Is the NHS COVID-19 data that the government is reporting as accurate as it could be? Howard Davis looks at what the reality is and how the NHS can be supported in gathering the right information.
Much has been written in the press about the potential inaccuracies in the coronavirus data being reported on a daily basis. My experience supporting the NHS and national organisations to record accurate clinical and mortality data suggests there are three pressing issues we need to consider when recording and reviewing information on the pandemic:
1 Is NHS COVID-19 data an accurate reflection of cases in hospital?
Has the national change in acute pathways and the media’s pre-occupation with the outbreak distorted our understanding of the quality of care delivered to patients?
2 Is it possible to adequately record and review data during a pandemic?
Coronavirus affects front-line and back-office staff, making what should be business-as-usual into a daily challenge, but we still need to record accurate clinical information now to give us a platform to return the service to normality once the crisis has passed.
3 Will we be able to measure the real impact of COVID-19?
Inconsistencies in NHS COVID-19 data and calculations combined with the wholesale change to hospital activity profiles will all have a direct impact on the value and intelligence Summary Hospital level Mortality Index (SHMI) will provide.
The need to record and then scrutinise accurate clinical information is as critical as it has ever been. This is to give us a clear and immediate understanding of the spread of the virus, to measure performance and outcomes, and ultimately to gauge how well the government and the NHS have responded to the pandemic.
Is NHS COVID-19 data an accurate reflection of mortality in hospital?
Beyond the drawbacks we already know about the current NHS COVID-19 data, particularly that the death rates only account for those patients that have been tested for the virus, there is a real danger that the persistent focus on the outbreak will lead to inaccurate or misleading data on care-quality across the service. My experience over the last few weeks supporting the NHS to record clinical and mortality information has highlighted issues with the clinical data.
My colleagues and I have seen patients with underlying conditions, such as lung cancer, who are on end-of-life palliative pathways and then test positive for COVID-19. As reported in the media, this is a coronavirus-related death, but from a clinical perspective, it is not.
There are also instances of patients recorded as dying from the virus that were admitted for other conditions. For example, immuno-suppressed transplant patients who acquire the virus during their hospital stay. Recording this solely as a coronavirus death does not give the full story about the how and why.
Reporting all coronavirus-positive deaths without more detail and granularity on other presenting conditions and how the patient acquired the virus will distort the true picture of the impact of COVID-19 in the short-term and impedes our understanding of the quality of care delivered to other patients.
There has also been much coverage this week about care-home deaths not being reported. The changes in emergency pathways have resulted in a reduction in admissions across the board. Frail and elderly patients who would normally be admitted and ultimately die in hospital are no longer in the hospital reported deaths. It is difficult to quantify the number of patients who should have been treated for, and potentially die from, other conditions, but weren’t because of the current situation. Or who had symptoms, but weren’t tested or treated. Early figures from the Office for National Statistics (ONS) on all certified deaths in and out of hospital suggest that deaths are increasing, and many are not being recorded as related to COVID-19 on the death certificate.
Is it possible to adequately record and review deaths in a pandemic?
COVID-19 is dominating all my working relationships with my NHS colleagues at the moment. We all know the operational issues that Trusts are facing, with not just front-line staff off work and self-isolating, but back-office coding and information staff too. All while additional requirements are being introduced to document, report and review care. My experience of new systems and processes, and reduced capacity, tell me that this will adversely affect the accuracy of clinical NHS COVID-19 data just at the time when we need it to be as detailed as possible.
The process of clinically reviewing mortality must also be maintained to provide individual organisations with assurance over all aspects of care delivery - not just those for COVID-19 patients - to identify lessons learned and to explore the issues that need further scrutiny. We must refrain from blaming everything on COVID-19.
Once the surge in deaths, and maybe further delayed surges, have passed, and we have addressed the elective backlog that has built up, the NHS will need to return to business-as-usual as quickly as possible. Care delivery has changed, patients have not been treated, and NHS COVID-19 data must be there to measure the impact and help the service grow. We need to ensure we record accurate clinical information now, and ensure it is fit-for-purpose to tell the whole story, to give us a platform to return the country to normality.
Will we be able to measure the real impact of COVID-19?
Due to delays in NHS COVID-19 data availability, it will be July before the impact of this month’s deaths will be seen in mortality indicators and, depending on how long the pandemic lasts, much longer until we understand the full extent of COVID-19’s effect.
How useful will SHMI be in the long run?
SHMI uses diagnosis groupings, such as malignant neoplasm of lung, to make comparisons meaningful and to enable organisations to identify areas of risk. Currently, COVID-19 is coded using an emergency diagnosis code and will fall into a generic SHMI diagnosis grouping of Allergic reaction, aftercare and screening. Unless there is a timely change to the SHMI definition then the NHS and the government will not be able to differentiate between COVID-19 and the other patients that fall into this wide and generic diagnosis basket. Add to this the fact that suspected coronavirus patients will be classified by SHMI in the same way as those with positive tests, and the starting point for recording hospital deaths is likely to be widely different from the one being reported at the moment.
This is only so long as suspected COVID-19 is identified early enough in the hospital stay to be included in the SHMI calculations, as it uses the diagnoses recorded in the first two episodes of the hospital stay. We’re seeing patients who are only tested for coronavirus after a period of having symptoms that may be related, and some are even discharged onto palliative pathways for other conditions before the tests are reported. This delay in testing and reporting will affect the resulting SHMI diagnosis grouping, unless there is clinical input to ensure that COVID-19 is identified as their true cause of death.
There has been a wholesale change in the way care has been delivered over the past month that will make SHMI an odd indicator to measure and interpret. Many patients that previously would have been admitted to hospital are no longer receiving care, either because elective surgery has been cancelled or because emergency pathways have been radically redesigned to accommodate the virus. This change has been implemented inconsistently and at different points across the country. The rate of COVID-19 deaths has varied wildly across the country, meaning some organisations may see a considerable reduction in deaths, making any linear comparisons of SHMI performance challenging.
This will all have a direct impact on the Care Quality Commission and other regional and national organisations that rely on SHMI and the alerting process to measure quality and identify risks. However, it will not necessarily provide any further clarity than the real-time NHS COVID-19 data on the quality of the NHS response to COVID-19.
Is this an opportunity to make NHS COVID-19 data more useful?
To ensure the NHS COVID-19 data we use to assess our response is meaningful and to inform our understanding of care delivered to all patients, national organisations will need to ensure that any future analysis will identify COVID-19 separately from other conditions in SHMI calculations and differentiate between hospital and community acquired COVID-19. They will also need to be able to reconcile the need to report on the hospital coronavirus mortality rate against an understanding of deaths in the community that may be linked to the change in acute pathways that have been introduced in response to the crisis.
In the end, this will rely on accurate data recorded by individual NHS organisations, with clinical input to ensure the data accurately reflects the conditions treated and the underlying cause of these conditions. We know clinicians want to do this and staff want to help them do this, it is just sometimes difficult to prioritise this alongside the real business of responding to COVID-19.