How can data really improve health and care?

The government has published the draft of its national data strategy. The document outlines how the government intends to use the data across the NHS and social care to improve health and care for patients, staff and researchers. There is a significant focus within the strategy on linking and making better use of data that already exists and outlining what needs to be put in place for this to happen. Much of what it contains is hard to disagree with and would likely lead to improvements. However, there is less emphasis on the type and quality of data that is gathered and having access to the right data at the right time to make the best possible decisions.

One of our recent reports focussed on this topic in the context of tackling health inequalities and the Covid vaccine take-up. Gathered from the experiences and perspectives of 50 people from a wide range of healthcare and lived experience backgrounds, it makes a number of practical recommendations about gathering and presenting data to aid understanding of the complexities of health inequalities. Whilst related to Covid vaccine hesitancy, much can be extrapolated from these findings to inform general principles.

Our findings indicated the following difficulties with data:

1. The classification of ethnicities is not accurate and detailed enough to be helpful, (for example, all Eastern Europeans are classified as “other White”). There is little contextualised and qualitative health information available and, whilst there is extensive data for some groups, there is little or none for others. As a result, the differing needs of particular communities are not known, and so greater targeted support and community-lead activities are not enabled. As a result, ineffective “one size fits all” approaches continue to be used. In terms of the Covid vaccine roll-out, it is not clear who we are missing, and we do not have data on what aspects of the current approach are not working for some.
2. Much of the data needed for a clear understanding of what is happening in this context is not being routinely captured, collated or analysed. As a result, it is not possible to identify who has not taken-up the free vaccination offer or why they have chosen not to do so. Also, there is no systematic collection of qualitative information about people’s Covid experiences, and how particular groups are being affected differently. As a result, there is little insight available into what messages would best positively influence take-up rates.
3. Granular data on ethnicity is not readily recorded as part of health treatment, so the levels of vaccine take-up by different BAME groups is not accurately known. The same is true for other structural factors that affect inequality, such as deprivation, disability and age or on non-health categories such as informal carers. This means that little insight is possible into the more complex analysis of rate intersectionality of these various factors, in different combinations.
4. Although the NHSE Race Observatory is developing a BAME communities data depository, there is no current clarity about what is being collected or about the levels at which different types of health data should be collected (including Primary Care providers, NHS Trusts, and wider population Public Health practitioners) Lower-level data is often valuable despite the risks that a smaller sample can affect statistical validity.
5. It is unclear who is gathering useful health information currently and how information could be better shared and some organisations are resistant to data sharing, which hinders a more comprehensive understanding of public needs. This, in turn, hinders the development of trust for some groups, where service deliverers were not “culturally competent” in their approach. The current tiered vaccine roll-out has also unintentionally reinforced the view of some groups that  they  are deemed to be less worthy than others, because despite having significant needs or disabilities, they are being made to wait and do not understand why.
6. There is insufficient BAME involvement in vaccine research and development, both inside the laboratories and in clinical trials. The relative absence of data on how the vaccine might affect particular groups differently also contributes unhelpfully to a climate of misinformation, mistrust, and hesitancy, which undermines vaccine take-up rates by some groups. Little appears to be being done to support those who are concerned or even frightened about vaccine safety, though appropriate peer support would be likely to help.

And, as a result, the following recommendations for action were made:

1. The NHS should capture more granular data on key factors affecting health such as ethnicity, disability, and those with long term conditions. This should also be in a form which allows intersection readily with data on other key health determinants, such as housing, unemployment, deprivation, and insecure employment.
2. The NHS needs to urgently undertake a comprehensive review of its data requirements in this context, to identify  what  is being collected and by whom, where there are gaps, and how such information can be better collated, analysed and reported,  to  support  improved  decision-making  on how best to meet the differing healthcare  needs  of  particular  communities and groups, including the use of local, place-based, peer  support arrangements, to encourage greater  trust  and engagement  with  For example, local  health  workers  should  be able  to  target  particular groups,  such  as  Bangladeshi  women  with  disabilities  living  in  poor housing, to better support health improvement and maintenance.  Local people need to be able to see a line of transparency and accountability on such information and how it is used.
3. A more sophisticated, multi-faceted, targeted approach be adopted to the use of appropriate information to support a greater vaccination take-up. This must include research on the use and  effectiveness  of  a wider range of  media to  convey  key messages and  reassurance,  using local messengers who are likely to be trusted by the recipients.
4. The next National Census in 2031 be extended to include additional information on ethnicity and on the wider determinants of health inequalities, to enable a greater correlation of the interplay of the different factors, and a better understanding of the scale and  effect of what they cause.

The full report is available here.