Substance Misuse Case Finding

Finding service users on mental health caseloads for addictions treatment, with a predictive machine learning model

Our client was a local council, who was commissioning addictions services from a mental health provider. Seeking to increase their caseloads in treatment, they believed there were people currently on MH caseloads that might benefit from addictions treatment.

Real World Health was commissioned to deploy our Substance Misuse Case Finding platform to tackle this challenge.

In partnership with the local council and mental health provider, we built a machine learning model to predict MH Service Users that might be eligible for substance misuse treatment. Our Data Science Platform (DSP) was used to combine multiple datasets, including EHR data from the MH trust and substance misuse caseload data.

Working closely with local clinicians, we defined clinical factors as features to predict eligibility.

Our model, once live, was shown to predict almost a third of patients before they were referred. This suggested significant capability to expedite referrals.

For flagged service users that had their cases reviewed by the Substance Misuse team, almost two-thirds (65.8%) had evidence of substance misuse. Half of these (31.7% of total flagged patients) were then appropriate for the substance misuse caseload.

Contact Us


Case Study Form

By submitting this form, I agree to receive communications from Real World Health in connection with my enquiry. My information will be processed in accordance with Real World Health’s Privacy Policy