In today’s fast-paced mental healthcare environment, clinicians are under more pressure than ever. Alongside the demands of patient care, they face growing administrative burdens – especially when it comes to documentation.
Notes from sessions, intake reports, and clinical evaluations often take up valuable time, reducing the hours left for direct patient engagement. Especially since many of these sessions are long, involve multiple participants, and require detailed, accurate reporting. The DAIsy project – with HealthTalk at GGz Oost-Brabant as a pilot study – offers a glimpse into a future where AI enhances care, rather than adds to the workload.
A helpful tool for a clinical administrative challenge
HealthTalk is a speech-to-text tool powered by a large language model (LLM). Conducted as part of the broader DAIsy initiative, which focuses on integrating AI into mental health workflows, this pilot represents a step toward transforming clinical documentation as we know it.
Manual notetaking, report generation, and electronic health record (EHR) maintenance present growing administrative challenges for mental health practitioners. The pilot was designed to test how well the tool could support real-world mental health professionals, particularly those working with eating disorder patients. HealthTalk transforms the conversations into structured summaries, aiming to correctly identify who said what and retain critical details. The goal? To lighten the documentation load on clinicians while improving the quality and consistency of reports.
Building a better future together
Accurate and timely documentation isn’t just an administrative task – it’s a vital part of quality care. When notes are rushed, inconsistent, or incomplete, it can affect treatment outcomes, care coordination, and even patient safety. Reducing unnecessary administrative burdens can give clinicians back something incredibly valuable: time. Time to connect, time to reflect, time to care. By supporting clinicians in creating clear and complete reports faster and easier, tools like HealthTalk support the healthcare providers and help ensure patients get the care they deserve.
HealthTalk was deployed in real clinical practice. That’s what makes its findings so valuable. By collecting data on real-world usage and clinician feedback, the pilot helps bridge the gap between potential and practice.
Numbers of the pilot study
Between December 2024 and the time of reporting, 32 clinicians from departments including eating disorders, emergency services, and neuropsychiatry took part in the pilot. Participation was voluntary, and some clinicians joined simply after hearing about the tool from colleagues. Over 363 conversations were recorded, with an average session lasting about 38 minutes. That’s more than 229 hours of recorded interaction, demonstrating real-life usage.
Clinicians offered rich qualitative feedback on their experiences with HealthTalk. They could see potential in how the tool could reduce documentation time, allowing them to focus more on their patients. For some, the shift away from traditional manual notes was quite a change – one that took time but was ultimately rewarding.
The project team used the valuable feedback as fuel for improvement. New conversation formats were co-developed with users, showing a strong commitment to user-centered design. This pilot is just the beginning and lays the foundation for a larger, more comprehensive study under the DAIsy framework: “Enhancing Clinical Documentation: Evaluating the Effectiveness of Large Language Models in Psychiatry”.
Looking ahead
The HealthTalk pilot will continue through August 2025, with ongoing improvements guided by clinician input. Its promise is clear: a smarter, more supportive approach to clinical documentation that works with clinicians, not instead of them. As the world of mental healthcare becomes more complex, we’ll need more intelligent tools to meet the challenge. But those tools must be tested, refined, and designed with people at the center. With HealthTalk, we’re taking a critical step toward that future. To learn more about the DAIsy initiative and the future of AI in mental health, stay connected for updates.
