Collaboration: working together to maximise value for users of data and statistics

What we said in October 2021:

Sharing and linking data can have life-saving impacts. This must be prioritised by governments beyond the pandemic.

Strong analytical collaboration resulted in valuable, high-quality, coherent statistics during the pandemic. Taking this approach to other topics will help overcome existing and future problems.

We encourage producers in the four nations to continue to engage on projects to provide comparable data across the UK. However, we also recognise that differences in policies in each country may mean that providing UK-wide comparability is not always possible. Producers should clearly explain to users any impact on the statistics and how they can be used, resulting from differences in policies across the UK.

What we found during October 2021-October 2022:

Strong collaboration was an essential part of the statistical system’s response to the pandemic. It demonstrated that through collaboration, producers can better meet the needs of users. We found two main purposes for collaboration during our follow-up: working together to share data and working together to produce coherent, and where possible comparable, statistics.

Collaborating to share data

On data sharing, our overall finding remains the same as in our 2021 report: while data sharing increased during the pandemic, barriers still exist. Over the last year (October 2021-October 2022), there have been more good examples of sharing data between producers in order to produce new statistics (see the case study below). We also found that data sharing initiatives established during the pandemic have resulted in producers gaining a greater appreciation of the benefits of sharing data with each other, such as improving data quality. Producers are therefore more likely to explore new data sharing projects. However, producers identified many of the same barriers we highlighted last year, including a lack of clear, consistent guidance on how to manage legal and information governance issues, and a reluctance among some to share data. These issues are broader than just health and social care data – they are relevant for the whole statistical system, as we set out in our 2022 State of the Statistical System report. Data sharing will be an area of focus for us over the next year as we follow up our Joining Up Data reports.

The Goldacre review, published in April 2022, is a key development in this area from the last year. The review explored ways to deliver better, broader, safer use of NHS data for analysis and research in England. The recommendations from the review include building a small number of secure analytics platforms (often referred to as Trusted Research Environments or TREs) and making these the norm for all analysis of NHS patient data, including by academics and NHS analysts. The use of TREs was endorsed by users we spoke to and was also strongly supported by the citizens’ juries on health data sharing in a pandemic jointly commissioned by the National Data Guardian for Health and Social Care, NHSX and the National Institute for Health and Care Research Applied Research Collaboration Greater Manchester. There are already some TREs across the UK which provide access to health data for research, such as the NHS Safe Havens in Scotland and the SAIL Databank in Wales. We hope to see further developments in this area to support safe health data sharing within and beyond the government analytical community.

Collaborating to produce coherent statistics

Our 2021 report identified the need for stronger analytical leadership and coordination of health and social care statistics. The most acute need for this was in England, where the organisational landscape is the most complex. In response to our recommendation, the English health statistics leadership forum was established, bringing together Heads of Profession for Statistics and lead statisticians from across the health statistics system in England. The forum aims to ensure that statistical outputs are aligned and complement one another, as well as taking a joined-up approach to prioritisation, horizon scanning and methodologies. Recently, an equivalent four nations leadership group has also been established, to complement other topic-based four nation working groups.

We found that topic-based working groups have been particularly successful in supporting a coordinated approach to health and social care statistics. During the pandemic, cross-UK groups were an essential part of the delivery of UK-wide statistics about COVID-19 deaths and vaccinations, as well as the UK COVID-19 dashboard. The cross-UK group on social care resulted in the creation of the UK adult social care statistics landscape page, which brings together statistics on social care for the whole of the UK. This includes a four nations matrix which informs users about which data are and are not comparable across the UK. In Scotland, the COVID-19 Data and Intelligence Forum was a successful way for producers to collaborate during the pandemic. This approach is now being used for social care data and statistics with the establishment of a Social Care Data and Intelligence Forum. The Forum oversees a joint work plan for Public Heath Scotland and the Scottish Government, aiming to improve cross-organisational coordination, data sharing and information governance.

As well as increasing coherence and comparability of statistics for users, these collaborations have benefited producers. Producers told us that they have more awareness of what other organisations are doing, know who to contact in order to collaborate and have improved joint horizon scanning. Producers want to continue this level of collaboration, expanding some of the four nations groups which are currently focused on COVID-19 to include other topics (for example, the vaccine group expanding from COVID-19 to include flu and other vaccines). The English Health Statistics Steering Group (EHSSG) has played a key role in supporting several of these topic-based groups. The EHSSG was established in 2016 with the aim of enhancing the coherence, production, dissemination and accessibility of health and care statistics in England. In its 2022-2027 workplan, the EHSSG states the importance of UK level coherence and where possible aspires to achieve this. We support further use and development of topic-based working groups in order to continue to improve statistics for users – for example, the cross-UK group on performance statistics, which includes waiting times, will be an important group coming out of the pandemic given this is a priority topic for users.

Users of health and social care statistics often need to compare statistics, for example between geographical areas. We found that users of health and social care statistics, such as parliamentary researchers and health think tanks, have a strong interest in comparable UK-wide data. Topics which users said they wanted comparable UK information on included waiting times and spending. Users do appreciate that comparability is not always possible due to differences in policies and operations in the four nations.

However, users told us that it can still be hard to determine exactly what is and is not comparable across the UK. For example, it is good to see that the EHSSG and statisticians across the UK have done a lot of work to understand and summarise differences in waiting time statistics. But this information, or a link to it, is not included in all of the relevant publications for each nation to easily inform users about comparability.

Some other examples users gave us where a lack of comparable data presents a challenge include inconsistent groups (e.g., different age bands used across organisations) and inconsistent measures and presentations of statistics. For example, we heard from researchers at the Nuffield Trust about the challenges they faced using statistics about deaths. In putting together comparable data on the location of deaths across the UK (see Annex 4 of the Nuffield Trust’s Support at the end of life report), the researchers found that obtaining sources with comparable time periods and definitions, and containing data on actual and 5-year averages, was difficult.

There remains more work to do to improve cross-UK comparability of statistics. We encourage producers in the four nations to continue to collaborate to publish comparable data where possible. This includes considering methodologies and definitions used, as well as the presentation of the statistics. Where UK-wide comparability is not possible, producers should clearly notify users that this is the case to aid user understanding and support appropriate use.