Dr Thomas Cowling
BSc MPH PhD
in Clinical Epidemiology
15-17 Tavistock Place
I am Assistant Professor in Clinical Epidemiology in the Department of Health Services Research and Policy, Faculty of Public Health and Policy.
My research currently focuses on measures of patient morbidity and the prognosis of cancer patients, using national datasets of electronic healthcare records (see 'Research' below). This work is supported by a Medical Research Council (MRC) Skills Development Fellowship (2018-2021).
I joined LSHTM in 2017 from Imperial College London where I completed degrees in Biomedical Sciences (BSc 2011), Public Health (MPH 2012), and Health Services Research (PhD 2016). My PhD was funded by a National Institute for Health Research (NIHR) Doctoral Research Fellowship.
MSc Health Data Science: I am part of the Programme Committee that leads this MSc and am a Module Organiser for the Data Challenge module.
MSc Public Health: I am a personal tutor and have taught on the Health Care Evaluation, Health Services, and Principles and Practice of Public Health modules.
I supervise three PhD students:
- Matthew Parry (Specialist Registrar in Urology) - Radiotherapy treatment strategies for patients with locally advanced prostate cancer: a national study using routinely collected hospital data and patient-reported outcome measures
- Jemma Boyle (Specialist Registrar in General Surgery) - Using national registry data to describe current practices and outcomes in the use of systemic anti-cancer therapy in the management of colorectal cancer in England
- David Wallace (Specialist Registrar in General Surgery) - Liver transplantation as treatment for hepatocellular carcinoma: a national study using linked electronic healthcare records
Current and prospective students are very welcome to contact me.
My research is contextualised by three main trends:
- Patients' health profiles have become more complex, on average, due to population ageing and greater levels of multimorbidity
- In the 'Big Data' era, large datasets of electronic healthcare records are increasingly available for research and linked to other datasets
- Flexible methods for estimating models, from both the statistics and machine learning communities, are receiving greater interest
I am investigating new approaches to modelling patient morbidity in electronic healthcare records, using more of the available data than is used by conventional methods.
This work has included approaches for selecting a small set of diagnosis codes from much larger sets of codes and for modelling many interactions between health conditions using machine learning approaches.
I am applying this methodological work to develop prognostic models for cancer patients, using national cancer registry and hospital records from England. These models have so far focused on colorectal cancer.
The PhD students whom I supervise are studying the factors that predict the treatments received by prostate, colorectal, and liver cancer patients and how outcomes differ between treatment groups.