top of page

Statistical Methods for use in Electronic Health Records

Signal Detection Methods utalising time-to-event methods to detect Adverse Drug Reactions

Team: Odile Sauzet, Victoria Cornelius

In this project we are developing and evaluating the performance of  signal detection tests to raise flags for adverse drug reactions in routinely collected data. The methods proposed utilise the time the event occurs and as a result has significant advantages with regards to performance compared to the frequently implemented disproportionality test approaches. 

Patient Reported Outcome MeasureS using Electronic informed consent and data capture (PROMISE)

Project led by Pragmatic Clinical Trials Unit at Queen Mary University of London

Chief Investiagors: Sandra Eldridge and Richard Hooper

Imperial Team: Victoria Cornelius, Hanaya Raad, Suzie Cro

 

PROMiSE project is a CTU collaboration led by PCTU aiming to deliver proof of concept for a scalable framework for the collection and analysis of patient-reported outcomes (PROMS) as time series, using electronic data capture and incorporating a robust and compliant electronic informed consent form (e-ICF) mechanism. 

Our work at ICTU invovles undertaking a methodological review of the analysis of non-randomised evaluations with multiple baselines., This survey of methods used to analyse data from studies with an observational stepped wedge or multiple baseline design, will allow us to make recommendations on the analysis of future large-scale, high quality, non-randomised healthcare evaluations of this kind.

bottom of page