How can big data improve humanity?
Alexander Hitchcock and Eleonora Harwich, researchers at UK think tank Reform, discuss the increasing importance of big data in the future of the healthcare industry, and how its use could benefit patients and pharmaceutical companies alike.
It is perhaps the most tantalising question for the fourth industrial revolution: how can pharma firms use big data to design medicines that treat patients more safely?
Put simply, big data means data sets so large or complex they cannot be processed using traditional methods. Genome sequencing is one example – the human genome is around three billion base pairs in size. Huge claims are made about big data’s value. A committee of MPs recently valued it at £216 billion for the UK economy between 2012 and 2017. In healthcare, they believed it added £14.4 billion of value.
Big data can improve patient care by refining treatments to make them more personalised. Understanding the effects of medicines once they are used allow regulators to better understand safety implications for patients. A molecular knowledge of patients’ characteristics allows companies to design tailored drugs.
This isn’t tomorrow’s world anymore. Alastair Kent, Director of Genetic Alliance UK, argues the only way to accurately understand diseases at a molecular level is to collect huge amounts of genomic data: “Diseases which look the same on the body, when you look at molecular level, are quite different,” he says. The cost of genome sequencing needed to do this has fallen (per genome) from $95 million in 2001 to $1,250 in 2015, according to data from the National Human Genome Research Institute Genome Sequencing Program.
The UK Government has also caught on to big data’s potential. John Manzoni, Chief Executive of the Civil Service, recently told Reform think tank that it underpins “the next phase of public-service reform and the transformation of government.” It now needs to create the infrastructure to allow pharmaceutical manufacturers to exploit big data for the benefit of patients.
Creating the right foundation
The healthcare system in the UK is extremely data rich, with NHS Digital holding a wealth of patient-level data on people that have interacted with community services, hospitals, mental-health services and maternity services. De-identified patient-level data are highly valuable for industry to understand the effectiveness of a drug, conduct research and identify the point in the patient pathway at which it would be most effective to administer the drug.
It is, however, extremely hard to access this sensitive information. Obtaining consent is costly, time-consuming and can generate issues of selection-bias, which implies that the sample might not be representative of the population. Silos also hamper the joining-up of certain datasets, which could be of great value, such as health and social-care data. The Law Commission describes the legal landscape of data sharing within public services to be ‘unduly complex and inconsistent’. Yet it can be done in the UK. GlaxoSmithKline partnered with healthcare organisations in Greater Manchester and is currently pioneering the world’s first randomised Phase 3 real-world effectiveness trial – the Salford Lung Study – using a single electronic medical record linking primary-care, secondary-care and pharmacy data.
Imitating this across the country in a cost-effective way could foster research, innovation and improve patient outcomes. For this to become a reality, the government needs to ensure that its data infrastructure is fit for purpose, as well as modernise the current data-governance regulations. Breaking down silos, as Greater Manchester organisations are doing, and ensuring that data are collected and presented in a standardised way are important steps the government can take. The Department for Health should also further develop the health data finder as a complete repertoire of all such data to facilitate researchers’ understanding of the information that is out there and what is available to request.
The NHS Research Ethics Committee could be used to deal with granting access to patient information in high-risk cases. Whether it is through the creation of a new body or through using existing ones, the government should play an active role in ensuring this ambition becomes a reality.
The UK has some of slowest uptake of new drugs in Western Europe – it typically takes 10 to 15 years to develop a medicine. George Freeman, former Minister for Life Sciences, proclaimed that “time is the great killer of value in this sector.”
Quicker access is the goal. The government needs to focus on enabling drugs to be introduced earlier – in particular to those with rare diseases, for whom randomised-control trials are not always feasible. Speedier access can also build a ‘real-world’ evidence base for effectiveness, reflecting different patient reactions and that patients will not always take the full course of drugs. Yet Kent says it is currently easier to be ‘canonised by the Pope’ than introduce novel therapies in the NHS. He points to the multi-layered commissioning approach before treatments are tendered, which takes ‘considerable time’. Genetics Alliance UK has called for ‘drastic streamlining’ of this process to get faster access to treatment and real-world data on its effects.
The UK’s Accelerated Access Review and the EMA’s adaptive pathways trial promised to rapidly introduce drugs to patients. The former offered regulatory approval in stages to adopt innovative treatment for ‘transformative’ drugs, while the latter aims to simplify some parts of the pathway to market and speed up other stages for restricted patient populations.
It may be ‘too soon to tell’ whether Accelerated Access will achieve its aim of taking years off drug introductions, according to Kent. The principle is a good one, however: Dr Shahid Hanif, Head of Health Data and Outcomes at the ABPI, argues it needs widening beyond a narrow range of medicines to rapidly increase access to data on the effectiveness of treatments, enabling their full access and improvement at a much faster rate.
Investment drives innovation. The pharma industry is one of the biggest investors in R&D in the UK, spending £4 billion a year. The government has an interest supporting this investment to maintain the best care for patients; the 2017 UK Industrial Strategy recognised this by injecting an additional £4.7 billion in R&D by 2020-21.
An important step in making sure this money is put to best use will be the coordination of efforts between government, industry and academia. The Industrial Strategy sets out to do this by helping link universities and the world of business. Innovative partnerships, such as the one between the University of Manchester, GSK and AstraZeneca, have led to a world-leading centre for inflammatory disease.
Initiatives like this should be further encouraged and supported – for example, a medicines manufacturing innovation centre which could provide a hub for SMEs, multinationals, academia and suppliers to research and collaborate.
A big data social contract
The benefits of big data will never be realised without public buy-in, as organisations such as the Nuffield Council on Bioethics have recently warned. Faith in the government is currently not high: a 2016 survey found that just 36% of people trust the government with their information, compared to 53% trusting banks.
A new ‘social contract’ is needed to increase trust, according to Dr Lydia Drumright of the University of Cambridge. As Dr Drumright told Reform this year, “when a patient goes to see a doctor, there is an unspoken social contract about how the information the patient divulged to the doctor will be used; we need to understand what that is in society.”
The Wellcome Trust’s independent data taskforce could be one way to codify this social contract. The Wellcome Monitor (Wave 3) study found that eight in ten people are willing for medical records to be used for medical research if anonymised. A national information campaign is crucial. Scotland’s SHARE initiative proves that an online and leaflet operation can recruit large numbers of people to share data for research. Social media offers an avenue for reaching certain groups in tailored ways. Such an initiative will clarify the purpose of data sharing – for research, not marketing or insurance.
Asking patients to opt-out of sharing this data is a powerful tool. In organ donation, it is well-established that opt-out processes for organ donations lead to higher donation rates, for example. However, as the recent care.data programme confusion showed, the public need to be fully informed about how their data is being used and feel secure in their rights to opt-out. Any decision not to share must form part of the public information campaign, and patients should be able to opt-out online or through their GP.
Role of the sector: Transparency and openness
The pharma industry also has a role to play in increasing trust over how patient data is used. This can be done through increased transparency of clinical trials, which as the Tamiflu tablet scandal highlighted, are currently opaque.
Openness about data will not only ensure that doctors have all the information they need when prescribing drugs – it will also speed up the development of new cures. The Project Data Sphere initiative was created as a medium to accelerate the development cures for cancer. Manufacturer transparency should be encouraged; by not sharing results, the research community cannot be aware of existing data and can incur huge unnecessary cost by repeating clinical trials.
The government and healthcare industry need to capitalise on the information age. Increasing knowledge of genetics and untapped NHS patient records offer huge opportunities for manufacturers to develop precision medicines and get them to market quickly. This creates a virtuous research cycle, with more data improving development of treatments. The government needs to create the framework to support manufacturers to exploit these opportunities safely and for the patients’ benefit – then big data might work for us all.