Camden & Islington Clinical Record Interactive Search (CRIS) Linkage with HES/Mortality Data

This agreement aims to link HES and Mortality data from NHS England with the Camden & Islington NHS Foundation Trust (C&I) Clinical Record Interactive Search (CRIS) Research Database for the purpose of research in the public interest.

Camden & Islington NHS Foundation Trust (C&I) is a large mental healthcare provider serving a geographic catchment area of two inner-city London boroughs, and approximately 470,000 residents. Based on social deprivation scores of 326 local authorities in England, Camden is the 74th and Islington is the 14th most deprived local authority. The variation in the levels of deprivation within both boroughs is large, highlighting the inequalities between different population groups and places. Within Camden there are areas that are within the top 10% most deprived areas in England and areas that are in the 20% least deprived. C&I provides mental health and substance misuse services to people living in Camden and Islington, substance misuse services to Westminster, and a substance misuse and psychological therapies service to residents in Kingston. The Trust has two inpatient facilities, at Highgate Mental Health Centre and St Pancras Hospital, as well as community-based services throughout the London boroughs of Camden and Islington. The Trust provides services for adults of working age, adults with learning difficulties, and older people in community or inpatient settings.

The objective of the data collection is to create a research resource to be used for research projects aiming to investigate physical health outcomes (including mortality) and receipt of health care in people with mental and behavioural health disorders attending secondary mental health care services provided by C&I.

The proposed linkage would significantly increase high-quality research outputs that examine the interface between mental and physical health. There is increasing emphasis in the health inequalities experienced by individuals diagnosed with severe mental illness (SMI), commonly defined as schizophrenia, bipolar disorder, schizoaffective disorder and other non-organic psychotic illnesses. These individuals have been found to have a reduced life expectancy of up to 20 years. What is less clear is how other extremely disabling psychiatric disorders, such as severe depression, post-traumatic stress disorder and personality disorders compare in terms of premature mortality, self-harm and physical health comorbidities. By linking HES and ONS mortality data with C&I CRIS data the study team will explore and quantify this currently under-researched health disparity. The study team will focus specifically on commonly occurring comorbidities and adverse complications such as; cardiovascular, respiratory, cancer, liver disease and self-harm.

This data linkage is supported by service user members of the C&I CRIS Oversight Committee, along with local patient and public involvement (PPI) groups the study team have consulted. People using secondary mental health services are rightfully concerned about their risk of premature mortality and morbidity, this linkage has the potential to answer outstanding questions.

Routine recording of Electronic Health Records (EHR)s at C&I commenced in mid-2008 using RiO, an electronic patient record system. RiO contains a comprehensive, longitudinal record of all clinical information recorded throughout patients’ contacts with Trust services, including socio-demographic information, dates and other details of referrals and admissions, detailed clinical assessments, care plans and standardized assessment forms. The record consists of both structured fields (such as dates and pick-lists) and unstructured free text (including progress notes and correspondence). The CRIS tool, developed by South London and Maudsley NHS Foundation Trust (SLaM) Biomedical Research Cluster (BRC) to extract information from their bespoke electronic Patient Journey System (PJS), consists of a series of data-processing pipelines which both structure and de-identify fields in the electronic patient record, rendering effectively anonymized data from the full clinical record available at the researcher interface. The system allows researchers to search against any combination of structured and unstructured fields that exists in the database. Users then specify the precise fields they want returned (such as specific diagnostic codes, demographic information and/or a particular text string in a clinical assessment).

University College London (UCL) is C&I’s long-standing research partner in clinical research and this is reflected in the development and operation of the C&I CRIS Research Database. The C&I CRIS Research Database administrator is formally employed with University College London with a substantive honorary research contract with C&I. The C&I CRIS Research Database clinical academic lead holds an academic appointment with UCL in the Division of Psychiatry and a consultant psychiatrist with C&I.

The C&I CRIS Research Database employs the same security model as that developed by SLaM to address the legal and ethical considerations attendant upon the use of confidential health data. Authorized researchers are provided with regulated access to anonymized information extracted from electronic patient records. The Research Database is used to support epidemiological and population-based research using only anonymized data, for which no patient consent is necessary though patients can opt out entirely if they choose.

The data subjects are individuals who have received treatment from the Trust between 2012/13 and 2021/22 and some treated earlier where records existed and could be migrated AND who have not notified C&I that they wish to opt out of having their data collected and/or linked.

C&I, the Data Controller, will carry out all necessary duties for the processing, supply and hosting of the distinct, C&I CRIS research database.

Specifically; the linkage activity, i.e. sending of Patient Identifers and receiving of HES and mortality attribute data, will be conducted by C&I and the C&I CRIS - HES and mortality linked data will stored within a C&I specific secure area. All reasonable security steps have been taken to protect data held by C&I, for example no standalone devices are used and no data is permitted to be stored on the local drives. Furthermore, security measures designed to protect data from being saved outside the C&I firewall are in place. Data cannot be accessed directly by researchers wishing to use linked data, this means that researchers will only have access to project specific extracts of the C&I-HES/ mortality data. Extractions will be carried out within the C&I secure area by C&I CRIS staff. If appropriate, support is provided to external researchers who wish to access the linked data to help them fulfil the necessary requirements to gain approved status.

Access to the C&I CRIS research database is limited to the C&I research database administrator and approved research users at C&I, only for research projects which are approved by the C&I research database oversight committee. Access can only be gained via the C&I network. This includes all data validation and quality checks for pseudonymisation which are conducted by C&I staff following data processing. All external researchers with no contractual arrangements with C&I are required to obtain a Research Passport and Honorary Research Contract prior to project approval. Honorary Research Contracts must be signed by approved research users, their substantive employers, and C&I with wording that the employee will be subject to their substantive employer's disciplinary process if they do anything that they shouldn't with the data. Researchers only have access to pseudonymised linked NHS England Data. All research projects are carried out within the C&I network and the linked data remain within the C&I NHS firewall at all times (as is the security model requirement for all analyses of C&I data, regardless of data linkage).

The C&I CRIS Oversight Committee will consider research proposals to use the linked dataset. The Oversight Committee includes research and development governance, Caldicott/ information governance, technical, clinical and service user representation. The Oversight Committee will be responsible for granting or denying approval for all applications to use the linked data. A key consideration of this panel will be to decide if a project poses an increased risk of providing de-anonymised results due to anticipated small cell sizes. Where the panel envisages this to be likely, additional reassurances will be asked for from the applicant and amendments to the application form may be requested. Any successful application will be provided with a bespoke dataset, i.e. one with only the specific variables identified in the application as necessary for the planned analysis. Speculative studies (“data dredging”) will not be permitted.

Research requests vary from year to year. Historically, CRIS has hosted between 5-12 research projects per year from MSc and PhD students at UCL and academic clinical staff at Camden & Islington NHS Foundation Trust. This number has risen in recent years, given the increasing profile of the CRIS research database as a research platform. The study team expect this number to increase following the successful linkage with HES-ONS data which will facilitate more robust, longitudinal analyses. The study team foresee an increase in research projects to 10-18 projects per year. They expect projects pertaining to their established areas of expertise in: severe mental illness, suicide and suicidality, substance use disorders, psychosis, eating disorders, and personality disorders. Understanding the physical health comorbidities of these patients, as well as the causes of mortality, are crucial to research which will elucidate the risk factors and potential interventions to improve care for these individuals.

C&I will only grant access to HES data as part of a linked dataset comprising a minimum of HES and CRIS data (i.e. not for analysis of HES data alone). Broadly, the studies using the linkage have adopted the following designs:

1. Investigations carried out on HES data from the C&I catchment, identifying a HES-derived outcome and comparing its occurrence between people with/without a given diagnosed mental and behavioural health disorders in order to derive standardised morbidity ratios (for example, some current research investigating respiratory disease admissions in people with learning disability compared to the local population);

2. Investigations restricted to people with a given HES-derived outcome and comparing subsequent events between people with/without a given diagnosed mental and behavioural health disorders (for example, further analyses of people with/without a learning disability who have a respiratory disease admission, comparing duration of hospitalisation and risk of readmission between the two groups);

3. Investigations restricted to people with a given diagnosed mental and behavioural health disorders investigating one or more HES-derived outcomes in relation to C&I-derived information (for example, investigating the relationship between mental health symptom profiles and physical health events in people with severe mental illness);

4. Investigations primarily carried out using C&I data, where HES-derived information is used to provide supplementary information (for example, the ability to adjust for serious physical illness in a number of analyses). This includes the use of mental healthcare data contained on HES for residents in the C&I catchment to capture mental health service use by providers other than C&I (e.g. out-of-catchment hospitalisations);

5. Investigations primarily carried out using C&I data where a HES outcome is used to define the sample (for example, a series of analyses investigating medication and health outcomes before and after childbirth in women with pre-existing severe mental illness).

The SLaM NHS Foundation Trust will act as the Data Processor for this linkage insofar that SlaM hosts data on behalf of C&I according to the DPA. All reasonable security steps have been taken to protect data hosted by SlaM, for example, no standalone devices are used and no data is permitted to be stored on the local drives. Furthermore, security measures designed to protect data from being saved outside the SlaM firewall are in place. Data cannot be accessed directly by researchers wishing to use linked data, this means that researchers will only have access to project-specific extracts of the C&I-HES/ mortality data.

The C&I Database itself will be hosted by SLaM, and secure data transfer is subject to a Data Processing Agreement (DPA). Whilst the data will be hosted by SLaM, the data will not be generally accessible to SLaM staff other than the necessary IT support personnel as detailed in the DPA. C&I data will not be available to users of any SLaM Research Databases (including the CRIS based ‘sister’ database operating there), and C&I retains ownership of the C&I data at all times.

The original terms of this data linkage were originally inclusive of data from 2012-18, however, delays as a result of COVID-19 would have meant the data would have been up to four years out of date once final approvals had been received from applicable regulatory bodies. Using such dated data may lead to erroneous research outputs where advancements in care or changes over time from 2018 onwards may not be reflected in historical data. Consequently, the HRA CAG approved on 5 August 2022 (CAG reference 20/CAG0069; REC reference 19/EE/0210) an extension of the cohort to include data up to and including 2022 to ensure that data was current, given that this data linkage was a one-time linkage without possibility for further renewals.

The primary output of the linkage is the production and maintenance of a research resource for the purpose of use in informative research analyses for publication in peer-reviewed journals and other standard routes of academic dissemination (e.g. conference presentations).

All secondary outputs (whether tables or visuals) will only include aggregated data suppressed according to the HES analysis guide. Outputs must also comply with the UK Data Service’s Handbook on Statistical Disclosure Control for Outputs including the rules around secondary suppression where applicable.

The study team expect that a minimum of two research papers would be published per year from 2024-2027 using the proposed data linkages. Examples of papers planned for publication include:

1. Co-morbidity and premature mortality in non-severe mental illness patients. Severe mental illness (SMI) is commonly defined as schizophrenia, bipolar disorder and other psychotic disorders. Published research has explored prevalence and incidence of physical co-morbidities among patients with SMI, but less is known about these comorbidities for other mental diagnostic groups such as depression, PTSD, and personality disorder. The paper will explore if common physical co-morbidities such as cardiovascular disease, diabetes, severe asthma, chronic obstructive pulmonary disease, and cancer are also overrepresented in non-SMI patient populations using secondary care mental health services. By linking C&I CRIS with HES-ONS mortality the study team will compare incidence and prevalence between SMI and non SMI patients, using both published population control estimates as well as matched HES-ONS mortality control data drawn from Camden and Islington boroughs, exploring both at co-morbidity and mortality from the physical conditions known to be overrepresented in SMI. The study team will also explore predictors of mortality and co-morbidity including demographic, social and clinical factors.

2. Mental illness - Pathways to physical healthcare. In order to reduce the physical health inequality experienced by patients suffering from mental health issues, availability and quality of treatment offered prior to and after receiving a comorbid physical diagnosis is of high importance. For example, are patients with mental health problems less likely to receive coronary angioplasty and stents? Are they less likely to receive transplants? Through the linkage of CRIS with HES-ONS mortality, the study team will be able to map out which treatments were offered to patients, and explore how such treatments (or lack of) impacted outcomes of physical and mental health and/or if treatments can be related to cause-specific mortality. The study team will also explore if there is inequality between mental health diagnostic groups in terms of treatment offers and pathways to physical healthcare, and the degree to which social deprivation, ethnicity, diagnosis, medication, age and sex explain any disparities.

All the potential uses of the linked data fall within the stated primary purpose of investigating physical health in people with diagnosed mental and behavioural health disorders. The data being requested will only be used for the purpose described. Any proposed changes will be submitted to NHS England for amendment and approval before implementation.

Publication targets will clearly depend on the nature of individual findings and the potential audience envisaged. Where possible, the study team will target general medical and/or public health journals with a broad audience, because analyses are likely to cross disciplines; however, they will also consider specialist journals within the mental health field as well as the individual medical specialties implicated. Dissemination at national and international conferences will adopt a similar strategy of aiming for as broad as possible a reach. They will include mental health focused meetings such as the Royal College of Psychiatrists and European Psychiatric Association congresses, and psychiatric epidemiology meetings such as the International Federation of Psychiatric Epidemiology (IFPE) but they will also seek presentations at medical specialty conferences where results have relevance to those audiences, as well as meetings where commissioners are likely to be represented.

For each application received, the CRIS Oversight Committee, considers the study design and advises on optimisation of benefits. The CRIS Oversight Committee also has a responsibility for publicity and dissemination of findings to relevant parties, media and patient groups.

Patient and public involvement (PPI) is central to the operation and ethical approval for the C&I CRIS research database. There are three service users on the C&I CRIS research database oversight committee. All applications for projects to access the C&I research database are reviewed by a service user.

Separately, the study team also have a Data Science PPI group (chaired by the McPin Foundation – a charity integrating experts by experience into research www.mcpin.org) who comment on and contribute to the design, conduct, and dissemination of studies using the C&I CRIS research database. While this group does not review applications for use of the C&I research database, their advisory role provides important guidance and insights into academic research, including ensuring that research questions are appropriately framed and that research findings are meaningfully contextualised. This PPI group continues to meet regularly to offer their guidance to C&I research database users. The McPin Foundation are not considered Data Controllers as they have no say over the data processing methodology. They provide facilitation support to the separate Data Science PPI group given their expertise in integrating lived experience into academic research. The Data Science PPI group provides important insights and guidance but do not regulate access to CRIS data – that is the remit of the CRIS Governance Board which also includes service user/carer representation.

All patient-facing materials within the Trust have been updated to reflect the data linkage. Research papers and publications using the linked data have not yet been realised. The linked data was received in 2023 Q4 and the team have been undertaking routine quality checks. As well, the team did not advertise the availability of the linked data to researchers until 2024 Q1.

Work is currently being undertaken to enable safe access to the production and maintenance of a research resource as outlined in the original application. It is expected that the first applications to use this linked data will be approved in 2024 Q1 with research publications published in later 2024.

Protocols to undertake the two specific analyses in our application (comorbidity and premature mortality in non-severe mental illness patients and; pathways to physical health care) are in the process of being drafted and applications submitted for review. The team expect that formal analysis for these two papers will commence no later than 2024 Q2 and final outputs to be published by 2025 Q1.

This agreement aims to link HES and Mortality data from NHS England with the Camden & Islington NHS Foundation Trust (C&I) Clinical Record Interactive Search (CRIS) Research Database for the purpose of research in the public interest.

Camden & Islington NHS Foundation Trust (C&I) is a large mental healthcare provider serving a geographic catchment area of two inner-city London boroughs, and approximately 470,000 residents. Based on social deprivation scores of 326 local authorities in England, Camden is the 74th and Islington is the 14th most deprived local authority. The variation in the levels of deprivation within both boroughs is large, highlighting the inequalities between different population groups and places. Within Camden there are areas that are within the top 10% most deprived areas in England and areas that are in the 20% least deprived. C&I provides mental health and substance misuse services to people living in Camden and Islington, substance misuse services to Westminster, and a substance misuse and psychological therapies service to residents in Kingston. The Trust has two inpatient facilities, at Highgate Mental Health Centre and St Pancras Hospital, as well as community-based services throughout the London boroughs of Camden and Islington. The Trust provides services for adults of working age, adults with learning difficulties, and older people in community or inpatient settings.

The objective of the data collection is to create a research resource to be used for research projects aiming to investigate physical health outcomes (including mortality) and receipt of health care in people with mental and behavioural health disorders attending secondary mental health care services provided by C&I.

The proposed linkage would significantly increase high-quality research outputs that examine the interface between mental and physical health. There is increasing emphasis in the health inequalities experienced by individuals diagnosed with severe mental illness (SMI), commonly defined as schizophrenia, bipolar disorder, schizoaffective disorder and other non-organic psychotic illnesses. These individuals have been found to have a reduced life expectancy of up to 20 years. What is less clear is how other extremely disabling psychiatric disorders, such as severe depression, post-traumatic stress disorder and personality disorders compare in terms of premature mortality, self-harm and physical health comorbidities. By linking HES and ONS mortality data with C&I CRIS data the study team will explore and quantify this currently under-researched health disparity. The study team will focus specifically on commonly occurring comorbidities and adverse complications such as; cardiovascular, respiratory, cancer, liver disease and self-harm.

This data linkage is supported by service user members of the C&I CRIS Oversight Committee, along with local patient and public involvement (PPI) groups the study team have consulted. People using secondary mental health services are rightfully concerned about their risk of premature mortality and morbidity, this linkage has the potential to answer outstanding questions.

Routine recording of Electronic Health Records (EHR)s at C&I commenced in mid-2008 using RiO, an electronic patient record system. RiO contains a comprehensive, longitudinal record of all clinical information recorded throughout patients’ contacts with Trust services, including socio-demographic information, dates and other details of referrals and admissions, detailed clinical assessments, care plans and standardized assessment forms. The record consists of both structured fields (such as dates and pick-lists) and unstructured free text (including progress notes and correspondence). The CRIS tool, developed by South London and Maudsley NHS Foundation Trust (SLaM) Biomedical Research Cluster (BRC) to extract information from their bespoke electronic Patient Journey System (PJS), consists of a series of data-processing pipelines which both structure and de-identify fields in the electronic patient record, rendering effectively anonymized data from the full clinical record available at the researcher interface. The system allows researchers to search against any combination of structured and unstructured fields that exists in the database. Users then specify the precise fields they want returned (such as specific diagnostic codes, demographic information and/or a particular text string in a clinical assessment).

University College London (UCL) is C&I’s long-standing research partner in clinical research and this is reflected in the development and operation of the C&I CRIS Research Database. The C&I CRIS Research Database administrator is formally employed with University College London with a substantive honorary research contract with C&I. The C&I CRIS Research Database clinical academic lead holds an academic appointment with UCL in the Division of Psychiatry and a consultant psychiatrist with C&I.

The C&I CRIS Research Database employs the same security model as that developed by SLaM to address the legal and ethical considerations attendant upon the use of confidential health data. Authorized researchers are provided with regulated access to anonymized information extracted from electronic patient records. The Research Database is used to support epidemiological and population-based research using only anonymized data, for which no patient consent is necessary though patients can opt out entirely if they choose.

The data subjects are individuals who have received treatment from the Trust between 2012/13 and 2021/22 and some treated earlier where records existed and could be migrated AND who have not notified C&I that they wish to opt out of having their data collected and/or linked.

C&I, the Data Controller, will carry out all necessary duties for the processing, supply and hosting of the distinct, C&I CRIS research database.

Specifically; the linkage activity, i.e. sending of Patient Identifers and receiving of HES and mortality attribute data, will be conducted by C&I and the C&I CRIS - HES and mortality linked data will stored within a C&I specific secure area. All reasonable security steps have been taken to protect data held by C&I, for example no standalone devices are used and no data is permitted to be stored on the local drives. Furthermore, security measures designed to protect data from being saved outside the C&I firewall are in place. Data cannot be accessed directly by researchers wishing to use linked data, this means that researchers will only have access to project specific extracts of the C&I-HES/ mortality data. Extractions will be carried out within the C&I secure area by C&I CRIS staff. If appropriate, support is provided to external researchers who wish to access the linked data to help them fulfil the necessary requirements to gain approved status.

Access to the C&I CRIS research database is limited to the C&I research database administrator and approved research users at C&I, only for research projects which are approved by the C&I research database oversight committee. Access can only be gained via the C&I network. This includes all data validation and quality checks for pseudonymisation which are conducted by C&I staff following data processing. All external researchers with no contractual arrangements with C&I are required to obtain a Research Passport and Honorary Research Contract prior to project approval. Honorary Research Contracts must be signed by approved research users, their substantive employers, and C&I with wording that the employee will be subject to their substantive employer's disciplinary process if they do anything that they shouldn't with the data. Researchers only have access to pseudonymised linked NHS England Data. All research projects are carried out within the C&I network and the linked data remain within the C&I NHS firewall at all times (as is the security model requirement for all analyses of C&I data, regardless of data linkage).

The C&I CRIS Oversight Committee will consider research proposals to use the linked dataset. The Oversight Committee includes research and development governance, Caldicott/ information governance, technical, clinical and service user representation. The Oversight Committee will be responsible for granting or denying approval for all applications to use the linked data. A key consideration of this panel will be to decide if a project poses an increased risk of providing de-anonymised results due to anticipated small cell sizes. Where the panel envisages this to be likely, additional reassurances will be asked for from the applicant and amendments to the application form may be requested. Any successful application will be provided with a bespoke dataset, i.e. one with only the specific variables identified in the application as necessary for the planned analysis. Speculative studies (“data dredging”) will not be permitted.

Research requests vary from year to year. Historically, CRIS has hosted between 5-12 research projects per year from MSc and PhD students at UCL and academic clinical staff at Camden & Islington NHS Foundation Trust. This number has risen in recent years, given the increasing profile of the CRIS research database as a research platform. The study team expect this number to increase following the successful linkage with HES-ONS data which will facilitate more robust, longitudinal analyses. The study team foresee an increase in research projects to 10-18 projects per year. They expect projects pertaining to their established areas of expertise in: severe mental illness, suicide and suicidality, substance use disorders, psychosis, eating disorders, and personality disorders. Understanding the physical health comorbidities of these patients, as well as the causes of mortality, are crucial to research which will elucidate the risk factors and potential interventions to improve care for these individuals.

C&I will only grant access to HES data as part of a linked dataset comprising a minimum of HES and CRIS data (i.e. not for analysis of HES data alone). Broadly, the studies using the linkage have adopted the following designs:

1. Investigations carried out on HES data from the C&I catchment, identifying a HES-derived outcome and comparing its occurrence between people with/without a given diagnosed mental and behavioural health disorders in order to derive standardised morbidity ratios (for example, some current research investigating respiratory disease admissions in people with learning disability compared to the local population);

2. Investigations restricted to people with a given HES-derived outcome and comparing subsequent events between people with/without a given diagnosed mental and behavioural health disorders (for example, further analyses of people with/without a learning disability who have a respiratory disease admission, comparing duration of hospitalisation and risk of readmission between the two groups);

3. Investigations restricted to people with a given diagnosed mental and behavioural health disorders investigating one or more HES-derived outcomes in relation to C&I-derived information (for example, investigating the relationship between mental health symptom profiles and physical health events in people with severe mental illness);

4. Investigations primarily carried out using C&I data, where HES-derived information is used to provide supplementary information (for example, the ability to adjust for serious physical illness in a number of analyses). This includes the use of mental healthcare data contained on HES for residents in the C&I catchment to capture mental health service use by providers other than C&I (e.g. out-of-catchment hospitalisations);

5. Investigations primarily carried out using C&I data where a HES outcome is used to define the sample (for example, a series of analyses investigating medication and health outcomes before and after childbirth in women with pre-existing severe mental illness).

The SLaM NHS Foundation Trust will act as the Data Processor for this linkage insofar that SlaM hosts data on behalf of C&I according to the DPA. All reasonable security steps have been taken to protect data hosted by SlaM, for example, no standalone devices are used and no data is permitted to be stored on the local drives. Furthermore, security measures designed to protect data from being saved outside the SlaM firewall are in place. Data cannot be accessed directly by researchers wishing to use linked data, this means that researchers will only have access to project-specific extracts of the C&I-HES/ mortality data.

The C&I Database itself will be hosted by SLaM, and secure data transfer is subject to a Data Processing Agreement (DPA). Whilst the data will be hosted by SLaM, the data will not be generally accessible to SLaM staff other than the necessary IT support personnel as detailed in the DPA. C&I data will not be available to users of any SLaM Research Databases (including the CRIS based ‘sister’ database operating there), and C&I retains ownership of the C&I data at all times.

The original terms of this data linkage were originally inclusive of data from 2012-18, however, delays as a result of COVID-19 would have meant the data would have been up to four years out of date once final approvals had been received from applicable regulatory bodies. Using such dated data may lead to erroneous research outputs where advancements in care or changes over time from 2018 onwards may not be reflected in historical data. Consequently, the HRA CAG approved on 5 August 2022 (CAG reference 20/CAG0069; REC reference 19/EE/0210) an extension of the cohort to include data up to and including 2022 to ensure that data was current, given that this data linkage was a one-time linkage without possibility for further renewals.

The primary output of the linkage is the production and maintenance of a research resource for the purpose of use in informative research analyses for publication in peer-reviewed journals and other standard routes of academic dissemination (e.g. conference presentations).

All secondary outputs (whether tables or visuals) will only include aggregated data suppressed according to the HES analysis guide. Outputs must also comply with the UK Data Service’s Handbook on Statistical Disclosure Control for Outputs including the rules around secondary suppression where applicable.

The study team expect that a minimum of two research papers would be published per year using the proposed data linkages. Examples of papers planned for publication include:

1. Co-morbidity and premature mortality in non-severe mental illness patients. Severe mental illness (SMI) is commonly defined as schizophrenia, bipolar disorder and other psychotic disorders. Published research has explored prevalence and incidence of physical co-morbidities among patients with SMI, but less is known about these comorbidities for other mental diagnostic groups such as depression, PTSD, and personality disorder. The paper will explore if common physical co-morbidities such as cardiovascular disease, diabetes, severe asthma, chronic obstructive pulmonary disease, and cancer are also overrepresented in non-SMI patient populations using secondary care mental health services. By linking C&I CRIS with HES-ONS mortality the study team will compare incidence and prevalence between SMI and non SMI patients, using both published population control estimates as well as matched HES-ONS mortality control data drawn from Camden and Islington boroughs, exploring both at co-morbidity and mortality from the physical conditions known to be overrepresented in SMI. The study team will also explore predictors of mortality and co-morbidity including demographic, social and clinical factors.

2. Mental illness - Pathways to physical healthcare. In order to reduce the physical health inequality experienced by patients suffering from mental health issues, availability and quality of treatment offered prior to and after receiving a comorbid physical diagnosis is of high importance. For example, are patients with mental health problems less likely to receive coronary angioplasty and stents? Are they less likely to receive transplants? Through the linkage of CRIS with HES-ONS mortality, the study team will be able to map out which treatments were offered to patients, and explore how such treatments (or lack of) impacted outcomes of physical and mental health and/or if treatments can be related to cause-specific mortality. The study team will also explore if there is inequality between mental health diagnostic groups in terms of treatment offers and pathways to physical healthcare, and the degree to which social deprivation, ethnicity, diagnosis, medication, age and sex explain any disparities.

All the potential uses of the linked data fall within the stated primary purpose of investigating physical health in people with diagnosed mental and behavioural health disorders. The data being requested will only be used for the purpose described. Any proposed changes will be submitted to NHS England for amendment and approval before implementation.

Publication targets will clearly depend on the nature of individual findings and the potential audience envisaged. Where possible, the study team will target general medical and/or public health journals with a broad audience, because analyses are likely to cross disciplines; however, they will also consider specialist journals within the mental health field as well as the individual medical specialties implicated. Dissemination at national and international conferences will adopt a similar strategy of aiming for as broad as possible a reach. They will include mental health focused meetings such as the Royal College of Psychiatrists and European Psychiatric Association congresses, and psychiatric epidemiology meetings such as the International Federation of Psychiatric Epidemiology (IFPE) but they will also seek presentations at medical specialty conferences where results have relevance to those audiences, as well as meetings where commissioners are likely to be represented.

For each application received, the CRIS Oversight Committee, considers the study design and advises on optimisation of benefits. The CRIS Oversight Committee also has a responsibility for publicity and dissemination of findings to relevant parties, media and patient groups.

Patient and public involvement (PPI) is central to the operation and ethical approval for the C&I CRIS research database. There are three service users on the C&I CRIS research database oversight committee. All applications for projects to access the C&I research database are reviewed by a service user.

Separately, the study team also have a Data Science PPI group (chaired by the McPin Foundation – a charity integrating experts by experience into research www.mcpin.org) who comment on and contribute to the design, conduct, and dissemination of studies using the C&I CRIS research database. While this group does not review applications for use of the C&I research database, their advisory role provides important guidance and insights into academic research, including ensuring that research questions are appropriately framed and that research findings are meaningfully contextualised. This PPI group continues to meet regularly to offer their guidance to C&I research database users. The McPin Foundation are not considered Data Controllers as they have no say over the data processing methodology. They provide facilitation support to the separate Data Science PPI group given their expertise in integrating lived experience into academic research. The Data Science PPI group provides important insights and guidance but do not regulate access to CRIS data – that is the remit of the CRIS Governance Board which also includes service user/carer representation.

There have been no yielded benefits from this data linkage yet as the Data Controller has not had a chance to work on the linked data to date.

This agreement aims to link HES and Mortality data from NHS Digital with the Camden & Islington NHS Foundation Trust (C&I) Clinical Record Interactive Search (CRIS) Research Database for the purpose of research in the public interest.

Camden & Islington NHS Foundation Trust (C&I) is a large mental healthcare provider serving a geographic catchment area of two inner-city London boroughs, and approximately 470,000 residents. Based on social deprivation scores of 326 local authorities in England, Camden is the 74th and Islington is the 14th most deprived local authority. The variation in the levels of deprivation within both boroughs is large, highlighting the inequalities between different population groups and places. Within Camden there are areas that are within the top 10% most deprived areas in England and areas that are in the 20% least deprived. C&I provides mental health and substance misuse services to people living in Camden and Islington, substance misuse services to Westminster, and a substance misuse and psychological therapies service to residents in Kingston. The Trust has two inpatient facilities, at Highgate Mental Health Centre and St Pancras Hospital, as well as community-based services throughout the London boroughs of Camden and Islington. The Trust provides services for adults of working age, adults with learning difficulties, and older people in community or inpatient settings.

The objective of the data collection is to create a research resource to be used for research projects aiming to investigate physical health outcomes (including mortality) and receipt of health care in people with mental and behavioural health disorders attending secondary mental health care services provided by C&I.

The proposed linkage would significantly increase high quality research outputs that examine the interface between mental and physical health. There is increasing emphasis in the health inequalities experienced by individuals diagnosed with severe mental illness (SMI), commonly defined as schizophrenia, bipolar disorder, schizoaffective disorder and other non-organic psychotic illnesses. These individuals have been found to have a reduced life expectancy of up to 20 years. What is less clear is how other extremely disabling psychiatric disorders, such as severe depression, post-traumatic stress disorder and personality disorders compare in terms of premature mortality, self-harm and physical health comorbidities. By linking HES and ONS mortality data with C&I CRIS data the study team will explore and quantify this currently under-researched health disparity. The study team will focus specifically on commonly occurring comorbidities and adverse complications such as; cardiovascular, respiratory, cancer, liver disease and self-harm.

This data linkage is supported by service user members of the C&I CRIS Oversight Committee, along with local patient and public involvement (PPI) groups the study team have consulted. People using secondary mental health services are rightfully concerned about their risk of premature mortality and morbidity, this linkage has the potential to answer outstanding questions.

Routine recording of Electronic Health Records (EHR)s at C&I commenced in mid-2008 using RiO, an electronic patient record system. RiO contains a comprehensive, longitudinal record of all clinical information recorded throughout patients’ contacts with Trust services, including socio-demographic information, dates and other details of referrals and admissions, detailed clinical assessments, care plans and standardized assessment forms. The record consists of both structured fields (such as dates and pick-lists) and unstructured free text (including progress notes and correspondence). The CRIS tool, developed by South London and Madusley NHS Foundation Trust (SLaM) Biomedical Research Cluster (BRC) to extract information from their bespoke electronic Patient Journey System (PJS), consists of a series of data-processing pipelines which both structure and de-identify fields in the electronic patient record, rendering effectively anonymized data from the full clinical record available at the researcher interface. The system allows researchers to search against any combination of structured and unstructured fields that exists in the database. Users then specify the precise fields they want returned (such as specific diagnostic codes, demographic information and/or a particular text string in a clinical assessment).

University College London (UCL) is C&I’s long-standing research partner in clinical research and this is reflected in the development and operation of the C&I CRIS Research Database. The C&I CRIS Research Database administrator is formally employed with University College London with a substantive honorary research contract with C&I. The C&I CRIS Research Database clinical academic lead holds an academic appointment with UCL in the Division of Psychiatry and a consultant psychiatrist with C&I.

The C&I CRIS Research Database employs the same security model as that developed by SLaM to address the legal and ethical considerations attendant upon the use of confidential health data. Authorized researchers are provided with regulated access to anonymized information extracted from electronic patient records. The Research Database is used to support epidemiological and population-based research using only anonymized data, for which no patient consent is necessary though patients can opt out entirely if they choose.

The data subjects are individuals who:

(i) have received treatment from the Trust between 2012/13 and 2017/18 and some treated earlier where records existed and could be migrated AND who have not notified C&I that they wish to opt out of having their data collected and/or linked, and/or

(ii) individuals who are or have been resident within the London boroughs of Camden & Islington geographic catchment between 2012/13 and 2017/18 and attended hospital for any reason whilst resident in that catchment area.

C&I, the Data Controller, have contracted with the SLaM Clinical Data Linkage Service (CDLS) to carry out certain necessary duties for the processing, supply and hosting of the distinct, C&I CRIS research database. The South London and Maudsley NHS Foundation Trust will act as the Data Processor for the application.

Specifically; the linkage activity, i.e. sending of Patient Identifers and receiving of HES and mortality attribute data, will be conducted by the SLaM Clinical Data Linkage Service (CDLS) and the C&I CRIS - HES and mortality linked data will stored and hosted by SLaM within the C&I specific secure area. The SLaM CDLS provides data processing services (linkage, storage, and data extraction) to external collaborators. All reasonable security steps have been taken to protect data held by SLaM, for example no standalone devices are used and no data is permitted to be stored on the local drives. Furthermore, security measures designed to protect data from being saved outside the SLaM firewall are in place. Data cannot be accessed directly by researchers wishing to use linked data, this means that researchers will only have access to project specific extracts of the C&I-HES/ mortality data. Extractions will be carried out within the C&I secure area within the SLaM firewall by C&I CRIS staff. If appropriate, support is provided to external researchers who wish to access the linked data to help them fulfil the necessary requirements to gain approved status.

In summary, SLaM CDLS, as a Data Processor, process C&I clinical data on behalf of and under the contractual obligation to C&I. C&I maintains exclusive control over access to the C&I CRIS research database as Data Controller. SLaM CDLS has no ability or permission to access C&I data. This contractual relationship is analogous to that of an NHS Foundation Trust and any third-party data processor and host insofar the NHS Foundation Trust remains the Data Controller while the third-party vendor acts as a Data Processor (for example C&I’s relationship with its electronic health record provider). As defined above, SLaM CDLS process the data, governed by the Data Processing Agreement, to fulfil the terms of its contractual obligation to C&I: to create and maintain the C&I CRIS research database. The lawful basis being relied upon to support the flow of confidential patient information from Camden and Islington NHS Foundation Trust (C&I) to SLaM to facilitate the creation of the C&I CRIS Database and this data linkage are: Article 6(1)(e) and Article 9(2)(j) of the General Data Protection Regulation (EU) 2016/679 (GDPR). SLaM are acting as a data processor (as defined in GDPR Article 28) on behalf of C&I who serve as the data controller. Per GDPR Article 28(3), C&I has established a data processing contract and processing agreement for the lawful flow of confidential patient information from C&I to SLaM specifically for the purpose of data processing. SLaM CDLS staff do not have access to the C&I research database.

Access to the C&I CRIS research database is limited to the C&I research database administrator and approved research users at C&I, only for research projects which are approved by the C&I research database oversight committee. Access can only be gained via the C&I network. This includes all data validation and quality checks for pseudonymisation which are conducted by C&I staff following data processing by SLaM CDLS. All external researchers with no contractual arrangements with C&I are required to obtain a Research Passport and Honorary Research Contract prior to project approval. Honorary Research Contracts must be signed by approved research users, their substantive employers, and C&I with wording that the employee will be subject to their substantive employer's disciplinary process if they do anything that they shouldn't with the data. Researchers only have access to pseudonymised linked NHS Digital Data. All research projects are carried out within the C&I network and the linked data remain within the C&I NHS firewall at all times (as is the security model requirement for all analyses of C&I data, regardless of data linkage).

The C&I CRIS Oversight Committee will consider research proposals to use the linked dataset. The Oversight Committee includes research and development governance, Caldicott/ information governance, technical, clinical and service user representation. The Oversight Committee will be responsible for granting or denying approval for all applications to use the linked data. A key consideration of this panel will be to decide if a project poses an increased risk of providing de-anonymised results due to anticipated small cell sizes. Where the panel envisages this to be likely, additional reassurances will be asked for from the applicant and amendments to the application form may be requested. Any successful application will be provided with a bespoke dataset, i.e. one with only the specific variables identified in the application as necessary for the planned analysis. Speculative studies (“data dredging”) will not be permitted.

Research requests vary from year to year. Historically, CRIS has hosted between 5-12 research projects per year from MSc and PhD students at UCL and academic clinical staff at Camden & Islington NHS Foundation Trust. This number has risen in recent years, given the increasing profile of the CRIS research database as a research platform. The study team expect this number to increase following the successful linkage with HES-ONS data which will facilitate more robust, longitudinal analyses. The study team foresee an increase in research projects to 10-18 projects per year. They expect projects pertaining to their established areas of expertise in: severe mental illness, suicide and suicidality, substance use disorders, psychosis, eating disorders, and personality disorders. Understanding the physical health comorbidities of these patients, as well as the causes of mortality, are crucial to research which will elucidate the risk factors and potential interventions to improve care for these individuals.

C&I will only grant access to HES data as part of a linked dataset comprising a minimum of HES and CRIS data (i.e. not for analysis of HES data alone). Broadly, the studies using the linkage have adopted the following designs:

1. Investigations carried out on HES data from the C&I catchment, identifying a HES-derived outcome and comparing its occurrence between people with/without a given diagnosed mental and behavioural health disorders in order to derive standardised morbidity ratios (for example, some current research investigating respiratory disease admissions in people with learning disability compared to the local population);

2. Investigations restricted to people with a given HES-derived outcome and comparing subsequent events between people with/without a given diagnosed mental and behavioural health disorders (for example, further analyses of people with/without a learning disability who have a respiratory disease admission, comparing duration of hospitalisation and risk of readmission between the two groups);

3. Investigations restricted to people with a given diagnosed mental and behavioural health disorders investigating one or more HES-derived outcomes in relation to C&I-derived information (for example, investigating the relationship between mental health symptom profiles and physical health events in people with severe mental illness);

4. Investigations primarily carried out using C&I data, where HES-derived information is used to provide supplementary information (for example, the ability to adjust for serious physical illness in a number of analyses). This includes the use of mental healthcare data contained on HES for residents in the C&I catchment to capture mental health service use by providers other than C&I (e.g. out-of-catchment hospitalisations);

5. Investigations primarily carried out using C&I data where a HES outcome is used to define the sample (for example, a series of analyses investigating medication and health outcomes before and after childbirth in women with pre-existing severe mental illness).

One further planned linkage with Public Health England and the National Cancer Registry is currently under review by Public Health England’s Office for Data Release. This dataset will be stored separately from the proposed HES-ONS-CRIS data linkage and there is no intention or technical ability to link data from the National Cancer Registry to HES-ONS-CRIS linked data.

The primary output of the linkage is the production and maintenance of a research resource for the purpose of use in informative research analyses for publication in peer-reviewed journals and other standard routes of academic dissemination (e.g. conference presentations).

All secondary outputs (whether tables or visuals) will only include aggregated data suppressed according to the HES analysis guide. Outputs must also comply with the UK Data Service’s Handbook on Statistical Disclosure Control for Outputs including the rules around secondary suppression where applicable.

The study team expect that a minimum of two research papers would be published per year using the proposed data linkages. Examples of papers planned for publication include:

1. Co-morbidity and premature mortality in non-severe mental illness patients. Severe mental illness (SMI) is commonly defined as schizophrenia, bipolar disorder and other psychotic disorders. Published research has explored prevalence and incidence of physical co-morbidities among patients with SMI, but less is known about these comorbidities for other mental diagnostic groups such as depression, PTSD, and personality disorder. The paper will explore if common physical co-morbidities such as cardiovascular disease, diabetes, severe asthma, chronic obstructive pulmonary disease, and cancer are also overrepresented in non-SMI patient populations using secondary care mental health services. By linking C&I CRIS with HES-ONS mortality the study team will compare incidence and prevalence between SMI and non SMI patients, using both published population control estimates as well as matched HES-ONS mortality control data drawn from Camden and Islington boroughs, exploring both at co-morbidity and mortality from the physical conditions known to be overrepresented in SMI. The study team will also explore predictors of mortality and co-morbidity including demographic, social and clinical factors.

2. Mental illness - Pathways to physical healthcare. In order to reduce the physical health inequality experienced by patients suffering from mental health issues, availability and quality of treatment offered prior to and after receiving a comorbid physical diagnosis is of high importance. For example, are patients with mental health problems less likely to receive coronary angioplasty and stents? Are they less likely to receive transplants? Through the linkage of CRIS with HES-ONS mortality, the study team will be able to map out which treatments were offered to patients, and explore how such treatments (or lack of) impacted outcomes of physical and mental health and/or if treatments can be related to cause-specific mortality. The study team will also explore if there is inequality between mental health diagnostic groups in terms of treatment offers and pathways to physical healthcare, and the degree to which social deprivation, ethnicity, diagnosis, medication, age and sex explain any disparities.

All the potential uses of the linked data fall within the stated primary purpose of investigating physical health in people with diagnosed mental and behavioural health disorders. The data being requested will only be used for the purpose described. Any proposed changes will be submitted to NHS Digital for amendment and approval before implementation.

Publication targets will clearly depend on the nature of individual findings and the potential audience envisaged. Where possible, the study team will target general medical and/or public health journals with a broad audience, because analyses are likely to cross disciplines; however, they will also consider specialist journals within the mental health field as well as the individual medical specialties implicated. Dissemination at national and international conferences will adopt a similar strategy of aiming for as broad as possible a reach. They will include mental health focused meetings such as the Royal College of Psychiatrists and European Psychiatric Association congresses, and psychiatric epidemiology meetings such as the International Federation of Psychiatric Epidemiology (IFPE) but they will also seek presentations at medical specialty conferences where results have relevance to those audiences, as well as meetings where commissioners are likely to be represented.

For each application received, the CRIS Oversight Committee, considers the study design and advises on optimisation of benefits. The CRIS Oversight Committee also has a responsibility for publicity and dissemination of findings to relevant parties, media and patient groups.

Patient and public involvement (PPI) is central to the operation and ethical approval for the C&I CRIS research database. There are three service users on the C&I CRIS research database oversight committee. All applications for projects to access the C&I research database are reviewed by a service user.

Separately, the study team also have a Data Science PPI group (chaired by the McPin Foundation – a charity integrating experts by experience into research www.mcpin.org) who comment on and contribute to the design, conduct, and dissemination of studies using the C&I CRIS research database. While this group does not review applications for use of the C&I research database, their advisory role provides important guidance and insights into academic research, including ensuring that research questions are appropriately framed and that research findings are meaningfully contextualised. This PPI group continues to meet regularly to offer their guidance to C&I research database users. The McPin Foundation are not considered Data Controllers as they have any say over the data processing methodology. They provide facilitation support to the separate Data Science PPI group given their expertise in integrating lived experience into academic research. The Data Science PPI group provides important insights and guidance but do not regulate access to CRIS data – that is the remit of the CRIS Governance Board which also includes service user/carer representation.

Yielded Benefits is not a requirement for new applications.