Mental Health Inpatient Census 2024 - Parts 1 and 2

Methodology and further information

The eighth Mental Health Inpatient Census was carried out on 10th April 2024 by the Scottish Government and NHS Boards. The census covered three parts: 

Part 1: Every patient occupying a psychiatric, addiction, or learning disability inpatient bed in an NHS Scotland facility on the census date. 

Part 2: Every mental health, addiction, or learning disability patient whose care is funded by NHS Scotland but is being treated in a facility outwith NHS Scotland. 

Part 3: Every patient who was in receipt of Hospital Based Complex Clinical Care (HBCCC) in general acute inpatient beds on the census date. 

 All 12 NHS Scotland territorial boards with psychiatric, addiction, or learning disability inpatient beds provided a return. The State Hospital (Special NHS Board) also provided a return. 

The Scottish Government’s Digital Acquisition Services provide data collection and validation support for a number of statistical returns. Digital Acquisition Services provided secure data collection software (procxed.net) and first stage data validation checks. Further information about the data collection software can be found in the Data Privacy Impact Assessment which can be obtained on request from MHIC@gov.scot 

The data collection, analysis and reporting of the census has been overseen and produced by Scottish Government statisticians. All statisticians in the Scottish Government are part of the Government Statistical Service (GSS) which comprises the statistics divisions of all major departments in the UK Government, and the devolved Governments of Scotland, Wales and Northern Ireland plus the Office for National Statistics, which has a coordinating role for the GSS. 

Data confidentiality

A Data Protection Impact Assessment was undertaken prior to the census which outlines how patient confidentiality is maintained. The Data Protection Impact Assessment can be obtained on request from MHIC@gov.scot 

In addition, statistical disclosure control has been applied to the analysis. Statistical Disclosure Control (SDC) covers a range of ways of changing data which are used to control the risk of inadvertent disclosure of sensitive information about a person or unit (such as a household or business). This publication has used the following methods where there are under 5 patients in a particular category: 

• Suppression of possibly disclosive cells (e.g. where the value is small) which means that the value for that cell in the table is not given and secondary suppression of cells which means at least one other value in the row or column is also not given to ensure that disclosive cells cannot be deduced through subtraction. 

• Table redesign and recoding, where cells are grouped together to protect small value cells. 

• Rounding – as an alternative to secondary suppression, but preserving some information about the relative size of the categories. 

Further information about Statistical Disclosure Control is available here: Statistical disclosure control - Office for National Statistics (ons.gov.uk) 

Revisions to 2022 report and archiving of pre-2022 publications.

As a result of a 2024 project to improve the internal coding used in the analysis of the census, some historical errors were discovered in the process used to collate the data needed to build the census dataset. 

The errors meant that a small amount of data relating to the number of inpatient beds was missing in some boards across the years 2014 (Forth Valley), 2016 (Fife and Lothian), 2017 (Grampian and Lothian), 2019 (Tayside) and 2022 (Lanarkshire and Lothian). The 2018 census was unaffected. Most health boards (8) were unaffected by any errors across the timespan of the census 

The missing number of beds and the health boards affected across the years are described in the tables below: 

Table 21: Inpatient beds missing from Mental Health Inpatient Census Part 1 at the time of publication, by health board. 

Mental health, learning disability and addiction beds, NHS Scotland, 2014-2019 census. 

Year	Geography	Beds Missing	Beds - Published	Beds - Revised	Percentage Missing
2014	Scotland	16	4532	4548	0.4%
2014	NHS Forth Valley	16	240	256	6.3%
2016	Scotland	98	4254	4352	2.3%
2016	NHS Fife	54	252	306	17.6%
2016	NHS Lothian	44	645	689	6.4%
2017	Scotland	32	4205	4237	0.8%
2017	NHS Grampian	10	353	363	2.8%
2017	NHS Lothian	22	649	671	3.3%
2019	Scotland	10	3922	3932	0.3%
2019	NHS Tayside	10	357	367	2.7%

Table 22: Inpatient beds missing from Mental Health Inpatient Census Part 1 at the time of publication, by ward type. 

Mental health, learning disability and addiction beds, NHS Scotland, 2014-2019 census. 

Year	Ward type	Beds missing	Beds - published	Beds - revised	Percentage missing
2014	All	16	4532	4548	0.4%
	Dementia care & treatment	16	*	*	*
2016	All	98	4254	4352	2.3%
	Acute	30	*	*	*
	Addiction	24	*	*	*
	Young people’s unit	32	*	*	*
2017	All	32	4205	4237	0.8%
	Continuing care / long stay / recovery (non-dementia)	14	500	514	2.7%
	Young people’s unit	18	54	72	25.0%
2019	All	10	3922	3932	0.3%
	Addiction	10	72	82	12.2%

*  data not published for census years 2014 and 2016. 

Key Messages 

• The missing data for beds is not impactful at a Scotland level and does not affect all health boards. 

• None of the key messages, trends or conclusions at a national level from previous publications are affected. 

• Some health board-level data and tables will be affected in previous publications, although broadly, the impact is limited. 

• Whilst the amount of missing data is small, it permeates through a range of tables and commentary which makes a complete overhaul of all previous year’s publications prohibitively, and disproportionately, costly in terms of the analyst time required to re-run and republish all of the data. 

The actions we have taken are: 

• Changes to the census data collection procedures and new validation checks have been implemented to prevent a re-occurrence in the future. 

• The publication and data for 2022 has been revised and re-published alongside the 2023 publication. 

• The latest publications (2023 onwards) use all of the revised data from 2022. 

• Tables and graphs in the revised 2022 and latest publications have been clearly marked for users to indicate that direct comparisons (for some health boards) are not possible across the entire time series. 

• Publications prior to 2022 have not been revised and have instead been archived in the National Records of Scotland web archive with clear messages for users about the missing data. 

• The decision to archive pre-2022 publications was taken after a consideration of the resource required to rebuild and re-run all of the analysis and republish all of the time series from 2014 onwards. The conclusion was that this work would have incurred a disproportionate cost, given the relatively small impact of the missing data. The best solution was therefore to archive the data and publications, still making them available to users but with clear signposting about the approximate nature of the results

Implementation of sex and transgender questions

In 2023 the questions used to collect data in this area were aligned with the Scottish Government, Office of the Chief Statistician’s guidance on sex, gender identity and trans status: Sex, gender identity, trans status - data collection and publication: guidance - gov.scot (www.gov.scot) 

This improved the language, design and consistency of the questions being used and has enabled people to be more appropriately represented in the survey. 

The change to reporting of sex instead of gender and the introduction of extra questions here also brings the Census questions into alignment with other Health and Social Care surveys such as the Scottish Health Survey, and the National Records of Scotland population census. 

The change in the questions are described below: 

Prior to 2023 the following question was asked: 

Q - Gender of patient: 

• Male 
• Female
• Other
• [for coding only] Not known 

From 2023 onwards, the following questions are now asked: 

Q - Sex of patient:

• Female
• Male
• Prefer not to say
• [for coding only] Not known 

Q - Does the patient consider themselves to be trans, or have a trans history? 

• Yes
• No
• Prefer not to say
• Not known 

Q - If the patient would like to, how would they describe their trans status (for example, non-binary, trans man, trans woman)? 

• Answer can be supplied as free text response. 

This means that from 2023 onwards data on sex that is reported in the census is not directly comparable with data collected on gender for prior years. Notes have been added to the relevant sections of the publication to make this clear. 

The census is one of the data collections that is part of the Equalities Data Improvement Programme (EDIP). The EDIP was a commitment to expand and make improvements to protected characteristics data. This was part of the wider Equality Evidence Strategy, published in March 2023. More details on this are provided here: Scotland’s Equality Evidence Strategy 2023-2025 - gov.scot 

Health Conditions

The International Classification of Diseases (10th Revision) has been used in the analysis for specific health conditions. The health condition and relevant ICD-10 code can be found in the following table:

Selected Health Conditions	ICD-10 Codes
Dementia	F00 – F03
Alcohol misuse	F10
Drug misuse	F11 – F19
Schizophrenia	F20
Schizotypal and delusional disorders	F21 – F25, F28 – F29
Manic episode	F30
Bipolar affective disorder	F31
Depression	F32 – F33
Persistent mood (affective) disorders	F34
Other mood (affective) disorders	F38 – F39
Neurotic, stress-related and somatoform	F40 – F45, F48
Behavioural syndromes	F50 – F55, F59
Personality Disorders	F60 – F66 and F68 – F69
Learning Disabilities	F70 – F73 and F78 – F79
Behavioural and emotional disorders with onset usually occurring in childhood and adolescence	F90 – F95 and F98

Autism

For the purpose of this report, the autism cohort is defined by the patient meeting at least one of the following criteria: 

Diagnosis of Autism	ICD-10 Code F84.0 or F84.1
NHS Boards answered Yes to the following health condition question: Autistic Spectrum Disorder	1 - Yes 0 - No

Alcohol misuse cohort

For the purpose of this report, the alcohol misuse cohort is defined by the patient meeting at least one of the following criteria:

Diagnosis of Alcohol Misuse	ICD-10 Code F10
NHS Boards answered either 1 or 3 to the following question:   Was there a history of alcohol dependence or substance abuse in the four weeks prior to admission to hospital/care home?	1 - Yes – alcohol dependence or harmful use of alcohol only 2 - Yes – substance abuse (excluding alcohol) 3 - Yes – both alcohol dependence and other substance abuse

Drugs misuse cohort

For the purpose of this report, the drug misuse cohort is defined by the patient meeting at least one of the following criteria:

Diagnosis of Drug Misuse	ICD-10 Code F11 – F19
NHS Boards answered either ‘2’ or ‘3’ to the following question:   Was there a history of alcohol dependence or substance abuse in the four weeks prior to admission to hospital/care home?	1 - Yes – alcohol dependence or harmful use of alcohol only 2 - Yes – substance abuse (excluding alcohol) 3 - Yes – both alcohol dependence and other substance abuse

Forensic patients

Forensic patients were identified if NHS Boards indicated ‘yes’ to the following Census question:

Is the patient being managed primarily by forensic services?

Access to the data for further research

To enable further research and statistical analysis, extracts of the Inpatient Census data may be made available for approved researchers.

Academic researchers must initially apply to the ‘Statistics Public Benefit and Privacy Panel’ or the ‘Public Benefit and Privacy Panel for Health and Social Care’ to gain access to the census data. 

Once an application has been approved, a copy of the original application form and a copy of the approval letter should be emailed to the following address MHIC@gov.scot for approval by the Scottish Government (Health & Social Care Analysis Division and the Principal Medical Officer for Mental Health). 

NHS Boards will have a version of their own census dataset which contains information about patients for whom they are responsible for providing treatment or are responsible for funding. NHS boards will have their own arrangements in place for researchers to access health data. 

All boards have a Caldicott Guardian who is responsible for assuring confidentiality and enabling appropriate data sharing, and a director responsible for research and development.