The Goldacre review is a milestone independent report, by a credible independent-minded expert, and is a call to action for the profound impact of data and analytics on improving healthcare and outcomes. Professor Ben Goldacre’s credentials are excellent and his views on the good-use and misuse of data and the power of analysis are well respected. His independence adds weight to the call for the NHS to see analysts as key members of the NHS healthcare team, previously championed by Simon Stevens and more recently Amanda Pritchard, who said in 2021 that “a good analyst can save more lives than a good anaesthetist”.
It is timely because it builds on growing the reputation of the analytics community and sets out a vision for the future. This seven-chapter comprehensive and accessible review covers everything from analytics to data curation, privacy and security to future strategy. It is most useful as a waypoint from which there is no return and sets a clear direction for the innovation and power that analytics brings to tackling the profound challenges to healthcare delivery we face today.
More open to debate, which the report usefully stimulates, is the how. The details of who and what, and where and when, the programmes, policies, investments, pilots, structures and authorities to act and deliver are where the debate now lands. As a practitioner on the ground I see the report as a call to action, rather than a prescription to advance analytical capability in the NHS and ICSs and ensure it is core and not a nice to have extra.
So there is lots to please the lifesaving analysts in the NHS, a call for professionalisation and funding for much loved grass roots organisations like AphA and the NHS R Community. A call for stronger collaboration between clinicians, decision makers and analysts to ensure the right questions are asked, answered and then actioned! And a call for no ‘black box’ analytics from private companies, we need to understand how the algorithm works to be able to evaluate if it is useful, and when we, the NHS, pay for a tool we should pay for it once for universal access for everyone that wants to use it.
There are two big themes I think the report trips over on. I can’t see the need for a new national analytics bureaucracy such as the NHS National Analytics Service to be housed in the centre. And there is not enough on how we get analysts and analytics to where it is needed and can have most impact, PCNs and elsewhere on the front line where decisions are being made but without the benefit of being as well informed as analytics can make them.
Our awesome analytical response to Covid came from jobbing grassroots analysts openly sharing code and their differing approaches to solving common problems. This was facilitated by NHS England through FutureNHS (the web-based information repository and sharing tool). NHS England got it right by enabling the sharing of ideas and problem solving, not by trying to bureaucratise and control it. A national analytical service’s prime goal must be to stimulate and champion data analytics to flourish in the NHS, and rapidly sharing and promoting achievements that everyone can implement.
I would love to have seen more support for local grassroots analytics, people that primary care and PCNs can lean-on for additional capability and capacity, be that in their ICS (surely that makes the most sense?), CSUs or internal consultancies like the Strategy Unit, the Health Economics Unit or networks like the Midlands Decision Support Network. Do we really want to concentrate analytical capability in the centre? Don’t we still need capable analysts in primary care to take the centrally created code and use it on local data?
The review suggests a competitive process for more funding. Does this mean that money flows to the more capable teams, as they write the best bids highlighting their already present capabilities, and that the teams with the best bid writers (rather than best analysts) win because many great NHS analytical teams struggle to find the time to dedicate to writing a comprehensive bid as they are already working at capacity. How about a targeted investment to support places where there is already capability to further grow but also money to flow to areas that need to grow their local analytic profession, addressing some of the inequalities in analytics provision? Being proud to work locally, to help make a difference locally, serving a population they know and love.
As you might suspect there is a lot about privacy and security, IG and accessing data. Primary care and GPs hold the crown jewels of health data. The comprehensive, granular data within GP systems has the potential to provide unapparelled insight. When it’s analysed with data from other healthcare settings, social and community care then the opportunities to understand and act on wider determinants of health become much more potent.
The review discusses how Trusted Research Environments (TREs), a secure environment that lets approved analysts access approved data for approved projects remotely, not being able to download or take data ‘away’ only the statistical/analytical outputs of the analysis, could be the answer to sharing our data. Professor Goldacre knows these well and set up OpenSAFELY a TRE that houses a national primary care data set to support urgent research on Covid.
The review suggests we shouldn’t have lots of TREs, as quite rightly that poses potentially more risk to the data being compromised or accessed inappropriately, and rather TREs should consolidate over time.
Goldacre rightly champions TREs to safeguard data security and protect against misuse. TREs are great, and very much appeal to me and my team, and when we are not supporting PCNs across the country we are often doing big national and regional projects that that are research or evaluation based and more consistent access to data would be a dream come true.
Many NHS analysts don’t do this kind of work. I would suggest the majority are still undertaking daily management, operational and performance management analytics, supporting decision makers to run their multimillion/billion pound organisations. The potential is enormous if we have the imagination to see analytics as a strategic asset and not just an operational tool.
Primary care, and especially PCNs, are seen as the ‘engine room’ for delivering PHM within an ICS. They are key to moving interventions up stream, having a better understanding of their communities and are therefore being best placed to decide on which intervention is most appropriate. Access to broader data (the wider determinants of health) and insight from analysts will be essential to help make the best informed decisions to support the best outcomes.
There is loads of wonderful, extremely positive and actionable insight in the review, and I am delighted that passionate people that clearly care about data and analytics wrote it. How much of this talks to the daily work of the majority of the NHS analysts I am less sure about, and how much of this will help provide much needed analytical support at the coalface of primary care remains to be seen.
Andi Orlowski is director of the Health Economics Unit at Midlands and Lancashire Commissioning Support Unit and senior adviser to NHS England on population health management.