Using patient data in research to improve cancer treatments and increase patient survival

What type of data do you need to conduct your research?

“For CanTest, the data will largely be on the symptoms that the patient has experienced, to see why a cancer test may be sensible. This will be supplemented by other important patient factors – age, sex, family and personal history of disease, and genetic predisposition. Then we need to know – in considerable detail – the test that has been offered/performed, and its result. Finally, of course, we will need to know if the patient transpires to have cancer, plus ideally details about the cancer, such as staging. That examines the test performance.

“There is another key aspect which is test acceptability and safety. It’s all very well knowing a test ‘works’. It’s just as important to know: is it acceptable to patients and clinicians? Is it safe? Are the results acted upon? Is it sensible from a health-economic viewpoint? Are there simply too many false-positives? The answers to these questions require a range of data: from qualitative work, to detailed health-economic modelling, to questionnaires/surveys.”

What are the major barriers to patient data access and sharing?

“There seem to be two. The first is simple: that some existing databases – such as the Clinical Practice Research Datalink (CPRD) – won’t allow their databases to be shared. That’s not them being mean – they’re a great helpful resource, but they do need to recoup enough user fees to keep going, and they fear that too liberal data sharing will allow researchers to do studies on the shared database that haven’t been approved.”

When it comes to sharing patient data, there are many legal and ethical issues to consider. There are very strict controls about the sharing of any of the data. No personally-identifiable information can be shared unless patients have explicitly consented to share it or specific permissions are given from a valid research ethics committee. Safeguards are in place to prevent re-identification of individuals, and penalty measures can be used in case of data misuse. Strong safeguards are essential for ensuring patients and the public trust how their data is used. 

Patient recruitment and full participation in research studies also a barrier 

“There is a proportion of patients who are uncomfortable about their personal data being used for research and have not wished us to access their medical records, meaning – at times – it has been difficult to get the maximum value out of their participation. They may still be happy to participate in studies, but will either withhold their consent to their medical records, or will bar the data being shared more widely. This is a completely legitimate stance; entry into research studies must be wholly voluntary.

“Some patients are uncomfortable with their records being released, because they feel that their records are personal, rather than a resource for researchers. Having conversations about how patient data is used, and why, can change these perceptions – but it takes time…”

Why is data sharing important?

A lot of progress in research is due to data sharing. In a previous story, we highlighted how the landscape is changing, with emerging technologies creating new opportunities and new ways to store large scale data or link patient data from different sources. Researchers, healthcare providers, funders and patients are increasingly placing more and more importance on data sharing and are already working to improve the way data are used for research. 

“Data sharing is important because we owe it to the patients: patients have freely given their time and data. We must squeeze the data lemon till every drop of juice comes out.

“It is fair to say that without research using data we would still be ignorant about the true symptoms of cancer, and their true importance. We would still be in the NICE 2005 era, rather than the NICE 2015 era –and we may not have seen the cancer survival improvements in the last decade. Of course, not all the improvements can be attributed to better diagnosis, but some can! So big data has saved lives – for sure!”