and psychotherapy for youth obsessive-compulsive disorder (OCD) through developing an artificial intelligence tool to support patients, parents and therapists in cognitive behavioural therapy. To give an innovative push in the public sector hospitals and research through integration of wearable sensors and machine learning techniques.

Methods: 10 patients (8-17 years) and one of their parents from a child and adolescent mental health center will be recruited as in the larger TECTO project. To examine whether the algorithms can distinguish between patients and typically developing children, 10 typically developing sex and age matched children and one of their parents or guardians will also be recruited from the catchment area. Passively sensed physiological indicators of stress are used as input to privacy preserving signal processing and machine learning algorithms, which predict OCD-episodes, clinical severity and family accommodation. Oxytocin, as a biomarker for family accommodation, is measured through saliva samples. Signal processing will be used to extract acoustic and physiological features of importance for therapeutic response.

Expected results: Results from the proposed project will be used to develop artificial intelligence (AI) tools that support clinicians, patients and parents, which will be implemented and evaluated in a public-sector hospital. Technology-enhanced therapy can be used in a stepped care model, in which subclinical symptoms are first monitored using passive sensors and then AI interventions are offered, supported by a healthcare professional, and when outpatient care is needed, the AI tool can support patient engagement. The results of this project will also advance research in computational science and psychiatry by testing biomarkers of clinical relevance.

Enrollment

36 patients

Sex

All

Ages

8 to 17 years old

Volunteers

No Healthy Volunteers

Inclusion criteria

OCD (ICD-10 F42) as the primary or secondary diagnosis, verified with a semi-structured psychopathological interview using K-SADS-PL.
CY-BOCS > 7: mild (8-15), moderate (16-23), severe (24-31), extreme (32-40)
A psychiatrist determined that the child is eligible for care within psychiatry for their primary diagnosis.
Patient is age 8 through 17 years (both inclusive).
Signed informed consent.

Exclusion criteria

Participation in other OCD trials.
Comorbid illness that contraindicates trial participation: pervasive developmental disorder not including Asperger's syndrome (ICD-10 F84.0-84.4 + F84.8-84.9)), schizophrenia/paranoid psychosis (ICD-10 F20-25 + F28-29), mania or bipolar disorder (ICD-10 F30 and F31), depressive psychotic disorders (F32.3 + F33.3), substance dependence syndrome (ICD-10 F1x.2).
Intelligence Quotient <70.
Any condition (e.g. allergies, eczema, hypersensitivity due to Asperger's syndrome) that would prevent the child or parent from wearing a wristband biosensor.

Trial design

36 participants in 4 patient groups

Patients

Description:

* OCD (ICD-10 F42) as the primary or secondary diagnosis, verified with a semi-structured psychopathological interview using Kiddie Schedule for Affective Disorders and Schizophrenia (K-SADS-PL). * CY-BOCS \> 7: mild (8-15), moderate (16-23), severe (24-31), extreme (32-40) * A psychiatrist determined that the child is eligible for care within psychiatry for their primary diagnosis. * Patient is age 8 through 17 years (both inclusive).

Treatment:

Behavioral: Exposure and response prevention (ERP)

Behavioral: Treatment as usual (TAU)

Device: wearable biosensor

Controls

Description:

* Ages 8 through 17 years (both inclusive). * Sex and age (+/- 3months) matched to an included patient.

Treatment:

Device: wearable biosensor

Caregivers of Patients

Description:

Parent or guardian of patient with OCD