UCL Institute of Health Informatics


Investigating adverse effects of psychiatric drugs through data-mining of electronic health records

Our Research

The South London and Maudsley NHS Foundation Trust (SLaM), is the largest mental health provider in Europe, spanning over four London boroughs. The SLaM Electronic Health Records (EHR) system, the Electronic Patient Journey System (ePJS), is typical of many such psychiatric hospital systems in that it stores much of its clinical records and prescribing information in an unstructured free text format. Adverse drug reactions are serious, and possibly life-threatening, side effects of medication treatment. Motivated by the abundance of medication and ADR information available in electronic health records, this Ph.D. project aims to investigate the prevalence of Adverse Drug Reactions (ADRs) in EHRs by developing natural language processing (NLP) tools to extract ADR-related knowledge from the free-text clinical notes. The tools to be developed should help stratify patients into different subpopulations on the basis of their diagnosis, medication, and ADRs. Integrating this data with other clinical and laboratory data may help to identify new risk factors or biomarkers for severe ADRs. The project is designed to achieve the following objectives:

  1. We have develop natural language algorithms to identify mentions of adverse drug events (ADEs) and distinguish: a) Positive from Negative mentions of ADEs; b) Current from past occurrences of ADEs by possessing temporal reasoning capabilities; c) General from patient-specific mentions of ADEs in the clinical notes.
  2. We have developed algorithms to create medication timelines from the patient EHRs to determine simple associations between ADEs and medications, as well as inferring causative relationships (ADRs).
  3. Further mode we conducted simple descriptive and inferential statistical analysis in order to stratify the patient population according to medication, diagnosis, ethnicity, age and gender for example.


Discovery Science
Precision Medicine
Public Health
Citizen Driven Health
Learning Health Systems


Dr Richard Dobson

Dr Zina Ibrahim

Dr Honghan Wu

Ehtesham Iqbal 

Daniel Rhodes

Alvin Romero

Robbie Mallah


Professor Robert Stewart; Banke Dzahini


Ehtesham Iqbal, Robbie Mallah, Richard George Jackson, Michael Ball, Zina M. Ibrahim, Matthew Broadbent, Olubanke Dzahini, Robert Stewart, Caroline Johnston, and Richard JB Dobson. Identification of adverse drug events from free text electronic patient records and information in a large mental health case register. PloS one 10, no. 8 (2015): e0134208. doi: http://dx.doi.org/10.1371/journal.pone.0134208

Honghan Wu,  Zina M. Ibrahim, Ehtesham Iqbal and Richard Dobson. Predicting Adverse Events from Multiple and Dynamic Medication Episodes – a preliminary result in a large mental health registry.  Accepted to the Conference of Specialist Group in Artificial Intelligence, 2016.

Ehtesham Iqbal, Robbie Mallah, Daniel Rhodes, Honghan Wu, Alvin Romero, Nynn Chang, Olubanke Dzahini, Chandra Pandey, Matthew Broadbent, Robert Stewart, Richard J. B. Dobson, Zina M. Ibrahim . ADEPt, a semantically-enriched pipeline for extracting adverse drug events from free-text electronic health records. PloS one  2017; 12(11): e0187121. doi: https://doi.org/10.1371/journal.pone.0187121