This blog was written for MQ by Dr. Amy Ronaldson.
May 2025, MQ Datamind Biennial Conference on Data Science It was held at Deutsche Bank’s London headquarters. As MQ researcherI am very excited to participate and receive insights from leading researchers, clinicians, decision makers, and people with experience in mental health challenges.
About me
I am MQ researcher Amy, who uses a large amount of frequently collected health data to understand infection results in patients with severe mental illness. Mental health data science is crucial to my work, making this conference a valuable space to exchange knowledge, share experiences and learn from others on the scene. I will never miss it!
The event presents, panel discussions, and Q&A sessions across career stages and disciplines. I believe there were several key topics at this conference that reflected the direction of travel in mental health data science:
-
The role of artificial intelligence (AI) in mental health data science
As AI develops rapidly, I’m eager to hear how it applies to mental health data science. Dr. Elizabeth Ford outlines how AI is currently used, from management applications to predicting patient needs and predictive modeling to innovations such as AI-driven therapy and medical scribes. While there is hope, there is still significant concern. Mental health data are often very sensitive, recorded in an unstructured format, and may contain unexpected identifiers. This makes data protection and informed consent crucial.
Public attitudes usually seem to support exit models if the data are firmly identified, but nuances in mental health records, such as previous misdiagnosis and changes in diagnostic criteria, present challenges to AI interpretation. Recorded biases, especially those regarding LGBTQ+ individuals, homeless people and gender differences, may exacerbate existing inequalities.
Ford believes that while artificial intelligence can assist clinicians, the final decision should be related to human experts to avoid exacerbating bias or unintended consequences.
-
Insights from early career researchers
For me, early career researcher (ECR) Flash presentations were always the highlight of MQ conferences. Showcase the next generation of talent in mental health data science, and appreciate valuable glimpses of emerging trends and future directions.
A key topic that emerged in the ECR presentation is the recurring challenge of mental health data science in multi-drug and subtle prescription patterns. Flash’s conversation touches on many aspects of this challenge, from evaluating drug interactions (such as metformin and antipsychotic-induced weight gain) to applying large language models to measure patterns in antidepressant treatment. A huge effort to understand the best way to leverage prescription data in mental health data science.
-
Machine Learning and Traditional Epidemiology
Professor Honghan Wu studied how deep learning models perform mental health prediction tasks, combining structured electronic health records with unstructured text. Unstructured texts in health records are a large number of, some untapped resources in the science of mental health data. At the end of the afternoon, discussions about the use of clinical texts in research sparked a lot of debate, especially in machine learning and traditional epidemiology. A key question arises: Can AI beat conventional methods when dealing with complex datasets? The consensus seems to be that machine learning has an advantage in handling large amounts of data, but still requires careful supervision to ensure that important nuances are not missed.
-
Data Sharing and the Future of Mental Health Sciences
Professor Andrew McIntosh gave a compelling speech on the future of collaborative mental health research in the UK. He raised challenges in data coordination and pointed out that increasing data set sizes through collaborative growth has unravelled insights into the genetic basis of mental illness. The discussion highlights the importance of replicability, the appropriate sample size, and the valuable efforts of organizations such as MQ and Datamind to improve data governance.
