Can your shopping receipt predict ovarian cancer?

Can your shopping receipt predict ovarian cancer? Do shopping habits provide clues about undiagnosed illness?  In 2019 Dr Elizabeth Dolan approached Ovacome with an idea.  Elizabeth is a researcher into Machine Learning, sometimes called AI. She tells us more about the project here:

“I’m researching the use of shopping data, including loyalty card data, for health research on ovarian cancer and respiratory diseases like COVID-19.

I first reached out to Ovacome for input on this project, knowing their good work supporting women with ovarian cancer. Ovacome let me attend events to meet women dealing with ovarian cancer, and supported developing my research proposal, ensuring the crucial patient perspective. Together we co-created an in-depth survey asking women about shopping experiences related to pre-diagnosis symptoms, like purchases aimed at managing bloating, pain, and fatigue.

I also assessed if computational techniques could analyse shopping data to predict ovarian cancer diagnosis. This successful study is now published in the open-access Journal of Medical Internet Research Cancer (JMIR). It revealed changes in diet, clothing, and healthcare purchases, and showed those self medicating on doctors’ advice had longer symptom duration before diagnosis. It also demonstrated potential for using AI (Artificial Intelligence) on this data to unveil symptom-related shopping habits and timing related to diagnosis.

My research centre, N/LAB at the University of Nottingham, investigates ethical uses of big data, behavioural science, and AI to generate health and retail insights for social good. We collaborated with the Cancer Loyalty Card Study (CLOCS) on their data linking ovarian cancer diagnosis to pain and indigestion medication purchases. Applying knowledge from our survey, we examined using machine learning (a type of AI) with this loyalty card data to improve predicting which women had ovarian cancer. Despite difficulties, it indicated potential, especially as sample sizes grow. Interestingly, key products were antacids (trapped wind products), clothes, and vitamins related to fatigue.

More research is still needed, but recent funding will continue this work and public partnerships. With my own family impacted by cancer, I know the importance of diverse perspectives for successful research that can ultimately aid early diagnosis."

You can find out more about this study and about how crucial the views and experiences of members of the Ovacome community have been in moving this project forward here, where Dr Dolan is interviewed by Phaedra Charlton from the World Ovarian cancer Coalition:  https://www.youtube.com/watch?v=Yfh0sJHJcwo 

Did you find this page helpful?

We welcome your feedback. If you have any comments or suggestions, please email ovacome@ovacome.org.uk or call 0207 299 6653.