First Project for the Winton Centre: Redesigning PREDICT

Jack Redden
Reporter, European Medical Journal
@EMGJackRedden

At the start of November 2016 the Winton Centre for Risk and Evidence Communication was opened within the Department of Pure Mathematics and Mathematical Statistics, University of Cambridge, Cambridge, UK. The centre has been set up as a place for quantitative evidence to be used to explore important issues and to be translated into clear and accessible information that can help with our own decision making. Its first major undertaking is to redevelop the PREDICT website, a valuable tool that provides doctors and their patients with risk estimates of treatment outcomes for early-stage breast cancer.

PREDICT was launched in 2010 and has since received thousands of visitors every month from all over the world. It works by feeding data into the website about the patient, including their age, their tumour size and grade, and whether their cancer is HER2 positive. In response, the website offers a prediction of 5-year and 10-year survival, as well as any additional benefits the patient might receive from drug treatments.

However, while freely available, a drawback of the site is that it is not very user-friendly and certain information can be difficult to access for its visitors. Therefore, design of the website and user-testing is to hold much of the focus for the small team of four at the Winton Centre. “What is really important, is that people can see both the risks and benefits of the treatments,” explained Dr Alexandra Freeman, the Executive Director of the Winton Centre, “and, at the moment, that is not that clear [on the website], so it is not that easy for patients to see what difference the treatments might make.”

Dr Freeman also spoke about the research carried out at the centre to better understand what information patients value once they receive a diagnosis and how these values can change. For example, Dr Freeman explained, at a recent oncology conference it was noted that women with young children may have different opinions about treatment than women without children. For the former group, the importance of longer survival may outweigh that of quality of life. It is learning about these different values that will help guide the project to ensure that the PREDICT tool can provide relevant information that will aid the patient and the clinician in their own decision making.

As part of PREDICT’s redevelopment, the team is also working towards implementing two different ways the website can be accessed and used: one catered more towards patients, and one better suited to clinicians. Sitting in on various multidisciplinary team meetings, they saw how cancer specialists used PREDICT as part of the brief couple of minutes or so that can be spared for each patient case to decide the appropriate treatment. Here they saw how quickly these teams needed PREDICT to work in such a fast-paced scenario and provide the necessary information. Meanwhile, they noticed how the tool, which was originally designed for clinicians, might not be as useful when trying to share this information with patients. Therefore, plenty of work is needed to ensure PREDICT is going to be able to address all users.

In future research, Dr Freeman is also looking to find out more about how online tools such as PREDICT can impact clinician’s behaviour. “For instance, where there isn’t a tool for Hodgkin’s lymphoma,” Dr Freeman said, “the difference between prescription rates for different treatments varies enormously between different centres because they are not using something that gives them a standard result, they are working purely on a clinician’s own training and preferences.” Instead, tools like PREDICT could provide more of an evidence-based approach using large clinical datasets to provide clinicians with guidance towards a more standardised approach.

As well as breast cancer, there are plans to design PREDICT so that it can provide risk estimates for treatment outcomes of other cancers, with plans already underway for Hodgkin’s lymphoma and prostate cancer. “We will be making all our software modules for visualisation open source,” Dr Freeman said, “and we will be developing them as generically as possible so that people in the near future can take them and use them for their own purposes. This tool will be [for] breast cancer but we are building it while keeping in mind that we, or others, may want to develop it for prostate cancer, Hodgkin’s lymphoma, and then possibly other cancers further down the line.” Until then, the relaunch date for PREDICT is planned for the final quarter of 2017.