PhD Studentship: Using supermarket loyalty cards data to study dietary patterns associated with known risk factors for cancer

The project:

Since 1990, the incidence of early-onset cancer has surged by nearly 80%, with high- and middle-income countries like the UK experiencing the most significant increase. Dietary choices have been identified as a primary risk factor associated with this rise. For instance, the rise in ultra-processed food (UPF) consumption, now accounting for over 50% of energy intake in the UK, has been linked to higher risks of mortality, cardiovascular disease, and several cancers. However, previous research has primarily focused on specific food groups, such as fruits, vegetables, sugary snacks, and red meats, relying heavily on self-reported data that introduce potential biases and fail to provide insights into processing ingredients like emulsifiers, nitrates, and nitrites that have been linked to cancer.

To address these gaps, the proposed research aims to leverage shopping history records collected through supermarket loyalty card programs. In the UK, where 85% of families purchase groceries from supermarkets, these records offer a promising alternative for studying dietary patterns and their links to cancer risks. Although shopping history data provide granular, objective, and longitudinal insights into lifestyle behaviours and risk factors, they are not without their biases, including the gap between purchase and consumption.

Despite these challenges, commercially collected transactional records offer an unprecedented opportunity to gain time-series data on dietary habits at the population level. This research project will focus on assessing the utility of shopping history data for cancer research.

Aims and Objectives

This research aims to assess the value of supermarket shopping history data in understanding the relationship between dietary choices and cancer. The specific objectives include:

• Analysing shopping data to identify dietary patterns over time.
• Linking patterns in shopping data with known risk factors for cancer through cross-sectional samples.
• Analysing patterns in the shopping data linked with known risk factors in national datasets of supermarkets
• Addressing potential data biases.
• Providing policy and health recommendations for cancer prevention.
• Disseminating findings among the scientific community and the public.

Methodology

The project will use two types of data: cross-sectional data collection and anonymised population-level data from industry partners (accessed through Smart Data Research UK data services centres). For the cross-sectional data, participants will be requested to respond to surveys about their diet (e.g., via food frequency questionnaire) and donate their shopping data using a data portability tool of a large UK supermarket. The student will analyse shopping data using standard statistical methods (e.g., regression) and machine learning categorisation methods to identify data patterns related to diet. This data will be used to investigate associations between dietary risk factors and cancer outcomes. The project will identify and mitigate potential biases inherent to the data, such as the time gap between purchase and consumption, to ensure the validity of research findings.

How to apply:

Please make an online application for this project at http://www.bris.ac.uk/pg-howtoapply. Select Population Health Sciences (PhD) on the Programme Choice page. You will be prompted to enter details of the studentship in the Funding and Research Details sections of the form.

Candidate requirements:  Well-suited for students with strong quantitative skills from a range of backgrounds, including epidemiology, health psychology, data science, nutrition, computer science, and social sciences.

Funding:

Covers tuition fees (for both home and international students), a competitive stipend, and research costs.

Contacts: Dr Anya Skatova (anya.skatova@bristol.ac.uk)

Advert information