Yesterday on the BBC’s Today programme, fronted bravely by Nick Robinson and Tessa Jowell, we heard Susan Short, Professor of Clinical Oncology and Neuro-Oncology (Leeds), speak about the need for combination therapies and treatments . Dr Colin Watts, Clinician Scientist & Hon Consultant Neurosurgeon (Cambridge), explored the impact that 5ALA has had on brain tumour surgery. And then there was discussion about adaptive trial design, so that everyone who is living with a brain tumour can be entered into a trial. And the need for good data and global data collection, and for people to talk across continents. Finally the programme talked about patient experience. Four strong themes!
We know that there truly is progress on all of these fronts – there is a stimulating environment out there. To support you in understanding what has been in the media we are providing a commentary: for the four days that follow the programme we promise to share with you the bigger picture about each of these four themes. Yesterday in part 1 we talked about adaptive trial design. Today the focus is data, data safety, and data quality. Helen Bulbeck, brainstrust’s Director of Services and Policy takes a look at what makes for good data.
Part 2: Brain tumour data
Great strides have been made in recent years in our understanding of what causes cancer and how best to diagnose and treat it. But we can only continue this progress if we have information about the people who are developing cancer. Public Health England collects this information through a process called cancer registration. Cancer registration is the only way we can keep track of how many people are getting cancer and what types of cancer they have. Healthcare teams use this information to continually evaluate and improve services and treatment options for patients. Cancer registration also fuels research into cancer helping scientists investigate possible causes and improve treatments. This information is crucial for progress against cancer and is supported by all the main UK cancer charities and cancer patient groups.
Professor Short talked about the need for data and the intelligence that sits around it.
One of the challenges with brain cancer is that it is a less common cancer (not rarer – there is a difference). This means that it is harder to gather meaningful data in large enough cohorts for anything to be done with it. And of course, if you don’t gather population level data then it really is meaningless if you are trying to solve problems for the whole population. So for example, an audit could be done about how far people travel for the radiotherapy treatment in London. This is not population level data. If you then tried to plan a model of radiotherapy delivery in Scotland, based on data from London, it just wouldn’t work – people have far greater distances to travel in Scotland. This data has its uses (for London) but it’s not meaningful for the whole population. It’s weak because it doesn’t give a complete picture. It’s weak because if you don’t have population level data it introduces bias – you won’t have a representative data set, just parts of a jigsaw, with missing bits.
“Population level data is good data”
Population level data is good data because it is data that is representative of the whole population. By isolating individual sectors of the population for brain cancer from the general population, health care strategies can be formulated specific to this disease area. Large, population-level healthcare databases enable us to discover answers to questions that can’t be found with individual studies or clinical trials alone. By aggregating data from many sources, population studies can help to evaluate health care treatment patterns, examine the quality of care being delivered to patients, identify disparities in treatment and outcomes, and estimate the economic burden of disease on individuals and society. This is the data that is held by Public Health England’s National Cancer Registration and Analysis Service (NCRAS).
Data is important in helping us understand cancer, and the only source of this data is the patient. This means that ultimately all data comes from you. It is your data. Therefore it is of the utmost importance that data is not just good, but also safe.
“It is your data”
Making sure your personal information is private and confidential is very important. Cancer registries take great care to keep the information they hold about you confidential and, as with other medical records, strict ethical and security safeguards are in place and access is strictly controlled.
Cancer registry officials and data analysts are permitted to access and use cancer registry data, to support the provision of health and social care or the promotion of health. When data are released outside the cancer registry there is a strict approval process.
Most releases of data from the cancer registry involve aggregated and anonymous data from which individual identifiers have been removed. Most of the brain tumour statistics presented on this website come from anonymous data provided by the UK’s cancer registries.
Data that identifies an individual may only be released if that person gives their consent, the data is being provided to those involved in their direct care or an independent advisory group grants special legal permission. Data will only be released if the recipient will use it for providing health or social care or for the promotion of health, and they can prove they can keep the data secure.
Public Health England is allowed to collect cancer data by law (though you can opt out of cancer registration at any time). There are other initiatives to collect data, but before agreeing to take part you should always make sure that in giving your data, and your consent, that safety and anonymity is guaranteed. You could ask:
- How good will the data be? Will it be population level data I am contributing to?
- How safe will my data be? Do the checks and safeguards match those put in place by Public Health England?
- What will sharing my personal information further do to build upon what is already being done?
For more information about cancer registration you can read this leaflet from Public Health England
Getting population level data into the public domain
If it’s your data, you should be able to access it. The data landscape is changing thanks to our work with NCRAS and PHE. Statistics influence many of the decisions you might make about your care and also help you understand the bigger picture, or landscape, in which you find yourself. Statistics influence the focus, and work of organisations like brainstrust, helping us understand the scale and impact of the problems we are setting out to solve.
The work we have been doing collaboratively with data since 2015 provides anonymised population level brain tumour data for public use in the form of standard output tables, accessible here: http://cancerdata.nhs.uk/standardoutput.
Since 2015 we’ve been:
- Looking collaboratively (with the research, clincial and patient communities) at Registry data for a less common cancer (brain) in new ways to address important questions for the brain tumour community.
- Establishing a process that can be used by other cancer sites to make data available and understandable.
- Developing partnership working to make data more meaningful.
And the progress on brain tumour data, to date?
- With Public Health England, the NCRAS and members of the brain and CNS tumour site specific clinical reference group, the national brain tumour registry was established (as was a national cancer registry). There is a population level bank of brain tumour data managed safely by NCRAS
- For the first time, in 2017 anonymised population level data on less common cancers has been made available for members of the public and charities to help them understand more about the conditions they interested in
- Emergent themes from this first statistical release have been identified and reports will follow
- This data brings more context to other datasets e.g. we can understand more about the need for early diagnosis.
- The data tells stories that bring weight some of the James Lind Alliance Top Ten Uncertainties and therein priorities for brain tumour research.
The first report from the 2017 data release will be out in the next few weeks. Regards sharing data, the next step is to present timely, good, safe population level data in way that we can all understand. Over the coming months and years we will be working collaboratively to add to the data available and publishing reports, with recommendations, as a result of what we learn from this data.