Review of mental health harmonisation

Policy details

Metadata item Details
Publication date:28 March 2023
Owner:Government Statistical Service (GSS) Harmonisation Team
Who this is for:Users and producers of statistics
Type:Harmonisation standards and guidance

The Government Statistical Service (GSS) Harmonisation team are based in the Office for National Statistics (ONS). The team supports harmonisation of data across the UK. This involves providing bespoke advice and harmonised standards and guidance about how data on a range of different topic areas are collected and presented. The standards and guidance are designed to improve the consistency, coherence and comparability of statistics.

Producers of statistics use these harmonised standards as a starting point in their data collection process. By using harmonised standards, they can:

Harmonised standards also allow people to effectively and accurately compare data that has been collected across different datasets. This means we can more easily understand what that data does, and does not, tell us. This ensures that statistics are used to their full effect for the public good.


We have harmonisation guidance for mental health which provides information related to definitions, important reports, legislation, and clinical measures. But there is currently no harmonised standard specifically for mental health.

Some limited data is collected through our standards for disability and wellbeing. Our harmonised standards for disability capture data on mental health as it aligns to UK legislation; these questions enumerate long-term conditions (at least 12 months) that impact daily activities, but they do not distinguish between physical and mental health conditions. In addition, our impairment standard has an option for “mental health”, although this cannot be separated or “disaggregated” by condition. A review of the harmonised disability standards has been published alongside this report. At the other end of the spectrum, we do provide a harmonised standard for personal wellbeing. Together, these standards provide narrow and fragmented mental health data that includes wellbeing and disability. They were not designed to collect mental health data, which means users cannot separate information about specific mental health conditions from the data. They also do not account for temporary episodes of poor mental health.

This review of mental health harmonisation fulfils a commitment published in our February 2022 GSS Harmonisation Team Workplan. We pledged to work with stakeholders and user groups to:

  • understand whether a harmonised standard for mental health is needed
  • develop and test a mental health harmonised standard if it is needed

This report fulfils the first pledge and contains our findings from our initial research and engagement work. This work included:

  • asking how mental health data was used
  • finding out about any gaps in the mental health data that is currently being collected
  • finding out what mental health data our users would ideally want to have

This work has highlighted many user needs that are challenging and complex. Some of these needs are being addressed as part of our work to update the disability standards. These have been indicated throughout this review. Once we have completed our work on disability we plan to investigate additional needs. This will help us consider a harmonised standard that captures high quality data specifically for mental health.

You can find information about all the methods mentioned on this webpage in our accompanying review of the disability harmonised standards.

Reasons for this review

No existing mental health standard

Mental health is broadly accounted for in the harmonised disability standards. The long lasting health conditions and illness harmonised standard and activity restriction harmonised standard gather data about mental health in line with UK legislation, including the Equality Act (2010) for Great Britain and the Disability Discrimination Act (1995) for Northern Ireland.

The questions are designed to gather data about people who report a physical or mental illness or condition that lasts at least 12 months and reduces their ability to do day-to-day activities.

The impairments standard includes non-physical impairments such as:

  • cognitive abilities — including difficulties with learning, understanding, or concentrating
  • mental health
  • social or behavioural impairments — including autism spectrum disorder and attention deficit hyperactivity disorder

There is also a harmonised standard for personal wellbeing which can be used to measure overall life satisfaction, happiness, and anxiety. But these standards were not designed to capture mental health data specifically, which means they are not suitable for this purpose. They only capture fragments of the data users need. For example, they do not sufficiently consider specific conditions, neurodiversity, multimorbidity, or the fact that mental health can fluctuate.

The need for a mental health standard

We have tried to establish the need for a specific mental health standard in our latest research and engagement work. We have worked with multiple stakeholders across public, private, and third-sector organisations to find out more about the demand for a mental health standard and what the user needs would be. We discovered that a dedicated standard for mental health would be useful to policy makers, charities, Disabled People’s Organisations, and service providers for a variety of reasons. For example, being able to gather data about people with specific neurodiverse conditions could help to establish employment rates and target specialist funding streams to address individual needs.

The data needs of our stakeholders are extensive, complex, and varied. Although we would not be able to address all these needs, we have conducted this review to begin to understand and collect information about them.


In 2019, the Office for Statistics Regulation (OSR) conducted systemic reviews of mental health statistics in England and Northern Ireland. They spoke to a wide range of statistics users who told them that the existing statistics on mental health did not give enough information about people and their conditions. Their research identified that, although the existing mental health statistics go some way to meeting user needs, there is much more that can be done. Their recommendations described the need for better data with improved quality. A harmonised standard for mental health could help to meet these recommendations by improving consistency, coherence and comparability of data.

Mental health context

Mental health is complex, which makes it particularly difficult to measure. There are three main reasons why this can be difficult:

  1. There is a spectrum of mental health that ranges from everyday wellbeing to clinically diagnosed disorders.
  2. People’s mental health can change daily from good to poor.
  3. People often have more than one condition. It is common to have more than one mental health condition, or to have a mental health condition and a physical health condition.

Because of this complexity we must first develop an accepted definition of mental health. We must define a scope that is both clear and consistent, and outline exactly which aspect, or aspects, of mental health our users want to capture.

“Mental health” is a term that can mean many things. In the UK there are many terms that are used interchangeably, such as “mental illness” or “mental health problem”. For example, the NHS refers to “mental health conditions” in their list of commonly accepted disorders such as depression and schizophrenia, but UK legislation uses the terms “mental disorder” and “mental impairment” in the Mental Health Act (1983) and the Equality Act (2010).

Health is a devolved issue across the UK, which means the terms and definitions we use differ across the four nations. The Office for Health Improvement and Disparities (OHID), Public Health Wales and the Department of Health Northern Ireland follow the definition of mental health proposed by the World Health Organisation (WHO). But the Scottish Government uses a broader definition in their Mental Health Strategy: 2017-2027. This definition refers to a range of things from emotional wellbeing to mental disorder, or an acute reaction to stress, and mental illness, such as schizophrenia.

1. A spectrum of mental health

The definition used by the Scottish Government highlights the spectrum of mental health. Everyday wellbeing applies to the entire population and is relatively straightforward to measure. There is a harmonised standard for wellbeing and the Scottish Government have planned a wellbeing census for children and young people, highlighting how this information can be captured relatively easily at the population level.

But clinically diagnosed disorders are experienced by a minority of people and are particularly difficult to measure. Highly skilled professionals are needed to diagnose conditions and high rates of clinical disorders are expected within hard-to-reach populations. This includes people who are homeless or living in institutionalised settings, such as prisons, hospitals, or sheltered housing. The 2014 NHS Digital Adult Psychiatric Morbidity Survey (APMS) is possibly the most comprehensive data collection on mental health in England and in recent years it has been limited to persons aged 16 and over who are living in private households.

2. Changing experiences of mental health

In addition, individuals’ experiences of mental health may fluctuate quickly from good to poor and the opposite. For example, personal wellbeing can depend on both internal and external factors that change from day to day. The harmonised standard for personal wellbeing reflects this by asking respondents how they felt “yesterday”. The temporary nature of mental health also applies to the experience of clinically diagnosed disorders. For example, at the extreme, bipolar affective disorder is characterised by occasions of elevated mood, lowering of mood, or both.

As indicated by the Scottish Government’s definition, there are also intermediate periods of acute mental distress. People may experience elevated stress, anxiety or depression, often in response to a major life event, such as postnatal depression. These periods will not always be diagnosed but it is still important to capture information about them. This means that any harmonised standard must consider both long-term and short-term experiences of mental health. The wording of questions will need to reflect the timeframe of interest.

3. More than one condition at the same time

We must also consider the high rates of people who experience more than one mental health condition at the same time, and people who have physical conditions alongside mental health conditions. For example, low levels of mental wellbeing are associated with having a chronic physical condition and high levels of psychiatric morbidity, which includes post-traumatic stress disorder and attention-deficit hyperactivity disorder. Similarly, the NHS Adult Psychiatric Morbidity Survey (APMS) found that people with severe symptoms of a common mental disorder are more likely to also report a chronic physical condition, compared to those with no or few symptoms. This means that statistical needs for respondents and users are unlikely to be met if we measure physical health, wellbeing, or psychiatric conditions separately.

Neurodiversity and mental health

A harmonised standard for mental health must also consider neurodiversity. Neurodiversity refers to the different ways that a person processes information. Neurodiversity is a spectrum, but people are often categorised as being neurotypical or neurodiverse. People may identify as neurodiverse if they have conditions like autism spectrum disorder (ASD), attention deficit hyperactivity disorder (ADHD) or attention deficit disorder (ADD), dyslexia, and dyspraxia.

The research community, policy professionals and employers are becoming more aware of neurodiversity. Our research shows there is a growing interest in including neurodiversity in a harmonised standard to help with data collection.

Currently, neurodiverse people are accounted for by the harmonised standards that measure disability, but only if they experience reduced ability to carry out daily activities in line with the Equality Act (2010). The impairment standard also includes a “socially or behaviourally” response option which directly references ASD and ADHD as examples. But this should not be used to gather data about the neurodiverse population because it is not exclusive to this population, and it does not apply to all individuals who have ASD, ADHD, or other neurodiverse conditions. Our stakeholders have also told us that the wording of this response option is misleading because it implies that people with ASD or ADHD demonstrate poor behaviour. We plan to address this issue in our future work. You can find out more about the impairment standard and the work we have planned in our review of disability harmonised standards.

During our research, stakeholders have explicitly stated a need to consider neurodiverse conditions within our disability standards. But, by definition neurodiversity means that differences are not deficits. This means neurodiversity may therefore be better considered alongside mental health. Although the two are separate concepts, there is some conceptual overlap. Historically, ASD and ADHD have been included as mental disorders in diagnostic tools such as the International Classification of Diseases and Related Health Problems (ICD) and the Diagnostic and Statistical Manual of Mental Disorders (DSM). The Adult Psychiatric Morbidity Survey (APMS) includes ASD within its definition of common mental disorders, and it references ADHD when discussing people who have more than one mental health condition. A recent systematic review and meta-analysis also found that mental health conditions were more prevalent in those with ASD when compared to the general population.

Although our disability standards may not be the best place to include questions on neurodiversity, it is the most feasible place at the moment. In the long-term we aim to design new questions so users can:

  • gather data about the neurodiverse population
  • separate, or “disaggregate” data so that users can investigate particular conditions

We will review where these questions are placed if and when a mental health standard is published.

Existing approaches

Although there is currently no harmonised standard for mental health, many organisations collect mental health data in a variety of different ways. You can find out more about how people are currently collecting mental health data in our mental health harmonisation guidance. The GSS Harmonisation team conducted a survey on data capabilities and processing. We found that 53% of our sample collect mental health data. Our sample included stakeholders across government, devolved administrations and private sector collectors. Full details can be found in our review of disability data harmonised standards.

Disability harmonised standards

At present, “mental health conditions or illnesses” are captured in the long-lasting health conditions or illnesses harmonised standard. But this standard combines mental and physical conditions with no ability to separate the two.

We also know that this harmonised standard can miss conditions that change over time because it only asks about conditions that last 12 months or more. It is likely that this question will underestimate mental health conditions because mental health can alternate from good to poor in a short time period.

The impairment harmonised standard is often presented to individuals who respond “yes” to the long-lasting conditions question. This standard looks at the activities a person can or cannot do because of their health condition. Respondents are given a list of impairments and are asked to select all that apply. Mental health is included as one of the response options, but this question is not a suitable way of gathering data about mental health conditions. There are two reasons for this. Firstly, mental health is likely to be undercounted in this two-step process due to its changing nature. Secondly, unlike physical impairments, there is only one response option for “mental health”. This wide interpretation of mental health does not allow for any separation of data to inform provision for specific funding and services relating to specific conditions. If multiple response options are selected, there is no way of identifying whether mental health is the main impairment.

Personal wellbeing harmonised standard

Information about personal wellbeing is captured in the personal wellbeing harmonised standard. This set of four questions covers life satisfaction, worthwhileness, happiness, and anxiety on a ten-point scale. The scale starts at 0, which means “not at all”, and ranges to 10, which means “completely”. The standard is useful for collecting an overall wellbeing score and it can be used to capture data at one end of the mental health scale. But users should be aware that satisfaction and worthwhileness are trait measures, whereas the happiness and anxiety questions are state measures, meaning they relate to specific timeframe.

Although wellbeing and mental health are interrelated, this standard should not be used as an indicator of mental health disorders across the spectrum. For example, it cannot be used to gather information about diagnosed anxiety disorders.

Medical condition-based methods

Mental health data is captured in surveys and administrative data using a mainly medical condition-based method. These surveys ask respondents whether they have conditions such as “depression”, “anxiety disorder”, or “bipolar disorder (or manic depression)”. You can find a detailed list of survey questions and administrative data sources in the mental health harmonisation guidance.

Clinical and validated measures of mental health

There are many clinical and validated measures of mental health used as monitoring, screening, diagnostic, or classification tools. The measurement tools are usually questionnaires that ask respondents to self-complete and score themselves on a variety of topics. Each questionnaire usually contains multiple items to assess various clinical disorders. An example of one of these questionnaires is the Generalised Anxiety Disorder.

These questionnaires are validated, replicable and reliable measures that have been thoroughly tested. But users should be aware they have not been designed for statistical use. The length of each questionnaire makes them often impractical for surveys, both in terms of cost and resource.

The Adult Psychiatric Morbidity Survey provides comprehensive coverage for a range or mental, behavioural and neurodiverse conditions but it involves considerable resources and methods. The APMS uses a variety of screening tools such as the Social Phobia Inventory, Psychosis Screening Questionnaire and the Clinical Interview Schedule. It can be used as a self-completion survey, or the surveys can be done as face-to-face interviews by trained and experienced data collectors. Because of the resources involved the survey is only completed every 7 years. Unless these tools are completed, coded, and fed into administrative data sources they are unlikely to meet our user needs.

Administrative data

Administrative, or “admin”, data sources of mental health do exist. For example, NHS digital hold information from a variety of sources including Hospital Episode Statistics which detail admissions and attendances at NHS hospitals in England. There is also the Mental Health Data Hub which hosts a variety of publications and datasets, including the Mental Health Data Services Data Set. Admin data sources are a useful resource to complement survey research. They are inexpensive, sustainable, and timely. They may also be the most accurate tool in the case of diagnosed clinical disorders. But they are unlikely to meet the harmonised standards of consistency, comparability and coherence because they are designed for operational purposes and not for research. The GSS Harmonisation team are currently exploring how to improve harmonisation of admin data.

Stakeholder needs for mental health data

Initial stakeholder engagement has given us some insight into the users of mental health data and what they need from the data. Our findings will be used to inform future work related to a harmonised standard for mental health data.

Definitions and guidance

Policy professionals and researchers clearly stated the need for better guidance or clear definitions on the difference between clinically diagnosed mental health, stress, and mental wellbeing. This included interest and clarity on how mental health intersects with disability, substance misuse, and neurodiversity. This would enable policy professionals to:

  • design more effective policies
  • make better decisions
  • provide more targeted services, preventative measures, and specialist services

Clearer definitions and guidance would enable researchers to:

  • improve their understanding of mental health
  • create more targeted research questions and funding bids
  • improve data quality – additional guidance would help respondents and improve the accuracy of data

We partially address this need within our disability work. We aim to design new questions that capture data about the neurodiverse population. Agreed definitions and scope will also help develop a harmonised standard for mental health.

There is also a need for further guidance or examples for the existing mental health response option in the impairment harmonised standard. Researchers from non-profit and government organisations have asked for this guidance. Social researchers asked for guidance on how to distinguish between common mental health problems and more serious mental health conditions that need treatment.

Policy makers also expressed a need for mild to moderate scales when looking at mental health impairments because the current standard is too broad. Many other data users agreed that this would help respondents answer more accurately, which would produce better data and improve the quality of their findings. We plan to design and test improvements for the impairment options as part of our work to revise the disability standard. This will include the mental health and socially or behaviourally response options.

A non-profit researcher asked whether we could remove the terms “conditions or illnesses” to describe mental health, so that it would be a more inclusive and accessible way of respondents discussing their mental health. The researcher explained that this would improve accuracy of responses. Such improvements in inclusivity would also increase alignment of data with the Equality Act (2010), but only where mental health has a substantial and long-term adverse effect on a person’s ability to carry out normal day-to-day activities. The language used in the long-lasting health conditions and illnesses standard has been updated from asking whether people have a disability, which placed the accountability on the individual. But our research shows the updated terms “conditions or illnesses” have also been criticised for being too medical compared to the preferred term “impairment”. The term “impairment” is used in the Equality Act (2010).

We plan to work with disabled people to better understand their experiences and preferred terminology as part of our work on the disability harmonised standard. Specifically, we will explore the use of the term “disabled” which is particularly divisive. We will also research new questions to help us gather more specific data about the neurodiverse population and align with the social model of disability. We will discuss preferred terminology as part of this work.

More detailed data about mental health

Data users, primarily policy and research professionals from government and non-profit organisations, typically need more detailed mental health data. There is a need to separate data according to mental health condition or illness to understand the effects in society and support decision-making around funding, service provision, and specialist services. For example, this could mean having a standard that includes detailed categories of mental health diagnoses, with higher-level categories which would allow for the analysis of non-disclosive data. Data producers and social researchers also wanted to be able to compare the different mental health groups to each other and stated that currently there was no further information than one tick box option. This would enable researchers to create more strategic plans and assessments for particular mental health impairments.

This aligns with the needs of people giving their data who have expressed a preference to report their physical and mental conditions in a single response. This would reduce the complexity of splitting conditions or illnesses into physical or mental categories and would allow respondents to share information about more than one condition.

Our research demonstrates a clear demand for very detailed data. This includes mental health conditions by type and severity, as well as information about how they interact with physical conditions. These needs will be investigated further as we consider the development of a harmonised standard for mental health.

Mental health of specific populations

Government analysts are mainly looking for mental health data on under-represented groups such as children and people in institutionalised settings, like care home residents and prisoners. Analysts need this data to:

  • ensure they are being inclusive in their findings and statistics
  • improve the identification of challenges that these people may face

Data users also showed a need for data linkage opportunities. For example, they suggested creating datasets that can be used to analyse mental health statistics together with personal characteristics such as education, income, sex, ethnicity, and other determinants of health. This would help get a wider understanding of mental health data and understand the influencing factors around it.

Policy advisors also asked for more information to inform their research appropriately and target services. This included the need to capture the lived experience of having a mental health condition and reality of being disabled by it. As part of our disability workstream we will be researching the lived experience of both disabled and non-disabled populations.


Our initial research and engagement shows there is a need for improved mental health data.

The current harmonised standards offer a limited overview of mental health that captures wellbeing at one end of the spectrum, and mental health as a disability at the other. These standards were not designed to gather data about mental health specifically. This means they do not consider the complexities of mental health, such as the diversity of mental health conditions or the fact that mental health can fluctuate.

We worked with a wide range of stakeholders to produce this review. We also used existing evidence and did our own research. We discussed a wide range of user needs and highlighted potential improvements to the relevant harmonised standards. There were several recurring themes in our work, including the need for inclusivity and granularity of data. Inclusivity includes the need to capture information about specific groups, such as people in institutionalised settings, and to consider how mental health can intersect with disability, substance misuse, and neurodiversity. In terms of granularity, stakeholders requested data according to the severity and type of mental health condition, with the ability to aggregate categories for the purpose of analysis.

What happens next

We have developed some short (spring 2023), medium (2023), and long-term (2024) plans based on our findings from our disability and mental health research. Our full plan is provided as part of our disability review. There are several parts of our plan that are relevant to mental health.

In the medium term (2023), we will:

  • amend and test the guidance on the impairment response option for “socially or behaviourally”
  • design and test guidance for the impairment response options which currently lack it – this includes the mental health response option
  • further explore the use of the term “disabled” and the implications this may have
  • engage with disabled people to better understand their experience, and engage with people who are not disabled to better understand how they answer existing questions

In the longer term, we will design new questions to meet identified gaps. This includes creating a question to count the neurodiverse population and account for diversity within this population.

We will work with stakeholders once this work has been completed to identify data gaps specifically related to mental health. We aim to conduct a thorough research programme with topic experts across government and the devolved administrations. This will help us create a high-quality mental health standard that meets user needs.

Whilst our work is ongoing, we encourage users to continue to use the existing standards for disability and wellbeing, as well as the harmonisation guidance for mental health. Despite their limitations, these standards have been developed as a result of thorough research and testing and are likely to outperform others that have not been through such a process.

Contact us

Government Statistical Service (GSS) Harmonisation Team

We are always interested in hearing from users so we can develop our work. If you would like to contact us, please email the team at

You can find more information about Harmonisation on our webpage.

Further information

Health harmonised standards

Other related health harmonisation standards and guidance that may be useful:

  • If you would like us to get in touch with you then please leave your contact details or email directly.
  • This field is for validation purposes and should be left unchanged.