Dear Dr Gulhane,
Thank you for your letter of 9 November asking us to investigate concerns about statistics on Accident and Emergency (A&E) waiting times in Scotland. The Scottish Government has a target that 95 per cent of people attending A&E should be seen within four hours.
Your letter refers to an article in the Scotsman, which points out that an estimated 2,000 patients who present at the Acute Assessment Unit (AAU) of Glasgow’s Queen Elizabeth University Hospital each month are excluded from Public Health Scotland’s (PHS) monthly waiting times statistics. The author suspects there may be inconsistencies in data collection here because patients presenting at an apparently similar Assessment Unit – the Medical Assessment Unit (MAU) at the Western General Hospital in Edinburgh – are included.
The monthly A&E statistics (and the Government target) cover all types of A&E site, including Emergency Departments, Minor Injury Units, and smaller community casualty sites. Virtual attendances and activity taking place in trolleyed areas of assessment units, which are often located alongside A&E departments, should also be included. Patients admitted to staffed beds in an Assessment Unit (rather than spending time on trolleys or chairs) are considered Emergency Admissions rather than A&E attendances and should be included in separate Inpatient and Day Case Statistics to which the four hour A&E access standard does not apply.
A&E statistics for the Western General Hospital include activity for both the Minor Injuries Unit and trolleyed areas of the MAU. For the Queen Elizabeth University Hospital, patients coming into trolleyed areas of the AAU via the Emergency Department should be included in the statistics. We understand from PHS that, due to limitations of the current data collection, activity in trolleyed areas of assessment units cannot be differentiated in the data. Therefore, statistics for the AAU are not reported separately but should be included as part of the overall A&E activity reported for the hospital. PHS has acknowledged this issue and is undertaking further work to assess whether or not all relevant activity in this assessment unit is being included in A&E submissions it receives.
In addition to the monthly A&E statistics, Public Health Scotland also publishes weekly waiting times statistics, which are often used by the media to report and compare hospital performance against the Scottish Government target. However, these are confined to Emergency Departments, which PHS defines as “large hospital departments which typically provide a consultant-led, 24-hour service with full resuscitation facilities and designated accommodation for the reception of emergency patients”.
The way in which services and facilities are defined in the monthly and weekly statistics is clearly very important to understand from a user perspective, especially in the context of the Government target. Background information and a glossary are available online, but the recent confusion suggests that they should be made more accessible and transparent for users. The Office for Statistics Regulation has suggested this to PHS. It has also asked PHS to communicate more clearly any caveats regarding data collection issues across various sites.
Separately, you raised a concern that initiatives adopted by individual hospitals may result in inconsistencies in the statistics. This recent Herald article claims that NHS Tayside fares well in the statistics due to a ‘continuous flow’ model at Ninewells Hospital, which means that some patients who would be waiting on trolleys to be seen in A&E elsewhere wait instead on trolleys in the acute medical receiving unit – where they do not count towards the waiting times estimates. It is not for us to say how hospitals should manage their emergency admissions policies, but the Office for Statistics Regulation has urged PHS to make it clear where this is likely to create difficulties in comparing waiting time statistics across hospitals and boards.
Finally, it is important to ensure that PHS guidance on data collection and classification is applied consistently across health boards. PHS has advised us that it is reviewing this guidance due to the increasing emergence of new clinical pathways to A&E. The Office for Statistics Regulation will continue to engage with PHS as it does so and as it responds to our feedback on the presentation of its statistics.
The issues raised in this case around difficulties making comparisons on NHS data in Scotland are indicative of a longstanding broader challenge in getting comparable data on healthcare provision across the UK, between nations and within them. It is important that users of statistics are able to compare the performance of the NHS across the UK on issues such as the waiting times for emergency care, and I encourage statistical producers to take this into account as they develop their statistics.
Sir Robert Chote