23 April 2018
The Chief Scientific Officer's (CSO) Annual Conference is a major event in the healthcare science calendar. Titled Partners in the Future,this year's conference saw delegates gather at the Royal Society for two packed days focusing on the importance of science in health (something PHG Foundation has always championed) and why collaboration is integral to the future of the NHS.
The main theme of partnership was accompanied by discussions on health data and artificial intelligence (AI); patient involvement; and skills and training – especially in the field of bioinformatics. Here, we present a short summary of the main messages from the conference.
Health systems have been collecting patient data for a long time. The Chief Editor of Nature, Helen Pearson, gave an example of how the collection of health data can lead to system-wide insights as she talked through the major findings of the British Birth Cohort Studies. These pioneering studies have been gathering data to inform policy and deliver change to address health inequalities since they began in 1946. What is new, argued Professor Peter Bradley of Public Health England (PHE), is the amount of available data. This data presents new opportunities, prompting ideas as to the problems it can be used to help solve – both at the individual level in the personalisation of treatments (enabling people to make better choices for care), and at the population level, for example, containing flu outbreaks.
Other speakers outlined reasons for collecting and using data in healthcare, with frequent reference to changing population demographics, and the need for long-term, connected health monitoring as individuals live longer and accumulate chronic conditions. Dr Richard Scott, the Clinical Lead for Rare Disease at Genomics England, emphasised the importance of systematic information gathering for patients with rare diseases; data concerning these conditions are inherently scarce, therefore sharing of information, with research right alongside, is especially important.
So, more data - whether on people or pathogens, rare diseases or common conditions - was said to be better for healthcare. There is a lot of it around; as later mentioned, the US healthcare system alone collected 150 exabytes (that’s 150 billion gigabytes) of patient data in one year. However, much of that data is poorly structured and goes unused. Questions around how data is collected and shared, and to what end, have prompted considerable discussion in the public arena; the conference was no exception.
Several speakers focused on the importance of patient involvement in healthcare, and emphasised the need for public approval of the use of data generated by and for the NHS; with some pointing out that, after all, ‘it is their data’. Although this is not strictly accurate, patients certainly do have rights regarding how data connected to them is used. The issue of obtaining comprehensive consent in the context of research using genomic data was raised and later expanded on to include its limitations and the potential of alternative models of consent.
In a stimulating talk, Prof Michael Parker, Director of the Wellcome Centre for Ethics and Humanities, argued that, with knowledge constantly changing, future uses of patient data cannot be specifically consented for. He proposed the use of alternative broad consent models, including the establishment of a social contract for responsible data use by all parties (including scientists) and derived from public discussion i.e. not left solely in the hands of the ‘experts’. Although intriguing, broad models of consent could present a challenge, given the imminent implementation of GDPR.
The importance of trust and understanding between collaborating partners in health was affirmed several times; data misuse, discrimination and exploitation, particularly by the commercial sector, were accepted as major points of concern for patients and the public.
The hype around new technologies - and how ready the system is to make use of them – was picked up by Prof Stephen Keevil of Guy’s and St Thomas’ NHS Foundation Trust, who asked, ‘artificial intelligence – is it all hype, or could it save the NHS?’.
Echoing concerns PHG Foundation outlined in our consultation response and evidence to House of Lords Select Committee on Artificial Intelligence, Professor Keevil and others emphasised the need for agreed standards within AI health trials and agreed regulatory approaches. Citing research, including the use of AI in the identification of disease such as tuberculosis and diagnosis of Alzheimer’s, Professor Keevil argued that, while prospects for AI are exciting, results to date are far from earth-shattering. There was general consensus that, although AI will undoubtedly be a useful tool for clinicians, and one that should be embraced in some circumstances, it will not replace skilled professionals, such as radiologists, within the NHS. Indeed, where it is in use already, AI is driving up the need for skilled bioinformaticians, who are currently in short supply (as are radiologists).
Training was another theme of the conference, with broad recognition that bioinformaticians and data-literate personnel are badly needed if UK healthcare is to take advantage of the potential improvements offered by new technologies and systems. Joanne Hannaford of Goldman Sachs stood out amongst the day’s speakers, as she provided insights into data use and management, including the use of data lakes, from outside the field of health. Ms Hannaford suggested that the organisations that survive will be those that excel at using data effectively. She added that GS are not replacing people with machine learning; with AI they simply aim to enhance individual’s abilities. The reference to data lakes was particularly timely given (as later mentioned) that the NHS plans to establish data lakes to facilitate the integration of data for research.
If the UK health system is to stop playing catch-up with technology, there is much work to be done. At the conference, the CSO pondered whether a ‘100,000 Genomes Project’ -style outcome (from establishment to commissioning WGS for the NHS within a few years) for AI could provide the means.
Having opened the conference talking about her recent experience as an NHS patient on a journey through diagnosis and treatment, the CSO closed the event with some suggestions as to the new technologies she thought likely to drive improvements in the service patients receive. As well as the anticipated establishment in the UK of a national genomic medicine service offering clinical whole genome sequencing later this year, the CSO listed:
Although some scepticism was reserved for virtual reality, the audience broadly agreed with her summary. For those not lucky enough to attend the conference, our healthcare futures infographics provide a ‘reality check at a glance’ for these and other emerging technologies – what they are, how they might benefit healthcare, and how close they are (or are not) to clinical application.