Dear Ed
I write with an update on NHS Digital’s activities in response to the Office for Statistics Regulation’s systemic review of Mental Health Statistics in England. The review highlights a large and immensely important agenda, on which we are committed to making very strong progress.
We have laid some very important foundations this year alongside focussing on key priorities around the pandemic. This includes extending our flagship Children’s Mental Health Survey to provide new insights from over the lockdown period in collaboration with ONS and others. We have plans to address many of the outstanding areas from the recommendations in the forthcoming year, either directly or through collaboration with other producers. User engagement will be at the heart of these plans and, having engaged with users on specific aspects, we will hold a broader user event later in the year to outline progress to date, future plans and gain their feedback on this.
This update is predominantly centred on areas that NHS Digital is delivering directly or through collaboration with other organisations. We will work with the English Health Statistics Steering Group to send a further in-year update covering more comprehensively the work taking place across all national bodies.
In presenting what we have done and plan to do, I will take the recommendations in a different order to reflect the data journey. I would welcome your feedback.
Recommendation two: we want to see continued activity to improve the quality of underlying statistics datasets, as well as clear communication with users about quality issues
This has been our biggest single immediate focus as the quality of the underlying data underpins all other aspects of their utility.
We now have close to 300 organisations routinely submitting data, having grown by more than 100 in the last 18 months. This is estimated to cover 97.5% of relevant data. Now we have coverage from all the major organisations, we are turning our attention to the smaller voluntary sector and independent sector bodies, asking providers why they are struggling to submit data and giving them more support via webinars and tailored one-to- one engagement.
Alongside this we are focussing on providing the tools and insight to data providers and more widely so they can understand the quality of the data and what to do about it.
To support providers, we have recently published two dashboards, which we presented to providers through a webinar:
- one profiling data quality for every piece of information submitted by every provider
- a second with a specific focus on the clinical information
We are also automating a process to provide plain English descriptions back to providers on issues with their data.
Links to the dashboards and webinars promoting them are below:
Mental Health Services Dataset – Data Quality Dashboard – NHS Digital
Mental Health Services Dataset SNOMED data quality dashboard – NHS Digital
Mental health services events – NHS Digital
While our statistical publications contain a lot of detailed information around the quality of data, we will be augmenting this with higher level summaries of what the data can and cannot tell you in our improved mental health data hub.
Recommendation one: producers and organisations should exploit statistics through better data, greater analysis and linking data
In 2021/22, we are looking to develop the Mental Health Data Hub, our dashboards and publications to enable users to monitor key trends more effectively. Users will be at the heart of this design.
As we onboard more datasets to our Data Processing Services, we will look to link mental health data across care domains and discussions are underway with other national bodies to prioritise data linkage to answer key questions for the benefit of health and care. We have currently linked Mental Health Services Dataset (MHSDS) to ONS mortality data to produce an excess mortality indicator working alongside colleagues in PHE. We link to the maternity services dataset to produce indicators on perinatal mental health, and we are exploring linking to the community services dataset to look at infant mental health. We are also discussing with colleagues in NHS E/I and PHE around linking a range of datasets together to provide vital analysis on people’s routes through mental health services, with an immediate focus on young people with eating disorders.
Alongside this, we are collaborating closely with ONS by providing Improving Access to Psychological Therapies data under the Digital Economy Act for linkage to data to look at the wider determinants of mental health.
Recommendation three: we want to see clearer leadership and greater collaboration across producers of mental health statistics
Most sets of statistics used to monitor performance in mental health are derived centrally from the detailed person-level dataset (the MHSDS). This is not always the case for other care settings where there are separate aggregate flows of performance data. Using the person-level dataset for performance statistics ensures consistency and transparency of reporting and improved ability to audit. It also enables people to drill into specific aspects of performance more easily, as well as a reduced burden on data providers. Finally, it makes data quality issues more apparent and is very dependent on the quality and completeness of the underlying data, on which I set out our work to date in the first section.
Considerable work has taken place across organisations to reach a position where nationally we have a single version of the truth for most topic areas. Duplicate collections have been removed on areas such as Early Intervention in Psychosis (EIP), Delayed Transfers of Care (DTOC) and creating an enhanced learning disabilities dataset. The final areas area where we are looking to drive up the quality of MHSDS to remove the need for a duplicate data flow is on Out of Area Placements (OAPs) and Young People’s Eating Disorders (CYP-ED).
There will be a strong focus this year on how we present the headline measures so they are more accessible to users “at a glance”.
Our NHS D mental health analytical lead chairs a Mental Health analyst national group who share their work and methodologies and feeds into the English Health Statistics Steering Group. The main initial focus has been on greater streamlining, coordination and collaboration and has aligned or is aligning methodologies on suicides, performance reporting and geographies and a work programme to build on this is being developed.
Recommendation four: access to NHS Digital data needs to improve
NHS Digital’s emerging data strategy will have a strong focus on improving metadata and mental health will be an area of early activity on this. We are already publishing more detailed information around datasets in the Technical Output Specification. The strategy will also include tools to enable users to slice and dice anonymous open data.
Over the course of the pandemic, we have already made open data available earlier on our supplementary information page.
NHS Digital is in the final stage of a data access review to identify and deliver against priority areas for improvement. We acknowledge in particular that the availability of data from the large mental health surveys has not been in line with users’ expectations. While some of the tighter process has been a necessary response to evolving legislation and technology, it has been too challenging for users with an appropriate legal basis to gain access. In addition to implementing the general improvements from the data access review, we are developing an agreed precedent by which access to survey data can be agreed through our Data Access Request Service which aims to improve ease and timeliness to survey data for users. This is an area we will continue to monitor closely and provide updates on progress.
Summary
I hope this gives you an overview of the work delivered, underway or planned in this very important area within NHS Digital. I would welcome your team’s feedback on any of this and we will write in-year with a broader update from across the English Health Statistics Steering Group.
Yours sincerely,
Chris Roebuck
Chief Statistician