Using collective intelligence design to unlock new insights about cancer
We ask - can more be done to exploit the health information commons and empower people as they navigate their own health and experience of cancer?
Using collective intelligence design to unlock new insights about cancer
Every year, nearly 360,000 people are diagnosed with cancer and nearly a quarter (28%) of all deaths are caused by cancer. Despite significant investment in screening, diagnosis and treatment over the past 20 years, five year cancer survival rates in the UK continue to lag behind other European countries.
This is why cancer has again been recognised as a priority in the NHS Long Term Plan, with a renewed focus on improving cancer survival through action on earlier diagnosis. It includes a welcome ambition to increase the proportion of cancers diagnosed at stages 1 and 2 from around half to three-quarters of cancer patients by 2028.
When people receive, or are waiting for, a cancer diagnosis, it can be a lonely and frightening time. The majority turn to the internet for information and advice. They find an abundance of information online - from the NHS, peer support groups and charities, as well as academic and medical research. Some of it is accurate and useful, but much of it is not. The sheer volume can be overwhelming and difficult to navigate. It also does little to address the issue of delayed diagnosis - with evidence showing that despite all this information, there remains a worrying lack of public awareness of many of the symptoms of cancer, especially amongst people under 50.
The National Cancer Registration and Analysis Service (NCRAS) holds one of the richest cancer datasets in the world - with extensive data on every person diagnosed with cancer in England. In recent years, they have made great strides to open up some of this data to researchers, clinicians and even the public, but still much of the data (which is by its nature, personal and sensitive) is not accessible, meaning that the full potential of the data is yet to be realised.
In this blog, we ask - can more be done to exploit all this information in ways that are useful to people as they navigate their own health and experience of cancer?
The information problem
Cancer is an extreme example of a common problem in the Internet era: information abundance can be almost as much of an issue as information scarcity. There are over 620 UK cancer charities offering information. A web search on ‘Cancer Support Forums‘ returns over 33m hits, Cancer Support Groups’ 173m, and a search for the latest research in Cancer on PubMed (the largest online repository of Biomedical research) revealed over 3.5m publications. Although qualitative studies have shown that the Internet can be helpful most of the time, it’s hard to pick out the relevant up-to-date information from the sheer scale of noise and opinions that are ungrounded in science.
A 2018 study on the Analysis of Cancer Survivors Network (which has over 3.6m unique users annually) found that most people (80%) were most interested in discussion boards (providing non medical information, human stories and personalised information), followed by 60% wanting a search function to look for specific information, and 50% using the internet as a library of medical information. As often with the Internet it’s human to human communication that comes out top and this implies the need to shape information services in ways that build on this need for personal connection and an understanding of people’s individual lived experiences.
The problem is a different one when it comes to late-stage presentation. Not only is awareness of potential symptoms worryingly low, screening uptake for all cancers also lags behind other countries with similar programmes and has fallen in some cases. Research shows that cancer campaigns may increase awareness of symptoms but do not lead to longer term improvements in cancer incidence or mortality. Even when people do seek advice from their GP, they may not get the additional diagnostic tests they need when they need them, and are instead sent home with false assurances.
“The patients we send up the 2-week wait are the easy ones. We can do that with our eyes shut. Have they got red flag symptoms? Yes. The really difficult ones are the grey areas, the ones with only a 1–2% risk of cancer.”Dr Fiona Walter, Reader in Primary Care Cancer Research, University of Cambridge (from Unfinished Business, Health Foundation)
So what is the solution? The Long Term plan commits the NHS to a number of important milestones over the next 10 years to improve diagnostic pathways, increase screening capacity and offer more personalised support. Here at Nesta, we have been exploring some of the latest trends in collective intelligence design, and we also wonder if some more innovative answers could be found if we looked to combine existing diagnosis and treatment data, with the qualitative insights from patients and the public, alongside clinicians and others involved in cancer research and care.
Opportunities for collective intelligence in health
Nesta has long been interested in how data and digital technology can empower people to improve their health and wellbeing. We have called for a ‘knowledge commons’ in health to enable multiple interest groups to input and utilise shared data to create new insights and value. We have also recognised the vital role that patient organisations can play in bringing together peer-to-peer communities, and unlocking the potential of data to drive more active participation in healthcare by those most affected.
We’ve also funded and supported practical work in this area. We’ve worked with people affected by dementia to design and use apps on smartphones and tablets to connect and share data with researchers. We’ve also supported the 100 for Parkinson’s project, which uses a citizen science approach to allow people across the UK and United States to track their health for 100 days using a dedicated smartphone platform called uMotif. To date over 2.2 million data points have been collected and donated to Parkinson’s research. Our latest programme, the Healthier Lives Data Fund, in partnership with the Scottish Government, is looking for new ways to use digital technology to make data available and useful for citizens.
Building on this work, Nesta’s newly formed Centre for Collective Intelligence Design is exploring how human and machine intelligence can be combined to make the most of our collective knowledge and develop innovative and effective solutions to social challenges. We know that new digital technologies such as artificial intelligence can help us to analyse, predict, learn, communicate, and make better decisions faster. They can enable groups to become smarter than the sum of their parts, and mobilise human intelligence at greater scale.
Building blocks for collectively intelligent approaches to health
It is possible to imagine a future where citizens can control their health data, have easy access to knowledge, in formats that are appropriate to their cognitive and cultural preferences, and are supported to be active decision makers in their treatment and health outcomes. This untapped collective intelligence is opened up by new digital capabilities such as continuous real-time monitoring, data visualisations, open source software and machine learning. A fully realised collectively intelligent health system will take time to establish but we believe the path to success will be built on some of the following components:
- Developing a new knowledge commons
A knowledge commons for health needs to be built on the integration of different sources of information, including datasets provided by public sector organisations, such as NCRAS, and people’s lived experiences. Pioneering approaches to the management and search of existing knowledge, such as the database of systematic reviews compiled by the non-profit Epistemonikos, also help to ensure that we are building an evidence-based understanding of which interventions hold the most promise.
- Digital peer-to-peer networks involving both citizens and clinicians
Enhancing communication channels between those delivering and receiving care could help to establish a new infrastructure for participation and problem solving. Crowdsourcing has become an increasingly recognised approach within health research, promotion and care - it might provide a stepping stone towards more balanced collaborations between clinicians, researchers and patients.
- Deep cultural change towards awareness of health behaviours
Further public engagement and ethnographic research is needed to better understand the diverse cultural, economic, social and cognitive influences on people’s behaviours, and how to increase awareness and support people to seek advice and take personal action. Nesta’s Good Help work provides useful clues about how to shift power and support people to feel more in control of their own health.
- Place-based as well as people-centred approaches
Health behaviours are embedded within circumstances, reflecting the communities and places that make up our everyday experience. “Smart health” initiatives tap into Internet of Things (IoT) sensors, mobile networks and community support, to promote health at the city and neighbourhood levels. For example, the SafeNeighbourhood project in Brazil merges multiple data sources and crowdsourcing to improve the independence of the elderly population.
- Building cross-sector data collaboratives
There is a need to establish new partnerships around data, that involves health and patient communities, as well as experts in data, technology and innovation. We envisage a role for new intermediary institutions or data stewards embedded within organisations, to coordinate relationships between private and public actors in the health data ecosystem. This must ensure adherence to agreed governance principles that take ethics and data privacy as their foundation.
While data-driven insights are undoubtedly important, they are only part of the story. We believe that significant and lasting changes in health outcomes will only be achieved through more human and personalised approaches that merely use digital technology and data as an enabler. There are therefore clearly benefits to be gained by engaging the broader public in deliberations about how to solve wicked healthcare problems such as late diagnosis in cancer.
Collective intelligence and cancer
There are already some promising examples of collective intelligence in the cancer field:
US based PatientsLikeMe give people diagnosed with a range of conditions, including cancer, tools to track their progress, compare themselves to others with the disease, and use this data to learn more about how to improve their outcomes.
IBM Watson for Oncology uses big data and artificial intelligence (AI) to allow clinicians to quickly identify key information in a patient’s medical record, surface best practices from relevant guidelines and medical journals, and explore and rank treatment options. The clinician can then apply their own expertise to identify the most appropriate treatment options.
One of the winners of the £1 million Cancer Innovation Challenge in Scotland is smartphone app Owise. Owise enables women with breast cancer to self-report their symptoms in real time, as well as see their treatment plan and options. Early evidence is showing that the collective intelligence from the pooled data from thousands of women can potentially increase survival rate as a result of improved quality of life.
So far we see very little innovation happening to address the specific challenges of late diagnosis. This is why Nesta’s Health Lab and Centre for Collective Intelligence Design have launched a pilot project working with human design agency, Humanly, to explore how we might use digital technology, data and collective intelligence design to reduce the number of people presenting with late stage cancer, with a focus on bowel cancer. We'll be publishing more information about this project soon.
If you’re interested in getting involved in this project in any way, please contact Sinead Mac Manus on [email protected].