Emulating target trials to estimate the impact of treatments on adverse outcomes related to metabolic syndrome using Electronic Health Records.

Lead Research Organisation: University College London
Department Name: Institute of Health Informatics


This project aims to use existing observational data from primary care databases to implement combined causal inference and machine learning techniques to evaluate the effect of 2 treatments on conditions related to metabolic syndrome.

The impact of bariatric surgery on diverse long-term cardiovascular outcomes: Obesity is an important risk factor for cardiovascular disease (CVD). Bariatric surgery has revolutionised the treatment of obesity, enabling sustained weight loss and frequently curing secondary conditions such as type 2 diabetes. To date, RCTs, have focused on the impact of bariatric surgery on Coronary Heart Disease or CVD, but have not considered the potential impacts on pathologically diverse CVD, including atrial fibrillation or Peripheral Vascular Disease. Furthermore, these studies included a limited number of participants, used different methods and had different periods of follow-up, so meta-analyses, found large amounts of heterogeneity. Moreover, the hazard ratio as a single measure, may not reflect the time- varying effect of this intervention . This project will aim to estimate the impact of bariatric surgery on coronary, cardiac, cerebral, and peripheral CVD outcomes.

Estimating the effect of ACEi and ARBs drugs on prevention of liver cirrhosis and liver cancer in patients with non-alcoholic fatty liver disease (NAFLD): While animal studies indicate that administering ACEi and ARBs reduces the incidence of liver cirrhosis and liver cancer through anti-fibrotic effects, observational studies report either the opposite effect, or no effect . This project will aim to estimate the impact of long-term prescriptions of ACEi and ARBs drugs on prevention of liver cirrhosis and liver cancer in patients with NAFLD.

This project fits into the CDT theme, AI enabled diagnostics, as it provides a framework for obtaining rapid and generalisable answers to complex clinical questions which can't be easily answered by RCTs. Research on personalised treatments relies predominantly on associations between risk factors and outcomes. However, ensuring that precision medicine research leads to trustable interventions requires both an understanding of the potential impacts of treatment on outcomes, and the ability to predict outcomes based on risk factors. Machine learning approaches have focused on the latter component, while causal inference attempts to answer questions related to the first component [8].
The project aims to implement methods blending causal inference and machine learning and compare the results with conclusions derived from traditional epidemiological studies, and causal inference studies which don't involve machine learning.

The project is part of the UCL Institute of Health Informatics (IHI) informatics consult project, which aims to provide clinicians with decision support by displaying near real time outcomes of hypothetical interventions, to enable them to make informed decisions about patient care.

The PhD will involve developing a framework to allow these methods to be generalised to answer causal clinical questions generated from the informatics consult project and data.


10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/S021612/1 01/04/2019 30/09/2027
2422508 Studentship EP/S021612/1 29/09/2020 30/09/2024 Aasiyah Rashan