10 years of CRIS: a timeline

The Clinical Record Interactive System (CRIS) was the first computer system of its kind to unlock data from the full electronic medical record. Developed by the National Institute for Health Research (NIHR) Biomedical Research Centre at South London and Maudsley NHS Foundation Trust (SLaM) and King’s College London, and with the direct involvement of patients, CRIS provided researchers with regulated access to anonymised information contained in SLaM’s medical records (Europe’s largest mental health provider).

CRIS allows researchers to extract and analyse an unparalleled quantity and quality of data, including detail behind the mental health symptoms people report, the treatments they receive, and the outcomes they experience. The system has stimulated vital research into how diseases develop, identifying people at risk, and determining the most cost-effective treatments—ultimately improving patient care and outcomes.

Professor Robert Stewart, Academic Lead for CRIS and Professor of Psychiatric Epidemiology & Clinical Informatics at King's College London said: “I’m very proud to be part of an initiative like CRIS that has genuinely broken new ground. There’s a lot of talk about the idea of using real-world information to improve real-world services, but rather fewer actual examples of this in action. At its heart, CRIS is an information governance model that enables access to clinical data for research purposes in a way that protects the legal and ethical rights of patients. From the start, patients have been directly involved in its design, implementation, and ongoing oversight, and the feedback has been universally positive.”

Read more about CRIS on our website.

A patient-led oversight committee was established to ensure that all research applications comply with ethical and legal guidelines, and to review, monitor, and audit all research conducted using CRIS. It is chaired by a service user and supported by the SLaM Trust’s Caldicott Guardian (the senior person in a UK NHS Trust responsible for safeguarding patient data), as well as representatives from child protection, clinicians, and researchers.

The NIHR Biomedical Research Centre (BRC) Nucleus was created through a transformational grant of £3 million from South London and Maudsley NHS Trust (SLaM) and Guy’s and St Thomas’ Charity. The Nucleus is a data management and informatics hub designed to support pioneering translational research into the prevention, treatment, and understanding of mental health and dementia, and to use those findings to benefit patients as quickly as possible.

The BRC Nucleus brings together specialists in informatics, biostatistics, epidemiology, psychiatry, psychology, and neuroscience to analyse complex biomedical data in a way not previously possible. The new research tools being developed in the Nucleus are designed to improve the quality of research data and produce faster, more focused results, from large and complex clinical, genomic, proteomic, and neuroimaging data. Over the years, work at the Nucleus has focused on developing new biomarkers and diagnostic tools, as well as novel therapies and treatments for mental health and related disorders. Patients and members of the public are fundamental in helping shape research priorities.

Read more about the BRC Nucleus.

In August 2009, a research paper describing the development of the Clinical Record Interactive Search (CRIS) system was published in the open access journal BMC Psychiatry.

A £1 million grant from the National Institute for Health Research (NIHR) was awarded to the Biomedical Research Centre (BRC) to support the further development of CRIS. It was spent enhancing computing resources to link the system with census data and other large datasets including neuroimaging, genomics, and pharmacy records.

Natural language processing (text mining) can extract very large amounts of in-depth information recorded in documents that would otherwise have not been available for analysis. A lot of the most valuable information in mental healthcare is recorded in text only, and it is vital to capture this detail for research to be effective.

Working with experts in natural language processing from the University of Sheffield, CRIS researchers have designed close to 100 individual programmes over the years. These identify important features from the electronic health record such as the symptoms someone is experiencing, or the treatments they are receiving. Using this technique, CRIS researchers have captured a great deal of detail about real world mental health issues in very large populations to examine which treatments work best for which people. They work closely with other research groups nationally and internationally on natural language processing applied in healthcare—for example through the Healtex national network.

During a speech to announce plans for the UK Government's Life Sciences Strategy to transform healthcare research and patient treatment, the Prime Minister made special reference to CRIS. He stressed its importance for unlocking a multitude of information that could help develop evidence-based treatments. Watch footage of David Cameron’s speech in this video.

The Clinical Data Linkage Service (CDLS) was created to connect mental health data from CRIS with a range of other data sets. This important resource is helping scientists gain insight into the important relationships between mental and physical health and care. People with mental health problems face a lot of disadvantages in their physical health, and this linkage helps pinpoint key areas where healthcare can be improved. Datasets are linked by unique identifiers such as an NHS number or date of birth. These are then removed, and the information fully anonymised, before researchers are given access to the data.

Since then, CRIS has been linked with a range of data sets including the Hospital Episode Statistics (HES) database that records all accident and emergency, general admissions, and outpatient visits to hospitals across England. This linkage has enabled researchers to compare physical illnesses and hospital use by patients within South London and Maudsley NHS Foundation Trust (SLaM) and the general population, as well as look at risk factors for physical illness within patient groups, and how pre-existing physical health problems may impact on the outcomes of mental health treatments.

Read more on our website.

Thanks to South London and Maudsley NHS Foundation Trust (SLaM), researchers now had access to a database of over 200,000 fully-electronic, detailed, and anonymised mental health records. This made CRIS the most in-depth mental health data resource in Europe.

In February 2013, peers from the House of Lords visited the NIHR Maudsley Biomedical Research Centre at King’s College London and the South London and Maudsley NHS Foundation Trust (SLaM) (as it is now known). The peers were interested in the critical role of service users in the design, implementation, and evaluation of research projects. The visit was led by Baroness Judith Jolly and included Baroness Sally Greengross and Baroness Sheila Hollins.

Anne Weston, Francis Crick Institute

The successful linkage between CRIS and the Thames Cancer Registry (TCR), covering 12 million people in London, Kent, Surrey, and Sussex, provided an important opportunity to investigate the relationship between mental health and cancer outcomes. This is driving research to identify opportunities to improve prevention, standards of care, and clinical outcomes.

In 2013, a study published in BMJ Open found that people with severe mental disorders, depression, dementia, and substance use disorders are more likely to die of cancer than those in the general population, even though the stage of cancer at diagnosis was not more advanced. The researchers could not say with certainty why more people with mental illness are dying from cancer than those without a history of mental disorders, but they identified possible contributing factors such as poorer care or unequal access to appropriate care, and differing tolerance to treatment regimens.

Today, this work is being extended to link CRIS with the National Cancer Registration and Analysis Service operated by Public Health England. This service collects data on all cases of cancer that occur in people living in England. The linkage is expected to be completed in early 2019, and will provide unprecedented depth of information on the neglected group of people coping with cancer and mental illness.

During a BBC Radio 4 Today Programme interview in February 2014, Professor Matthew Hotopf, Director of the NIHR Maudsley Biomedical Research Centre (BRC), discussed the importance of using electronic medical records for research into improving patient care. Making special reference to CRIS, he stressed that the rewards of medical record research could be enormous.

The National Institute for Health Research (NIHR) Maudsley Biomedical Research Centre (BRC) Nucleus celebrated 5 years of pioneering innovation and achievement in mental health research at a showcase in September 2014. The event was attended by leading academics and representatives of major funding organisations, including the Chief Medical Officer of NHS England Professor Dame Sally Davies.

Read more on our website.

The Improving Access to Psychological Therapies (IAPT) initiative began in 2008 and has transformed the way in which talking treatments like cognitive behavioural therapy are delivered in the NHS. South London and Maudsley NHS Foundation Trust (SLaM) provided IAPT services to its catchment using a separate electronic health records system. In 2015 this was incorporated into CRIS, alongside the other SLaM records and with the same security and governance structures. This increased the size of the CRIS database by around 100,000 people and has opened up a wealth of data on psychotherapies for research and service development.  

By analysing the CRIS health record of more than 7,500 individuals with schizophrenia from the South London and Maudsley NHS Foundation Trust (SLaM), researchers identified a link between negative symptoms experienced by people with schizophrenia and poor clinical outcomes. People with schizophrenia usually experience a combination of negative symptoms (e.g. poor motivation, emotional withdrawal, and a reduction in speech and activity) and positive symptoms (e.g. hallucinations and delusions).

Using natural language processing (text mining), the research team led by Dr Rashmi Patel from the Institute of Psychiatry, Psychology & Neuroscience (IoPPN) at King’s College London, found that negative symptoms are a relatively common feature of schizophrenia and are associated with increased hospital admission, readmission, and length of inpatient stay. Published in BMJ Open, the findings highlight the importance of developing effective treatments for the negative symptoms of schizophrenia.

This paved the way for a range of automated information extraction tools to be developed over the following 2 years (see 2017) to help CRIS scientists move beyond diagnosis to symptom profiling of other mental health conditions.

The Centre for Translational Informatics (CTI) was established in June 2016 to enable CRIS to have a more direct impact on clinical care. The aim was to find new ways of improving mental healthcare using digital technologies, and to accelerate their delivery to patients.

Building on research breakthroughs made at the NIHR Maudsley Biomedical Research Centre, the CTI provides a front door to collaboration with the digital health industry. Partners include technology companies such as IBM and FitBit; academic institutions including the University of Cambridge and the University of Michigan; and many other research organisations including UK Biobank and the Farr Institute of Health Informatics Research.

Recognising that digital products and services have often failed in the NHS due to a lack of buy-in from frontline staff and service users, the CTI puts service users, carers, and clinicians at the heart of projects as designers and developers. Research at the CTI has resulted in the development of digital solutions that are transforming the way healthcare is provided including Healthlocker, Agents in RecordsCogStack and Medichec.

Read more about the CTI on our website.

Richard Jackson and the teams at King’s College London Institute of Psychiatry, Psychology & Neuroscience (IoPPN) and the University of Sheffield developed natural language processing (text mining) models that can capture a broad range of severe mental illness symptoms from free text contained in electronic mental health records. By analysing discharge summaries of over 15,000 patients with severe mental illnesses contained in CRIS, they identified large amounts of novel information on 46 symptoms, more recently supplemented by 20 depression symptoms. This CRIS-CODE project, published in BMJ Open, offers a new powerful tool for the collection of diagnostic and outcome data as well as symptom profiles from electronic health records.

Scientists are currently undertaking research to provide evidence on the entire life course of some of the most common diseases from pregnancy to adulthood, and how they might be prevented. Led by Professor Lucilla Poston, using CRIS, and commencing this year, the eLIXIR (early-Life data cross-Linkage in Research) project links information from maternity, neonatal, and mental health records with blood samples from mothers and children from South London and Maudsley NHS Foundation Trust (SLaM). Over many years this will develop into a powerful resource to help scientists discover when and how health problems begin, how they might be prevented, and what treatments work best for which people. In the future, researchers plan to extend this work to link with primary care data, child clinical health data, and national hospital and school data. Read more about the eLIXIR project on our website.

The successful linkage of CRIS with census record data from the Office for National Statistics (ONS) will provide an important opportunity to understand social disadvantages and inequalities faced by people living with mental disorders. This followed the earlier CRIS linkage with the National Pupil Database in 2013, which has allowed important questions to be answered on interrelationships between education and mental health. A number of informative linkages with data outside healthcare that allow investigations of the much wider causes, contexts, and consequences of mental disorders are anticipated.

Read more on our website.

Ten years since its launch, CRIS is now the most in-depth mental health resource in Europe, containing over 400,000 anonymised mental health records. It has been deployed across several other mental health trusts in the UK, and over 130 research papers have been published using data from the system.

Matthew Broadbent, CRIS Clinical Informatics Lead said: “The value of CRIS is reflected in the increasing demand to use it over the past decade, rising to over 120 project applications per year. More than anything, the success of CRIS is down to the imagination, skill, and determination of the team of people who have set it up, keep it running, and support its day-to-day use.”  

Commenting on future plans, Professor Robert Stewart, Academic Lead for CRIS said: “There are real opportunities for the UK to achieve a unique position in ‘healthcare data science’ by drawing on routine information from clinical care to better understand health and disease and improve services. Using CRIS, successful broader platforms like eLIXIR can be readily replicated in other Trusts and built into strong national networks. This will help the NHS to become a true world leader in clinical informatics.”