By Melinda Rees, Chief Executive Officer, Psyomics
Today, mental health assessments rely heavily on clinician judgment and self-reported symptoms. While these remain essential, advances in neuroscience, genomics, and digital health now offer the potential to transform how we identify and manage complex conditions such as bipolar disorder – improving both the accuracy and speed of diagnosis.
At Psyomics, we believe it’s time to redefine what’s possible. Our mission is to lead in a new era of precision mental health, integrating biological insights with digital automation to enhance understanding, detection, and management of mental health and neurodivergent conditions.
Through our latest research, we’re beginning to show how this vision can become reality and what the future of truly personalised mental health care might look like.
Bipolar disorder (BD) affects around 40 million people worldwide and is among the most frequently misdiagnosed mental health conditions – often mistaken for major depressive disorder (MDD). Up to 40% of individuals are initially misclassified, leading to delayed or inappropriate treatment and significant personal, clinical, and economic consequences, including increased suicide risk, premature mortality, and greater healthcare utilisation.
Current diagnostic methods rely heavily on self-report, clinician judgment, and retrospective recall, all of which are prone to bias and symptom overlap across conditions. As a result, existing tools show variable accuracy, leaving a pressing need for objective, scalable diagnostic methods.
The Cost of Misdiagnosis
In the UK, people wait an average of almost 10 years for a correct diagnosis. More than a third of undiagnosed individuals attempt suicide, and up to 60% of cases are initially misclassified as major depression.
These figures underscore the profound real-world impact of delayed or inaccurate diagnosis. The core challenge remains that complex biopsychosocial conditions are still categorised largely through behavioural observations and self-reported experiences.
Biology Meets Digital Intelligence
Our flagship research initiative, beseenTM for bipolar, unites two crucial components: structured digital assessment and biological analysis. The digital component is a detailed, clinician-informed questionnaire based on best-practice diagnostic frameworks such as DSM-5 and ICD-11, capturing a full picture of symptoms, history and lived experience.
Participants also provide a simple finger‑prick blood sample, which we analyse using advanced metabolomic profiling. This allows us to identify a distinct biological signature, a panel of metabolites that can differentiate bipolar disorder from major depression. The combination of these two data streams yields a triangulated diagnostic that outperforms either source on its own.
Early Results and Future Potential
The results so far are highly promising:
- Delta Trial (2018–2020): AUROC 0.71 compared with gold-standard diagnostic interviews.
- Combined data: AUROC increased to 0.96 in retrospective analysis.
- 2024 Validation Study: Revised biomarkers achieved AUROC 0.75 alone, and 0.97 with digital tools.
These findings show that integrating biology and digital assessment could significantly improve diagnostic accuracy and speed up access to effective treatment.
Our ambition goes beyond bipolar disorder. We see a future where structured assessments, biological markers and AI-driven analysis work together to guide clinicians, support research and personalise care.
- Clinicians can make quicker, more accurate needs assessments
- Researchers can better stratify patients and design clinical trials
- Patients receive the right treatment faster and enjoy a better quality of life
The gap between what’s possible and what’s standard practice in mental health is widening. Bridging it is not only a scientific challenge – it’s a public health imperative.
At Psyomics, we’re committed to making that vision a reality.