Advancements in technology and the use of data within the NHS have enormous potential to improve efficiency and patient outcomes through new areas such as artificial intelligence (AI). With the likely proliferation of these approaches in coming years, the Department of Health and Social Care (DHSC) has published a new code of conduct for data driven health and care technology. It points out that
Data-driven technologies must be harnessed in a safe, evidenced and transparent way. We must engage with patients and the public on how to do this in a way that maintains trust.
Based on our dialogue research1 with patients and the public about the uses of their data in our DeepMind Summit work, forthcoming work for the Academy of Medical Sciences on data-driven technologies in healthcare, and our dialogue with the HRA about consent for data linkage, we welcome the initial code of conduct, and the principles of engagement with the public that the DHSC sets out.
Firstly, it is important to note that the NHS starts from a position of strength. Our dialogues with the public show that they tend to trust the NHS implicitly. They give the NHS the responsibility, and the mandate, to work out how to share our data in the public interest. This is a heavy responsibility as well as an indicator of the esteem in which the public holds the NHS – it means that the NHS bears more of the burden of 'getting it wrong' in the public’s eyes than the tech companies do.
It is therefore good to see the DHSC taking this initiative, outlining an initial code for the NHS and its suppliers to sign up to, providing guidance about what is expected of them. And not only this, but the code fits well with what the public thinks are the main issues.
In our dialogues with the public, patients and stakeholders for DeepMind, we developed six principles for how the NHS and tech companies should work together. These very much overlap with the code’s 10 principles.
Three of our principles were about impacts: that tech should be used to solve the problems that need solving now, not future-oriented problems (participants wanted to resist a Silicon Valley mindset that everything can be solved by the application of tech). They also thought any project using tech should work towards progressive social outcomes (e.g. reducing health inequalities). Finally, they thought these projects should have a measurable impact on individual care and NHS staff and systems.
The focus in DHSC’s initial code on evaluating impacts is strong. It emphasises the importance of defining the user, the value proposition and the commercial strategy. It also asks for evidence of how effective a product or innovation is for its intended use – something we know will resonate with patients, who were keen to ensure that tech projects 'stayed within their remit', and spoke about co-designed work which keeps returning to its objectives so the benefits to patients don’t get lost somewhere along the way.
The other three principles defined in our DeepMind Summit were about behaviours: participants said that how things are done is as important as the ultimate aims. They thought that the NHS and tech companies should adhere to principles of transparency about exactly what they are doing; focusing on data security; and also work in collaboration with frontline clinicians, patients and the public.
Again, the initial code reflects these behaviours. Focusing on transparency, it says organisations should be 'fair, transparent and accountable about what data is being used'.
However, our research shows that transparency is easier to talk about than to achieve.
In theory, it is all very good to be transparent about what exactly you are doing, but it needs to land in a context of public understanding. And the public currently do not know much about tech and data; furthermore, they don’t know how the NHS currently works and the role of public and private services with the delivery of care.
We find that dialogue participants often question why the NHS would need to work with external tech companies at all. Even the idea that the private sector is involved in delivering NHS care is, in itself, news for some people. The public will need to appreciate the need for new tech company partnerships with the NHS, or transparency could backfire. They might seek to opt out of data sharing agreements just on principle as they fear their data might be shared indiscriminately with commercial organisations who might be associated with data in other non-health contexts.
Having said this, our work for the Wellcome Trust suggests that most people are willing to share their data with commercial organisations in healthcare – if they believe the data is being used for a social good. So this returns us to the need to communicate the intended impact of a project, and the importance of defining the social good that it hopes to achieve.
The final principle the new code sets out is that algorithms underlying these projects should be explained. Again, given low technical knowledge of algorithms and data science generally among the public, it would be challenging to communicate the complexities of these algorithms to a lay audience. However, the code is realistic about what the public can be expected to know; it provides a good workaround, stating that organisations should outline 'how you are evaluating the algorithm's effectiveness'. The evaluation process, and the questions asked, could be scrutinised by an interested public, even without technical computing knowledge.
So, overall, we welcome the introduction of a new initial code of conduct and it fits with what people tell us are the main issues when we spend time explaining it to them in dialogue research. But we must be mindful of the need to bring the public along with the tech and data explosion gently, providing the right information at the right time to enable a fuller understanding of the role of new tech in healthcare.
Perhaps supplementary guidance could also be given on how to communicate these new approaches to patients and the public; which might also help the NHS and tech companies themselves navigate the challenges associated with complying with the code. Having the public on side is vital – as the document says,
Our ability to unlock the benefits of data-sharing relies on the public having confidence in the health and care system’s appropriate and effective use of data.
Note 1: “Public Dialogue brings together members of the public, policy makers, scientists and other expert stakeholders to deliberate and come to conclusions on national public policy issues involving science and technology”