Improving meta-analysis of complex healthcare interventions: Application to chronic obstructive pulmonary disease

About the Project

Health and social care interventions—like those for quitting smoking or managing weight—are often made up of several parts. For example, a smoking cessation programme might include motivational sessions, behaviour-change strategies, and different delivery formats (online, in-person, group sessions) [1]. These programmes can vary in intensity, setting, and who delivers them.

When researchers and decision-makers evaluate these programmes, they want to know things like:

• Which parts of the programme work best?
• Can we remove parts that don’t help to make the programme cheaper and easier?

Answering these questions can improve future clinical trials and influence healthcare policy, helping clinicians provide better care.

What’s the challenge?

These interventions are usually evaluated in randomised controlled trials (RCTs). Across trials, interventions often have components in common which are given alongside other components which differ across trials. To combine evidence from different trials, researchers use a method called network meta-analysis (NMA), which compares multiple interventions at once. However, treating every unique combination of components as a separate intervention can create a lot of uncertainty.

A better way?

Component network meta-analysis (CNMA) estimates the effect of each individual component [2]. This approach can:

• Reduce uncertainty in results.
• Predict how effective new combinations might be—even if they haven’t been tested yet.
• Help design smarter trials and improve healthcare decisions.

What will this PhD do?

This project will develop new methods for CNMA and apply them to real-world data on integrated care programmes for people with chronic obstructive pulmonary disease (COPD). Specifically, you will:

• Develop CNMA models that account for diminishing returns when adding more components and evaluate their performance alongside existing CNMA models through a simulation study considering a range of scenarios and varying data characteristics.
• Develop an algorithm to check if interactions between components can be estimated from the available data.
• Develop novel interactive visualisations for sharing results with non-specialists.
• Apply these methods to COPD care programmes to answer important clinical questions and explore links to realist synthesis.

Training opportunities:

You’ll learn advanced statistical modelling, data analysis, and programming in R and RShiny. You’ll work with large datasets, develop visual tools, and collaborate with researchers at the University of Glasgow. You’ll also join the Biostatistics Research Group, attend seminars and journal clubs, and present your work at conferences.

Impact and outputs:

Your research will help identify the most effective parts of complex health interventions, guiding better trial design and healthcare policy. The outputs—publications, open-source R packages, and a COPD case study—will support clinical decision-making in areas like smoking cessation, weight management, and diabetes prevention, ultimately improving patient care.

Enquiries

Project Enquiries to suzanne.freeman@le.ac.uk

To apply please refer to

https://le.ac.uk/study/research-degrees/research-subjects/health-sciences