Announcement of methodological changes to the ‘Statistics on people with a learning disability and autistic people in mental health hospitals’ publication

From the May 2026 (up to April 2026 reporting) publication onwards, the methodology used for the specified measures below within the ‘Statistics on people with a learning disability and autistic people in mental health hospitals’ publication will include the following changes.

1. Determining whether a patient is an inpatient at the end of the month

The Assuring Transformation (AT) dataset records a person’s stay in hospital as 'episodes of care', a stay may be made up of several episodes with a new episode starting if the person transfers to a different hospital or type of ward. Occasionally an episode is recorded as ending on one day and the new one starting the next day, but other information in the data indicates that the patient did not actually go home between the 2 – it was a continuous stay in hospital.

Where this happens at the end of a month (that is, patients who have an episode end date of the last day of a month and then next episode start date of the first day of the next month), the current methodology treats these patients as not being an inpatient at the end of the month. This means that the counts of ‘inpatients at the end of the month’ are lower than the true value. There are 14 instances of this in the data, affecting reporting periods between 2016 and January 2026.

The new methodology will treat these patients as being an inpatient at the end of the month, as the data indicates a continuous stay in hospital. The change will mean a more accurate view of the number of inpatients at the end of the month.

2. Determining a patient’s total length of stay

The current methodology looks only at the current record in the data for patients who are in hospital at the end of the month. If the record has a value in question 11b for the first admission to hospital in the current continuous inpatient stay (that is, date the patient first came into hospital), that date is used. If not, the date that the patient was admitted to the hospital they are currently in (Q11a) is used.

The new methodology looks at all records in AT for the patient and combines contiguous episodes of care into spells and considers all values of the ‘hospital admission date’ and ‘earliest known admission date’ in the spell to find the original admission date.

As some patient ‘spells’ contain several episodes of care, it is more appropriate to consider all values of the ‘hospital admission date’ and ‘earliest known admission date’. Furthermore, there are examples where the ‘earliest known admission date’ recorded in AT for the current episode conflicts with other records in AT where there are different ‘hospital admission’ dates for the same patient but different episodes of care.

The change will mean a more accurate view of the total length of stay for inpatients.

The revised methodologies affect the below tables we publish from the publication in May 2026 onwards.

Change 1: Inpatient counts

Change 2: Total length of stay

Change 1: Inpatient counts

Change 2: Total length of stay

The first publication to be affected by this change will be the May 2026 release, which covers Assuring Transformation data in the period from March 2015 – April 2026.

Questions and feedback on the publication are welcomed and should be sent to [email protected] or alternatively call 0300 303 5678.