Pain and Mental Health
About the project
The aim of this project is to better understand the overlap between pain and mental health. In order to do this, we will be trying to get information about pain from the clinical notes within electronic health records of patients with mental health diagnoses. Access to this information will allow us to answer numerous research questions about pain and mental health.
This project will utilise computer-based methods called Natural Language Processing which will employ some form of machine learning to identify relevant mentions of pain from the clinical text.
Why electronic health records?
Electronic health records (EHR) are an essential data source for information gathered at point of contact with patients.
They are usually composed of structured information (such as charts, tests, forms), and unstructured information (such as free-text in the form of clinical notes, correspondence to GPs and other specialists).
Why clinical notes?
Information such as pain is not generally coded as a disorder, and so might not be as easily available within structured fields of electronic health records. This is especially true with mental health records where most information is recorded in the form of free-text.
What type of information?
Clinical notes are records of face to face interactions with patients and generally contain information such as the reason the patient is seeing the clinician, follow-up notes from previous appointments, description of patient symptoms, preliminary diagnoses, recommendations for medication prescriptions, and so on.
Why pain?
Pain is a major global healthcare problem, with chronic pain affecting one in five adults. An interest in understanding pain and its impact has been growing over the past decade.
Why mental health records?
The WHO conducted a 14-nation study which found that there is an overlap between pain and mental health issues.
Mental health EHRs are therefore a good source to gather infomration about pain and mental health.
What is CRIS?
The Clinical Record Interactive Search (CRIS) database consists of anonymised records from The South London and Maudsley NHS Foundation Trust (SLaM), one of the largest mental health care providers in Western Europe.
CRIS consists of almost 30 million notes and correspondence letters, with an average of 90 documents per patient
CRIS follows a robust, patient-led governance model and has ethical approval for secondary analysis (Oxford C Research Ethics Committee, reference 18/SC/0372).
NHS patients who have opted out of sharing their information are not included within this database.