A basic NHS data gap: how common are common conditions?
The NHS does not publish statistics on many common conditions, especially conditions diagnosed by GPs like neurological and respiratory conditions.
This means charities don’t know how many patients need their services - some even spend money to get this basic data. Not even the NHS itself knows for certain.
The NHS does have a service that could produce these statistics, but it is ‘at capacity’.
We recommend that the forthcoming UK Statistics Assembly should discuss the need for these statistics, and explore ways they could be produced.
The problem
The NHS does not publish statistics on how often many common conditions are diagnosed (aka their ‘incidence’), especially those diagnosed by GPs. This makes it harder for charities to support patients and campaign for them. Some charities currently pay for this data, using time and money that could be used to care for patients.
As one example, to plan its support services, the charity Parkinson’s UK needs to know:
How many new Parkinson’s diagnoses have there been this year?
What are the demographics of people diagnosed, and where are they?
How many people are living with the condition overall?
But the NHS does not produce statistics on these questions, so Parkinson’s UK pays for data to produce its own estimates. Amazingly, official NHS planning bodies now rely on the charity’s data, rather than vice versa - while Parkinson’s UK can only afford to update its estimates once a decade because this research is so expensive.
The cause
This happens because many diagnoses happen in primary care, and GP systems are not set up to share data routinely - as the Sudlow Review, published this week, explains in detail. Data is fragmented across different systems.
This means we have poor data on the incidence of conditions that are diagnosed mainly by GPs, ranging from Alzheimer’s to respiratory conditions.
Why it matters
This problem affects many healthcare charities, as we found in our report on reported gaps in health statistics. Some examples we were told about:
Asthma + Lung UK cannot find out how many respiratory diseases are diagnosed every year, which limits their ability to plan and make the case for investment.
The MS Society faces similar problems and has also paid for prevalence data.
Versus Arthritis cannot get regional data on the prevalence of musculoskeletal conditions, so estimates this in its own statistical model, which again reduces the resources available for supporting patients.
The lack of data also affects researchers, public health bodies and MPs. Versus Arthritis told us: “By not measuring these conditions, it became difficult for local planners to take them into account when planning their services, and risked the conditions being overlooked”.
Potential solutions
Although the underlying data is fragmented, there are existing initiatives in the NHS to share data across GP systems. Some of these are already used to produce diagnosis data:
The General Practice Extraction Service (GPES) run by NHS England collects patient data from GP practices. This is used to produce official statistics on dementia diagnoses. However, GPES is ‘at capacity’, and cannot accept new data requests.
OpenSafely, run by the University of Oxford and LSHTM also aggregates GP data, but can only be used for Covid-related research.
The Clinical Practice Research Datalink, a non-profit service run by the MHRA, collects anonymised patient data from GP practices,and can be also used to produce statistics. This was the source used by Parkinson’s UK for its estimates. However, it charges for access to data (starting at £20k, with no charity discount).
What we need to change
In the short term: the UK Statistics Authority should include a discussion on the need for better statistics for charities at its forthcoming Statistical Assembly.
In the medium term: the GPES should be asked to produce new experimental statistics. Given that it is already produces dementia diagnosis statistics, this is presumably relatively straightforward.
In the longer term: as the Sudlow Review recommends, we need better national management of GP data - which should include access for charities as well as academics - and appropriate official statistics.