JMIR Publications

ABSTRACT

Background:

Child and adolescent mental health services (CAMHS) and child development clinics (CDCs) in the UK are facing unprecedented demand, resulting in prolonged waiting times for assessment of neurodevelopmental conditions, including autism spectrum disorder and attention deficit hyperactivity disorder. Delayed diagnosis and intervention are associated with poorer outcomes, with administrative burden being a key contributor to limited assessment capacity. Emerging digital health technologies like ambient voice technology assisted documentation (AVT-A) can reduce this burden and improve service efficiency. This second phase study builds upon the authors’ previous AVT proof-of-concept study.

Objective:

1. To assess the performance of AVT-A across a broad range of use cases in CAMHS outpatient clinics and a CDC 2. To assess the impact of AVT-A on quantitative and qualitative clinical outcomes, including administrative burden 3. To explore clinician, patient and carer perceptions towards AVT-A in clinical settings

Methods:

A mixed-methods pre-post service development pilot was conducted from December 2024 to March 2025, comparing AVT-A with manual documentation. The study was conducted across three sites: two CAMHS outpatient clinics and one CDC. Clinicians spanning mental health and neurodevelopmental roles were invited to participate, testing 12 different use cases. The primary outcome measure was self-reported time taken to complete administrative tasks per clinical encounter. Secondary outcome measures included qualitative clinician experience and patient/carer perception of AVT-A.

Results:

37 clinicians provided baseline and intervention timesheet data. Most were full-time working mental health nurses, aged 25-34 and female. A total of 1,085 clinical appointments were recorded, with 50% (n=543) using AVT-A. Across all use cases, the mean administration time was 28% less with AVT-A compared to manual documentation (p<0.001). However, satisfaction with documentation accuracy and confidence in documentation quality were higher with manual documentation compared with AVT-A (85% vs 69%, p=0.11, and 81% vs 73%, p=0.46, respectively), with no statistically significant differences observed. No significant differences were found in self-reported efficiency of seeing patients or completing administrative work. Data from patient/carer surveys revealed no significant differences between AVT and manual documentation across any measures.

Conclusions:

In the face of significant demand, AVT-A presents an opportunity to improve service capacity, clinician wellbeing and patient experience. AVT significantly reduces documentation burden in neurodevelopmental assessments and is acceptable for patients and carers. However, challenges including documentation accuracy and quality must be addressed.