Niamh Doherty is PhD student within the Cancer Epidemiology Research Group at the Centre for Public Health (CPH), QUB. She is supervised by Dr Blánaid Hicks and Dr Chris Cardwell, who have a particular interest in using Real World Evidence to study the safety and effectiveness of treatments and medications in cancer.

Cancer epidemiology is the study of the factors involved in cancer development in specific populations. We utilise large population-based databases to investigate the relationship between common exposures (such as diet, reproductive factors, and use of medications) and specific types of cancers (for example prostate cancer, bowel cancer, and bladder cancer).

This research helps identify targets for health policy and clinical practices that may be useful in the prevention of cancer. Examples of the impact of these kinds of studies are the public health advice for individuals on smoking behaviour or exercise, because the relationship between those factors and cancer risk has been clearly identified, and as such can be targeted for reducing cancer risk in the population.

My PhD aims to investigate the relationship between key common exposures and the risk of kidney and bladder cancers.

Specifically, I try to understand correlations between the use of medication, in particular 5-alpha reductase inhibitors, usually prescribed to treat benign prostate enlargement or male pattern hair loss, sex hormones (oestrogen, testosterone), and inflammatory factors (kidney stones, biomarkers), and identify how those affect cancer incidence.

To achieve this, I have conducted several cohort studies to quantify the risk of these cancers with each exposure, using two large population-based databases of anonymised health data from the UK. This method is summarised in Figure 1.  

Figure 1. A simplified cohort study design in cancer epidemiology of exposed vs unexposed individuals (for example, with and without kidney stones) with the outcome on a particular cancer type.

 

The use of “real world evidence” from large databases allows more robust predictions which will be applicable to a greater and more diverse range of patients, by using a larger sample size, and longer follow-up times compared to standard experimental studies (randomised control trials) that typically have small numbers and restricted populations.

My research aims to investigate causal effects of these exposures on cancer by building statistical models with confounders (factors that can interfere with a direct influence between exposures and cancer outcome) and using other methods to accurately capture the associations in the population.

When Genetic data is combined with population data it allows estimating effects of exposures on cancer risk based on individuals genetic background and how it affects their predisposition, using a method termed Mendelian randomisation - studying the effects of particular exposures on people with specific genetic profiles.

We hope our research will contribute to our knowledge on cancer in the UK population and provide direction for further research in cancer risk with respect to hormonal and inflammatory factors in kidney and bladder cancer development.