Insights into population health research
3 January 2020
Richard Martin, Professor of Clinical Epidemiology at the University of Bristol, led the Cancer Research UK (CRUK)-funded Cluster Randomised Trial of PSA Testing for Prostate Cancer (CAP) trial. The CAP trial attracted national recognition for providing definitive answers about the value of Prostate-Specific Antigen (PSA) testing in prostate cancer screening. Richard spoke to Catherine Cowell, Research Funding Manager at CRUK, about the impact of the study and the advice he’d give researchers seeking funding for population research.
Can you briefly describe the funded project and the set-up and rationale of the trial?
We set up the CAP trial to examine the effectiveness of population-based screening and treatment for prostate cancer. Our first step was to undertake a systematic review of the existing evidence on screening and treatment of prostate cancer. The review revealed uncertainty about the effectiveness of screening, particularly the balance between the potential benefits of screening in reducing prostate cancer-specific mortality and its potential harms resulting from overdetection and subsequent overtreatment. Overdetection occurs when cancers are detected by screening that would never affect a person in their lifetime. Overtreatment happens when people then receive treatment that won’t improve their health (as their cancer is so slow-growing, it could be left untreated without causing them harm) but may cause unpleasant side effects.
The review showed that research was vitally needed to address the controversy around both prostate cancer screening and treatment of localised prostate cancer. We knew such research had the potential to change clinical practice, so we set up CAP, the screening trial, and, nested within that, the ProtecT trial of treatment for prostate cancer detected through screening.
What are the key findings of the study so far?
We followed over 400,000 men from nearly 600 general practices across the UK for an average of 10 years. The practices were randomised to either invite all eligible men aged 50–69 years in their practice for a single PSA screening test, or to a control arm in which the men were provided standard National Health Service (NHS) management. The number of practices in each arm of the trial was the same. There was some PSA testing in the control arm, as patients in the NHS can request a PSA test, but less than 6% of men in the control arm had one.
Our primary outcome measure was prostate cancer mortality after 10 years of follow-up, but we also examined the difference in prostate cancer incidence, and the difference in all-cause mortality between the screened and unscreened groups. In the first 18 months of screening, 5 times as many prostate cancers were detected in the screened arm than in the control arm. Although the difference in prostate cancer detection between the two study arms was quite large, prostate cancer mortality and all-cause mortality were the same in both groups after 10 years.
We showed that the PSA blood test is a blunt tool for screening, as it was overdetecting low-risk prostate cancers, but also missing what proved to be fatal prostate cancers. It’s important to note we aren’t saying that screening for prostate cancer isn’t effective; we’re saying that screening for prostate cancer using the PSA test doesn’t seem to produce much benefit after 10 years of follow-up.
Your team was awarded an Office of National Statistics (ONS) Research Excellence Award for the study. What do you think made the research stand out?
We were one of 10 teams shortlisted for the ONS Research Excellence Award in 2018. All 10 teams were invited to give presentations at an awards ceremony in London, and our presentation won. I think we won because of the impact of the results on prostate cancer screening policy, but also because of how innovative the trial was when it was set up in2003. It was an extremely large trial and we were able to do long-term follow-up with low rates of loss to follow-up (less than 1%) because of the electronic data linkage we could do through the NHS. This sort of trial would be difficult to do in countries without a unified national health system. We were able to link all the participants’ NHS numbers to Hospital Episode Statistics (HES) data, cancer registration data and mortality data, and those data have allowed us to assess our primary and secondary outcomes of mortality and cancer detection.
What other outputs has this research had?
We were actively engaged with policymakers prior to publication and continue to work with them, including the UK’s National Screening Committee and the US Preventative Services Taskforce. We talked to CRUK, the National Institute for Health Research and the Royal College of General Practitioners, who all issued communications, including blog posts and press statements, about the results. We also made a 3-minute video explaining the results of the trial, which continues to be invaluable as a quick summary of the trial and its results. Even if your research doesn’t have such direct policy implications, it can still be worth putting something on a webpage or blog about what your results mean in lay terms. If you receive CRUK funding I think you have a duty to disseminate your results widely and explain them to the public, not just to clinicians or in scientific publications.
What do you think are the most exciting recent developments in prostate cancer screening?
There have been exciting developments in the last couple of years that mean we will be able to consider options for prostate cancer screening other than PSA testing. Promising advances in the use of multiparametric magnetic resonance imaging (MRI) to select people with aggressive prostate cancer who need a biopsy could reduce some of the overdetection and harms that can result from the PSA test. However, we wouldn’t be able to image all men aged 50–69 years, because this would be far too expensive for the NHS, so we will need tests to triage men into having MRI scanning or not.
The genetic basis of prostate cancer is also becoming increasingly well-defined. In fact, many of the samples from the men in our trial have contributed to understanding which germline genetic variants are associated with increased risk of prostate cancer. Consequently, genetics may help in identifying men who are at greater risk of prostate cancer and targeting them for MRI imaging. Other tests, such as Stockholm3, that combine protein and genetic analyses have shown promise in identifying men who need further imaging and biopsy.
In terms of treatment, active surveillance compared very favourably with radical treatment (surgery or radiotherapy) in our ProtecT trial, the treatment trial nested within the CAP study, particularly in men with low risk disease. The increasing use of active surveillance in low-risk disease would reduce the overtreatment of less aggressive prostate cancer.
What is the future of the CAP trial?
CRUK have funded our study for a further 5 years to see if any difference in prostate-specific cancer mortality emerges after a period of 15 years. Long-term follow-up is important; given what we know about the natural history of prostate cancer, we may not see any difference in prostate cancer mortality for many years.
We are supplementing this activity with modelling studies, using data from CAP and ProtecT. The idea is to estimate what the impact of MRI scanning, active surveillance, genetic risk stratification and other strategies would be on patient outcomes and NHS resources. There is a lot of work going on in the area – both new empirical studies and continued follow-up on the CAP trial.
Advice for funding applications
Applications require a lot of time and effort to put together. You need to explain the objectives clearly and describe your methods in a careful and robust way. For example, people often forget to include sample size calculations and it’s very important to get those right. The committee won’t be convinced if you haven’t provided the information, or if it’s wrong. Funders are increasingly interested in understanding the potential impact of any findings, so you should make that very clear. Working with a patient and public involvement group can be invaluable in prioritising and framing research questions that are important to patients and will have clinical or public health impact.
You should also consider data sharing when putting together your application. Data sharing is important to make inroads into cancer prevention and treatment – without it we’re not making maximum use of the funding and data the research generates. However, sharing data isn’t easy and requires funding. You need the right software, hardware and people to prepare the data, which might include asking for funding for trial coordinators, data managers, software experts or bioinformaticians. Even though they are extra costs in the application, they really add value to the research.
Finally, make sure your proposed research is linked to CRUK’s research strategy. The strategy and other guidance such as the principles of the James Lind Alliance priority setting partnerships can be useful resources for thinking about the types of high-priority research questions that need answering.