An approach to linking education, social care and electronic health records for children and young people in South London: a linkage study of child and adolescent mental health service data

A scoping review of fair machine learning techniques when using real-world data

OBJECTIVE

The integration of artificial intelligence (AI) and machine learning (ML) in health care to aid clinical decisions is widespread. However, as AI and ML take important roles in health care, there are concerns about AI and ML associated fairness and bias. That is, an AI tool may have a disparate impact, with its benefits and drawbacks unevenly distributed across societal strata and subpopulations, potentially exacerbating existing health inequities. Thus, the objectives of this scoping review were to summarize existing literature and identify gaps in the topic of tackling algorithmic bias and optimizing fairness in AI/ML models using real-world data (RWD) in health care domains.

METHODS

We conducted a thorough review of techniques for assessing and optimizing AI/ML model fairness in health care when using RWD in health care domains. The focus lies on appraising different quantification metrics for accessing fairness, publicly accessible datasets for ML fairness research, and bias mitigation approaches.

RESULTS

We identified 11 papers that are focused on optimizing model fairness in health care applications. The current research on mitigating bias issues in RWD is limited, both in terms of disease variety and health care applications, as well as the accessibility of public datasets for ML fairness research. Existing studies often indicate positive outcomes when using pre-processing techniques to address algorithmic bias. There remain unresolved questions within the field that require further research, which includes pinpointing the root causes of bias in ML models, broadening fairness research in AI/ML with the use of RWD and exploring its implications in healthcare settings, and evaluating and addressing bias in multi-modal data.

CONCLUSION

This paper provides useful reference material and insights to researchers regarding AI/ML fairness in real-world health care data and reveals the gaps in the field. Fair AI/ML in health care is a burgeoning field that requires a heightened research focus to cover diverse applications and different types of RWD.

Shared goals for mental health research: what, why and when for the 2020s

Mental health problems bring substantial individual, community and societal costs and the need for innovation to promote good mental health and to prevent and treat mental health problems has never been greater. However, we know that research findings can take up to 20 years to implement. One way to push the pace is to focus researchers and funders on shared, specific goals and targets. We describe a consultation process organised by the Department of Health and Social Care and convened by the Chief Medical Officer to consider high level goals for future research efforts and to begin to identify UK-specific targets to measure research impact. The process took account of new scientific methods and evidence, the UK context with a universal health care system (the NHS) and the embedded research support from the National Institute for Health Research Clinical Research Network, as well as the views of individual service users and service user organisations. The result of the consultation is a set of four overarching goals with the potential to be measured at intervals of three, five or ten years.

Association between depression diagnosis and educational attainment trajectories: an historical cohort study using linked data

BACKGROUND

Depression symptoms are thought to be associated with lower educational attainment, but patterns of change in attainment among those who receive a clinical diagnosis of depression at any point during childhood and adolescence remain unclear.

METHODS

We conducted a secondary analysis of an existing data linkage between a national educational dataset (National Pupil Database) and pseudonymised electronic health records (Clinical Record Interactive Search) from a large mental healthcare provider in London, United Kingdom (2007 to 2013). A cohort of 222,027 pupils were included. We used Growth Mixture Modelling (GMM) and stakeholder input to estimate trajectories of standardised educational attainment over School Years 2, 6 and 11. Multinomial logistic regression analyses were then used to investigate the association between resulting educational attainment trajectory membership (outcome) and depression diagnosis any time before age 18 (exposure).

RESULTS

A five-trajectory GMM solution for attainment was derived: (1) average/high-stable, (2) average-modest declining, (3) average-steep declining, (4) low-improving and (5) low-stable. After adjusting for clinical and sociodemographic covariates, having a depression diagnosis before age 18 was associated with occupying the average-modest declining trajectory (RRR = 2.80, 95% CI 2.36-3.32, p < .001) or the average-steep declining trajectory (RRR = 3.54, 95% CI 3.10-4.04, p < .001), as compared to the average/high-stable trajectory.

CONCLUSIONS

Receiving a diagnosis of depression before age 18 was associated with a relative decline in attainment throughout school. While these findings cannot support a causal direction, they nonetheless suggest a need for timely mental health and educational support among pupils struggling with depression.

The association between early childhood onset epilepsy and attention-deficit hyperactivity disorder (ADHD) in 3237 children and adolescents with Autism Spectrum Disorder (ASD): a historical longitudinal cohort data linkage study

Children and young people with Autism Spectrum Disorder (ASD) have an increased risk of comorbidities, such as epilepsy and Attention-Deficit/Hyperactivity Disorder (ADHD). However, little is known about the relationship between early childhood epilepsy (below age 7) and later ADHD diagnosis (at age 7 or above) in ASD. In this historical cohort study, we examined this relationship using an innovative data source, which included linked data from routinely collected acute hospital paediatric records and childhood community and inpatient psychiatric records. In a large sample of children and young people with ASD (N = 3237), we conducted a longitudinal analysis to examine early childhood epilepsy as a risk factor for ADHD diagnosis while adjusting for potential confounders, including socio-demographic characteristics, intellectual disability, family history of epilepsy and associated physical conditions. We found that ASD children and young people diagnosed with early childhood epilepsy had nearly a twofold increase in risk of developing ADHD later in life, an association which persisted after adjusting for potential confounders (adjusted OR = 1.72, CI95% = 1.13-2.62). This study suggests that sensitive monitoring of ADHD symptoms in children with ASD who have a history of childhood epilepsy may be important to promote early detection and treatment. It also highlights how linked electronic health records can be used to examine potential risk factors over time for multimorbidity in neurodevelopmental conditions.

Patient and public involvement to inform priorities and practice for research using existing healthcare data for children's and young people's cancers

BACKGROUND

In the United Kingdom, healthcare data is collected on all patients receiving National Health Service (NHS) care, including children and young people (CYP) with cancer. This data is used to inform service delivery, and with special permissions used for research. The use of routinely collected health data in research is an advancing field with huge potential benefit, particularly in CYP with cancer where case numbers are small and the impact across the life course can be significant. Patient and public involvement (PPI) exercise aims: Identify current barriers to trust relating to the use of healthcare data for research. Determine ways to increase public and patient confidence in the use of healthcare data in research. Define areas of research importance to CYP and their carers using healthcare data.

METHODS

Young people currently aged between 16 and 25 years who had a cancer diagnosis before the age of 20 years and carers of a young person with cancer were invited to take part via social media and existing networks of service users. Data was collected during two interactive online workshops totalling 5 h and comprising of presentations from health data experts, case-studies and group discussions. With participant consent the workshops were recorded, transcribed verbatim and analysed using thematic analysis.

RESULTS

Ten young people and six carers attended workshop one. Four young people and four carers returned for workshop two. Lack of awareness of how data is used, and negative media reporting were seen as the main causes of mistrust. Better communication and education on how data is used were felt to be important to improving public confidence. Participants want the ability to have control over their own data use. Late effects, social and education outcomes and research on rare tumours were described as key research priorities for data use.

CONCLUSIONS

In order to improve public and patient trust in our use of data for research, we need to improve communication about how data is used and the benefits that arise.

Longitudinal pathways between emotional difficulties and school absenteeism in middle childhood: Evidence from developmental cascades

Emotional difficulties are associated with both authorized and unauthorized school absence, but there has been little longitudinal research and the temporal nature of these associations remains unclear. This study presents three-wave random-intercepts panel models of longitudinal reciprocal relationships between teacher-reported emotional difficulties and authorized and unauthorized school absence in 2,542 English children aged 6 to 9 years old at baseline, who were followed-up annually. Minor differences in the stability effects were observed between genders but only for the authorized absence model. Across all time points, children with greater emotional difficulties had more absences, and vice versa (authorized: ρ = .23-.29, p < .01; unauthorized: ρ = .28, p < .01). At the within-person level, concurrent associations showed that emotional difficulties were associated with greater authorized (β = .15-.17, p < .01) absence at Time 3 only, but with less unauthorized (β = -.08-.13, p < .05) absence at Times 1 and 2. In cross-lagged pathways, neither authorized nor unauthorized absence predicted later emotional difficulties, and emotional difficulties did not predict later authorized absence at any time point. However, greater emotional difficulties were associated with fewer unauthorized absences across time (β = -13-.22, p < .001). The implications of these findings are discussed.

A national multiple baseline cohort study of mental health conditions in early adolescence and subsequent educational outcomes in New Zealand

Young people experiencing mental health conditions are vulnerable to poorer educational outcomes for many reasons, including: social exclusion, stigma, and limited in-school support. Using a near-complete New Zealand population administrative database, this prospective cohort study aimed to quantify differences in educational attainment (at ages 15-16 years) and school suspensions (over ages 13-16 years), between those with and without a prior mental health condition. The data included five student cohorts, each starting secondary school from 2013 to 2017 respectively (N = 272,901). Both internalising and externalising mental health conditions were examined. Overall, 6.8% had a mental health condition. Using adjusted modified Poisson regression analyses, those with prior mental health conditions exhibited lower rates of attainment (IRR 0.87, 95% CI 0.86-0.88) and higher rates of school suspensions (IRR 1.63, 95% CI 1.57-1.70) by age 15-16 years. Associations were stronger among those exhibiting behavioural conditions, compared to emotional conditions, in line with previous literature. These findings highlight the importance of support for young people experiencing mental health conditions at this crucial juncture in their educational pathway. While mental health conditions increase the likelihood of poorer educational outcomes, deleterious outcomes were not a necessary sequalae. In this study, most participants with mental health conditions had successful educational outcomes.

De-identified Bayesian personal identity matching for privacy-preserving record linkage despite errors: development and validation

BACKGROUND

Epidemiological research may require linkage of information from multiple organizations. This can bring two problems: (1) the information governance desirability of linkage without sharing direct identifiers, and (2) a requirement to link databases without a common person-unique identifier.

METHODS

We develop a Bayesian matching technique to solve both. We provide an open-source software implementation capable of de-identified probabilistic matching despite discrepancies, via fuzzy representations and complete mismatches, plus de-identified deterministic matching if required. We validate the technique by testing linkage between multiple medical records systems in a UK National Health Service Trust, examining the effects of decision thresholds on linkage accuracy. We report demographic factors associated with correct linkage.

RESULTS

The system supports dates of birth (DOBs), forenames, surnames, three-state gender, and UK postcodes. Fuzzy representations are supported for all except gender, and there is support for additional transformations, such as accent misrepresentation, variation for multi-part surnames, and name re-ordering. Calculated log odds predicted a proband's presence in the sample database with an area under the receiver operating curve of 0.997-0.999 for non-self database comparisons. Log odds were converted to a decision via a consideration threshold θ and a leader advantage threshold δ. Defaults were chosen to penalize misidentification 20-fold versus linkage failure. By default, complete DOB mismatches were disallowed for computational efficiency. At these settings, for non-self database comparisons, the mean probability of a proband being correctly declared to be in the sample was 0.965 (range 0.931-0.994), and the misidentification rate was 0.00249 (range 0.00123-0.00429). Correct linkage was positively associated with male gender, Black or mixed ethnicity, and the presence of diagnostic codes for severe mental illnesses or other mental disorders, and negatively associated with birth year, unknown ethnicity, residential area deprivation, and presence of a pseudopostcode (e.g. indicating homelessness). Accuracy rates would be improved further if person-unique identifiers were also used, as supported by the software. Our two largest databases were linked in 44 min via an interpreted programming language.

CONCLUSIONS

Fully de-identified matching with high accuracy is feasible without a person-unique identifier and appropriate software is freely available.

Linking electronic mental healthcare and benefits records in South London: design, procedure and descriptive outcomes

OBJECTIVES

To describe the process and outcomes of a data linkage between electronic secondary mental healthcare records from the South London and Maudsley (SLaM) NHS Foundation Trust with benefits records from the Department for Work and Pensions (DWP). We also describe the mental health and benefit profile of patients who were successfully linked.

DESIGN

A deterministic linkage of routine records from health and welfare government service providers within a secure environment.

SETTING AND PARTICIPANTS

Adults aged≥18 years who were referred to or accessed treatment at SLaM services between January 2007 and June 2019, including those who were treated as part of Improving Access to Psychological Therapies (IAPT) services between January 2008 and June 2019 (n=448 404). Benefits data from the DWP from January 2005 to June 2020.

OUTCOME MEASURES

The linkage rate and associated sociodemographic, diagnostic and treatment factors. Recorded primary psychiatric diagnosis based on International Classification of Diseases (ICD)-10 codes and type of benefit receipt.

RESULTS

A linkage rate of 92.3% was achieved. Women, younger patients and those from ethnic minority groups were less likely to be successfully linked. Patients who had subsequently died, had a recorded primary psychiatric diagnosis, had also engaged with IAPT and had a higher number of historical postcodes available were more likely to be linked. Overall, 83% of patients received benefits at some point between 2005 and 2020. Benefit receipt across the psychiatric diagnosis spectrum was high, over 80% across most ICD-10 codes.

CONCLUSIONS

This data linkage is the first of its kind in the UK demonstrating the use of routinely collected mental health and benefits data. Benefit receipt was high among patients accessing SLaM services and varied by psychiatric diagnosis. Future areas of research are discussed, including exploring the effectiveness of interventions for helping people into work and the impact of benefit reforms.

Identifying Military Service Status in Electronic Healthcare Records from Psychiatric Secondary Healthcare Services: A Validation Exercise Using the Military Service Identification Tool

Electronic healthcare records (EHRs) are a rich source of information with a range of uses in secondary research. In the United Kingdom, there is no pan-national or nationally accepted marker indicating veteran status across all healthcare services. This presents significant obstacles to determining the healthcare needs of veterans using EHRs. To address this issue, we developed the Military Service Identification Tool (MSIT), using an iterative two-staged approach. In the first stage, a Structured Query Language approach was developed to identify veterans using a keyword rule-based approach. This informed the second stage, which was the development of the MSIT using machine learning, which, when tested, obtained an accuracy of 0.97, a positive predictive value of 0.90, a sensitivity of 0.91, and a negative predictive value of 0.98. To further validate the performance of the MSIT, the present study sought to verify the accuracy of the EHRs that trained the MSIT models. To achieve this, we surveyed 902 patients of a local specialist mental healthcare service, with 146 (16.2%) being asked if they had or had not served in the Armed Forces. In total 112 (76.7%) reported that they had not served, and 34 (23.3%) reported that they had served in the Armed Forces (accuracy: 0.84, sensitivity: 0.82, specificity: 0.91). The MSIT has the potential to be used for identifying veterans in the UK from free-text clinical documents and future use should be explored.

Assessing the feasibility of a web-based outcome measurement system in child and adolescent mental health services - myHealthE a randomised controlled feasibility pilot study

BACKGROUND

Interest in internet-based patient reported outcome measure (PROM) collection is increasing. The NHS myHealthE (MHE) web-based monitoring system was developed to address the limitations of paper-based PROM completion. MHE provides a simple and secure way for families accessing Child and Adolescent Mental Health Services to report clinical information and track their child's progress. This study aimed to assess whether MHE improves the completion of the Strengths and Difficulties Questionnaire (SDQ) compared with paper collection. Secondary objectives were to explore caregiver satisfaction and application acceptability.

METHODS

A 12-week single-blinded randomised controlled feasibility pilot trial of MHE was conducted with 196 families accessing neurodevelopmental services in south London to examine whether electronic questionnaires are completed more readily than paper-based questionnaires over a 3-month period. Follow up process evaluation phone calls with a subset (n = 8) of caregivers explored system satisfaction and usability.

RESULTS

MHE group assignment was significantly associated with an increased probability of completing an SDQ-P in the study period (adjusted hazard ratio (HR) 12.1, 95% CI 4.7-31.0; p = <.001). Of those caregivers' who received the MHE invitation (n = 68) 69.1% completed an SDQ using the platform compared to 8.8% in the control group (n = 68). The system was well received by caregivers, who cited numerous benefits of using MHE, for example, real-time feedback and ease of completion.

CONCLUSIONS

MHE holds promise for improving PROM completion rates. Research is needed to refine MHE, evaluate large-scale MHE implementation, cost effectiveness and explore factors associated with differences in electronic questionnaire uptake.

Moving from development to implementation of digital innovations within the NHS: myHealthE, a remote monitoring system for tracking patient outcomes in child and adolescent mental health services

Objective

This paper aims to report our experience of developing, implementing, and evaluating myHealthE (MHE), a digital innovation for Child and Adolescents Mental Health Services (CAMHS), which automates the remote collection and reporting of Patient-Reported Outcome Measures (PROMs) into National Health Services (NHS) electronic healthcare records.

Methods

We describe the logistical and governance issues encountered in developing the MHE interface with patient-identifiable information, and the steps taken to overcome these development barriers. We describe the application's architecture and hosting environment to enable its operability within the NHS, as well as the capabilities needed within the technical team to bridge the gap between academic development and NHS operational teams.

Results

We present evidence on the feasibility and acceptability of this system within clinical services and the process of iterative development, highlighting additional functions that were incorporated to increase system utility.

Conclusion

This article provides a framework with which to plan, develop, and implement automated PROM collection from remote devices back to NHS infrastructure. The challenges and solutions described in this paper will be pertinent to other digital health innovation researchers aspiring to deploy interoperable systems within NHS clinical systems.

Assessing machine learning for fair prediction of ADHD in school pupils using a retrospective cohort study of linked education and healthcare data

OBJECTIVES

Attention deficit hyperactivity disorder (ADHD) is a prevalent childhood disorder, but often goes unrecognised and untreated. To improve access to services, accurate predictions of populations at high risk of ADHD are needed for effective resource allocation. Using a unique linked health and education data resource, we examined how machine learning (ML) approaches can predict risk of ADHD.

DESIGN

Retrospective population cohort study.

SETTING

South London (2007-2013).

PARTICIPANTS

n=56 258 pupils with linked education and health data.

PRIMARY OUTCOME MEASURES

Using area under the curve (AUC), we compared the predictive accuracy of four ML models and one neural network for ADHD diagnosis. Ethnic group and language biases were weighted using a fair pre-processing algorithm.

RESULTS

Random forest and logistic regression prediction models provided the highest predictive accuracy for ADHD in population samples (AUC 0.86 and 0.86, respectively) and clinical samples (AUC 0.72 and 0.70). Precision-recall curve analyses were less favourable. Sociodemographic biases were effectively reduced by a fair pre-processing algorithm without loss of accuracy.

CONCLUSIONS

ML approaches using linked routinely collected education and health data offer accurate, low-cost and scalable prediction models of ADHD. These approaches could help identify areas of need and inform resource allocation. Introducing 'fairness weighting' attenuates some sociodemographic biases which would otherwise underestimate ADHD risk within minority groups.

Learning outcomes in primary school children with emotional problems: a prospective cohort study

BACKGROUND

Academic difficulties are common in adolescents with mental health problems. Although earlier childhood emotional problems, characterised by heightened anxiety and depressive symptoms are common forerunners to adolescent mental health problems, the degree to which mental health problems in childhood may contribute independently to academic difficulties has been little explored.

METHODS

Data were drawn from a prospective cohort study of students in Melbourne, Australia (N = 1239). Data were linked with a standardised national assessment of academic performance at baseline (9 years) and wave three (11 years). Depressive and anxiety symptoms were assessed at baseline and wave two (10 years). Regression analyses estimated the association between emotional problems (9 and/or 10 years) and academic performance at 11 years, adjusting for baseline academic performance, sex, age and socioeconomic status, and hyperactivity/inattention symptoms.

RESULTS

Students with depressive symptoms at 9 years of age had lost nearly 4 months of numeracy learning two years later after controlling for baseline academic performance and confounders. Results were similar for anxiety symptoms. Regardless of when depressive symptoms occurred there were consistent associations with poorer numeracy performance at 11 years. The association of depressive symptoms with reading performance was weaker than for numeracy if they were present at wave two. Persistent anxiety symptoms across two waves led to nearly a 4 month loss of numeracy learning at 11 years, but the difference was not meaningful for reading. Findings were similar when including hyperactivity/inattention symptoms.

CONCLUSIONS

Childhood anxiety and depression are not only forerunners of later mental health problems but predict academic achievement. Partnerships between education and health systems have the potential to not only improve childhood emotional problems but also improve learning.

Debate: 'A rose by any other name' would smell as sweet - myths peddled about the ills of diagnosing conduct disorders

Using a diagnosis of ODD/CD enables the clinician to apply a huge amount of helpful information about what has caused the concerns and then to offer effective treatment. More often than not, they do not use a diagnostic label with the family; the point is for the clinician to share the expertise. Myths about the nature of psychiatric diagnosis and the harms of a label of ODD/CD are debunked in this article. It is society who stigmatises these individuals because of their antisocial behaviour, and withholding skilfully applied benefits of the accumulated knowledge of the condition is cruel and especially harmful to disadvantaged groups in society where ODD/CD is far more prevalent.

Rapid systematic review to identify key barriers to access, linkage, and use of local authority administrative data for population health research, practice, and policy in the United Kingdom

BACKGROUND

Improving data access, sharing, and linkage across local authorities and other agencies can contribute to improvements in population health. Whilst progress is being made to achieve linkage and integration of health and social care data, issues still exist in creating such a system. As part of wider work to create the Cambridge Child Health Informatics and Linked Data (Cam-CHILD) database, we wanted to examine barriers to the access, linkage, and use of local authority data.

METHODS

A systematic literature search was conducted of scientific databases and the grey literature. Any publications reporting original research related to barriers or enablers of data linkage of or with local authority data in the United Kingdom were included. Barriers relating to the following issues were extracted from each paper: funding, fragmentation, legal and ethical frameworks, cultural issues, geographical boundaries, technical capability, capacity, data quality, security, and patient and public trust.

RESULTS

Twenty eight articles were identified for inclusion in this review. Issues relating to technical capacity and data quality were cited most often. This was followed by those relating to legal and ethical frameworks. Issue relating to public and patient trust were cited the least, however, there is considerable overlap between this topic and issues relating to legal and ethical frameworks.

CONCLUSIONS

This rapid review is the first step to an in-depth exploration of the barriers to data access, linkage and use; a better understanding of which can aid in creating and implementing effective solutions. These barriers are not novel although they pose specific challenges in the context of local authority data.

Autism spectrum disorders as a risk factor for adolescent self-harm: a retrospective cohort study of 113,286 young people in the UK

BACKGROUND

Individuals with autism spectrum disorder (ASD) are at particularly high risk of suicide and suicide attempts. Presentation to a hospital with self-harm is one of the strongest risk factors for later suicide. We describe the use of a novel data linkage between routinely collected education data and child and adolescent mental health data to examine whether adolescents with ASD are at higher risk than the general population of presenting to emergency care with self-harm.

METHODS

A retrospective cohort study was conducted on the population aged 11-17 resident in four South London boroughs between January 2009 and March 2013, attending state secondary schools, identified in the National Pupil Database (NPD). Exposure data on ASD status were derived from the NPD. We used Cox regression to model time to first self-harm presentation to the Emergency Department (ED).

RESULTS

One thousand twenty adolescents presented to the ED with self-harm, and 763 matched to the NPD. The sample for analysis included 113,286 adolescents (2.2% with ASD). For boys only, there was an increased risk of self-harm associated with ASD (adjusted hazard ratio 2·79, 95% CI 1·40-5·57, P<0·01). Several other factors including school absence, exclusion from school and having been in foster care were also associated with a higher risk of self-harm.

CONCLUSIONS

This study provides evidence that ASD in boys, and other educational, social and clinical factors, are risk factors for emergency presentation with self-harm in adolescents. These findings are an important step in developing early recognition and prevention programmes.

Public opinion on sharing data from health services for clinical and research purposes without explicit consent: an anonymous online survey in the UK

OBJECTIVES

UK National Health Service/Health and Social Care (NHS/HSC) data are variably shared between healthcare organisations for direct care, and increasingly de-identified for research. Few large-scale studies have examined public opinion on sharing, including of mental health (MH) versus physical health (PH) data. We measured data sharing preferences.

DESIGN/SETTING/INTERVENTIONS/OUTCOMES

Pre-registered anonymous online survey, measuring expressed preferences, recruiting February to September 2020. Participants were randomised to one of three framing statements regarding MH versus PH data.

PARTICIPANTS

Open to all UK residents. Participants numbered 29 275; 40% had experienced an MH condition.

RESULTS

Most (76%) supported identifiable data sharing for direct clinical care without explicit consent, but 20% opposed this. Preference for clinical/identifiable sharing decreased with geographical distance and was slightly less for MH than PH data, with small framing effects. Preference for research/de-identified data sharing without explicit consent showed the same small PH/MH and framing effects, plus greater preference for sharing structured data than de-identified free text. There was net support for research sharing to the NHS, academic institutions, and national research charities, net ambivalence about sharing to profit-making companies researching treatments, and net opposition to sharing to other companies (similar to sharing publicly). De-identified linkage to non-health data was generally supported, except to data held by private companies. We report demographic influences on preference. A majority (89%) supported a single NHS mechanism to choose uses of their data. Support for data sharing increased during COVID-19.

CONCLUSIONS

Support for healthcare data sharing for direct care without explicit consent is broad but not universal. There is net support for the sharing of de-identified data for research to the NHS, academia, and the charitable sector, but not the commercial sector. A single national NHS-hosted system for patients to control the use of their NHS data for clinical purposes and for research would have broad support.

TRIAL REGISTRATION NUMBER

ISRCTN37444142.

Public opinion on sharing data from health services for clinical and research purposes without explicit consent: an anonymous online survey in the UK

OBJECTIVES

UK National Health Service/Health and Social Care (NHS/HSC) data are variably shared between healthcare organisations for direct care, and increasingly de-identified for research. Few large-scale studies have examined public opinion on sharing, including of mental health (MH) versus physical health (PH) data. We measured data sharing preferences.

DESIGN/SETTING/INTERVENTIONS/OUTCOMES

Pre-registered anonymous online survey, measuring expressed preferences, recruiting February to September 2020. Participants were randomised to one of three framing statements regarding MH versus PH data.

PARTICIPANTS

Open to all UK residents. Participants numbered 29 275; 40% had experienced an MH condition.

RESULTS

Most (76%) supported identifiable data sharing for direct clinical care without explicit consent, but 20% opposed this. Preference for clinical/identifiable sharing decreased with geographical distance and was slightly less for MH than PH data, with small framing effects. Preference for research/de-identified data sharing without explicit consent showed the same small PH/MH and framing effects, plus greater preference for sharing structured data than de-identified free text. There was net support for research sharing to the NHS, academic institutions, and national research charities, net ambivalence about sharing to profit-making companies researching treatments, and net opposition to sharing to other companies (similar to sharing publicly). De-identified linkage to non-health data was generally supported, except to data held by private companies. We report demographic influences on preference. A majority (89%) supported a single NHS mechanism to choose uses of their data. Support for data sharing increased during COVID-19.

CONCLUSIONS

Support for healthcare data sharing for direct care without explicit consent is broad but not universal. There is net support for the sharing of de-identified data for research to the NHS, academia, and the charitable sector, but not the commercial sector. A single national NHS-hosted system for patients to control the use of their NHS data for clinical purposes and for research would have broad support.

TRIAL REGISTRATION NUMBER

ISRCTN37444142.

Identification of child mental health problems by combining electronic health record information from different primary healthcare professionals: a population-based cohort study

OBJECTIVES

To investigate the potential value of combining information from electronic health records from Dutch general practitioners (GPs) and preventive youth healthcare professionals (PYHPs) in predicting child mental health problems (MHPs).

DESIGN

Population-based retrospective cohort study.

SETTING

General practice, children who were registered with 76 general practice centres from the Leiden University Medical Centre (LUMC) primary care academic network Extramural LUMC Academic Network in the Leiden area, the Netherlands. For the included children we obtained data regarding a child's healthy development from preventive youth healthcare.

PARTICIPANTS

48 256 children aged 0-19 years old who were registered with participating GPs between 2007 and 2017 and who also had data available from PYHPs from the period 2010-2015. Children with MHPs before 2007 were excluded (n=3415).

PRIMARY OUTCOME

First MHPs based on GP data.

RESULTS

In 51% of the children who had MHPs according to GPs, PYPHs also had concerns for MHPs. In 31% of the children who had no MHPs according to GPs, PYHPs had recorded concerns for MHPs. Combining their information did not result in better performing prediction models than the models based on GP data alone (c-statistics ranging from 0.62 to 0.64). Important determinants of identification of MHPs by PYHPs 1 year later were concerns from PHYPs about MHPs, borderline or increased problem scores on mental health screening tools, life events, family history of MHPs and an extra visit to preventive youth healthcare.

CONCLUSIONS

Although the use of combined information from PYHPs and GPs did not improve prediction of MHPs compared with the use of GP data alone, this study showed the feasibility of analysing a combined dataset from different healthcare providers what has the potential to inform future studies aimed at improving child MHP identification.

Estimating the impact of child and early adolescent depression on subsequent educational attainment: secondary analysis of an existing data linkage

AIMS

Depression is thought to be associated with lower subsequent educational attainment during school. But, without longitudinal studies which take account of prior attainment and other potential confounders, estimates of the impact of clinically recognised depression in childhood and early adolescence are unknown. We investigated whether a clinical diagnosis of depression is associated with lower subsequent educational attainment, and whether the association is modified by gender, ethnicity and socioeconomic status.

METHODS

We conducted a secondary analysis of an existing administrative data linkage between national educational data and a large mental healthcare provider in London, UK (2007-2013). Depression diagnosis before age 15 (exposure) was measured from electronic health records, and subsequent educational attainment at age 15-16 (outcome) was measured from educational records. We fitted logistic regression models and adjusted for gender, ethnicity, socioeconomic status, relative age in school year, neurodevelopmental disorder diagnosis and prior attainment. We investigated effect modifiers using interaction terms.

RESULTS

In total, n = 63 623 were included in analysis, of whom n = 242 had record of a depression diagnosis before age 15. Depression was associated with lower odds of subsequently achieving expected attainment levels in national exams, after adjustment for all covariates (odds ratio = 0.60, 95% confidence interval = 0.43 to 0.84, p = 0.003). There was no evidence that gender, ethnicity or socioeconomic status modified this association.

CONCLUSIONS

These findings support a relationship between depression and lower subsequent educational attainment. This highlights the need for tailored educational interventions to support children and adolescents with depression, particularly in the lead up to key educational milestones.

Factors affecting access to administrative health data for research in Canada: a study protocol

INTRODUCTION

In Canada, most provinces have established administrative health data repositories to facilitate access to these data for research. Anecdotally, researchers have described delays and substantial inter-provincial variations in the timeliness of data access approvals and receipt of data. Currently, the reasons for these delays and variations in timeliness are not well understood. This paper provides a study protocol for (1) identifying the factors affecting access to administrative health data for research within select Canadian provinces, and (2) comparing factors across provinces to assess whether and how they contribute to inter-provincial variations in access to administrative health data for research.

METHODS

A qualitative, multiple-case study research design will be used. Three cases will be included, representing three different provinces. For each case, data will be collected from documents and interviews. Specifically, interviews will be carried out with (1) research stakeholders, and (2) regulatory stakeholders (10 individuals/group * 2 groups/province * 3 provinces = 60). During within-case analysis, interview data for each stakeholder group will be analyzed separately using constant comparative analysis. Document analysis will occur iteratively, and will inform interview guide adaptation, and supplement interview data. Cross-case analysis will involve systematic comparison of findings across cases.

DISCUSSION

This study represents the first in-depth examination of access to administrative health data in Canada. The main outcome will be an overarching mid-range theory explaining inter-provincial variations in access to administrative health data in Canada. This theory will be strengthened by the inclusion of the perspectives of both researchers and those involved in the regulation of data access. The findings from this study may be used to improve equitable and timely access to administrative health data across provinces, and may be transferable to other jurisdictions where barriers to access to administrative health data have been reported.

Linking education and hospital data in England: linkage process and quality

INTRODUCTION

Linkage of administrative data for universal state education and National Health Service (NHS) hospital care would enable research into the inter-relationships between education and health for all children in England.

OBJECTIVES

We aim to describe the linkage process and evaluate the quality of linkage of four one-year birth cohorts within the National Pupil Database (NPD) and Hospital Episode Statistics (HES).

METHODS

We used multi-step deterministic linkage algorithms to link longitudinal records from state schools to the chronology of records in the NHS Personal Demographics Service (PDS; linkage stage 1), and HES (linkage stage 2). We calculated linkage rates and compared pupil characteristics in linked and unlinked samples for each stage of linkage and each cohort (1990/91, 1996/97, 1999/00, and 2004/05).

RESULTS

Of the 2,287,671 pupil records, 2,174,601 (95%) linked to HES. Linkage rates improved over time (92% in 1990/91 to 99% in 2004/05). Ethnic minority pupils and those living in more deprived areas were less likely to be matched to hospital records, but differences in pupil characteristics between linked and unlinked samples were moderate to small.

CONCLUSION

We linked nearly all pupils to at least one hospital record. The high coverage of the linkage represents a unique opportunity for wide-scale analyses across the domains of health and education. However, missed links disproportionately affected ethnic minorities or those living in the poorest neighbourhoods: selection bias could be mitigated by increasing the quality and completeness of identifiers recorded in administrative data or the application of statistical methods that account for missed links.

HIGHLIGHTS

Longitudinal administrative records for all children attending state school and acute hospital services in England have been used for research for more than two decades, but lack of a shared unique identifier has limited scope for linkage between these databases.We applied multi-step deterministic linkage algorithms to 4 one-year cohorts of children born 1 September-31 August in 1990/91, 1996/97, 1999/00 and 2004/05. In stage 1, full names, date of birth, and postcode histories from education data in the National Pupil Database were linked to the NHS Personal Demographic Service. In stage 2, NHS number, postcode, date of birth and sex were linked to hospital records in Hospital Episode Statistics.Between 92% and 99% of school pupils linked to at least one hospital record. Ethnic minority pupils and pupils who were living in the most deprived areas were least likely to link. Ethnic minority pupils were less likely than white children to link at the first step in both algorithms.Bias due to linkage errors could lead to an underestimate of the health needs in disadvantaged groups. Improved data quality, more sensitive linkage algorithms, and/or statistical methods that account for missed links in analyses, should be considered to reduce linkage bias.

The role of mental health symptomology and quality of life in predicting referrals to special child and adolescent mental health services

BACKGROUND

Children and adolescents' mental health problems have been largely assessed with conventional symptom scales, for example, Strengths and Difficulties Questionnaire (SDQ) given that it is one of the mostly widely used measures in specialist Child and Adolescent Mental Health Services (CAMHS). However, this emphasis on symptom scales might have missed some important features of the mental health challenges that children and young people experience including day to day functioning and life satisfaction aspect (i.e. qualify of life).

METHOD

The study examined longitudinal association between a young person's self-perceptions of quality of life and mental health difficulties and referral to specialist CAMHS service using a population cohort study (Targeted Mental Health in Schools service data) nested within a large-scale linkage between school (National Pupil Data base) and child mental health service administrative data (South London and Maudsley NHS Foundation Trust children and adolescent mental health services health records). Cox proportional hazard regression to estimate crude and adjusted hazard ratios (HRs) for the association between participant psychopathology, and incidence of CAMHS referral.

RESULTS

Pupils experiencing more behavioural difficulties, had an increased incidence of CAMHS referral (adjusted hazard ratio 1.1, 95% confidence interval 1.0-1.2). However, pupils who reported higher health related quality of life had a lower incidence of CAMHS referral over the follow-up period (adjusted hazard hario 0.94, 95% confidence interval 0.9-0.98).

CONCLUSION

Children and young people's perception of their quality of life should be considered at the stages of a clinical needs assessment.

Educational attainment trajectories among children and adolescents with depression, and the role of sociodemographic characteristics: longitudinal data-linkage study

BACKGROUND

Depression is associated with lower educational attainment, but there has been little investigation of long-term educational trajectories in large cohorts with diagnosed depression.

AIMS

To describe the educational attainment trajectories of children with a depression diagnosis in secondary care, and to investigate whether these trajectories vary by sociodemographic characteristics.

METHOD

We identified new referrals to South London and Maudsley's NHS Foundation Trust between 2007 and 2013 who received a depression diagnosis at under 18 years old. Linking their health records to the National Pupil Database, we standardised their performance on three assessments (typically undertaken at ages 6-7 years (school Year 2), 10-11 (Year 6) and 15-16 (Year 11)) relative to the local reference population in each academic year. We used mixed models for repeated measures to estimate attainment trajectories.

RESULTS

In our sample of 1492 children, the median age at depression diagnosis was 15 years (interquartile range = 14-16). Their attainment showed a decline between school Years 6 and 11. Attainment was consistently lower among males and those eligible for free school meals. Black ethnic groups also showed lower attainment than White ethnic groups between Years 2 and 6, but showed a less pronounced drop in attainment at Year 11.

CONCLUSIONS

Those who receive a depression diagnosis during their school career show a drop in attainment in Year 11. Although this pattern was seen among multiple sociodemographic groups, gender, ethnicity and socioeconomic status predict more vulnerable subgroups within this clinical population who might benefit from additional educational support or more intensive treatment.

Inequalities in referral pathways for young people accessing secondary mental health services in south east London

Differences in health service use between ethnic groups have been well documented, but little research has been conducted on inequalities in access to mental health services among young people. This study examines inequalities in pathways into care by ethnicity and migration status in 12-29 years old accessing health services in south east London. This study analyses anonymized electronic patient record data for patients aged 12-29 referred to a south east London mental health trust between 2008 and 2016 for an anxiety or non-psychotic depressive disorder (n = 18,931). Multinomial regression was used to examine associations between ethnicity, migration status, and both referral source and destination, stratified by age group. Young people in the Black African ethnic group were more likely to be referred from secondary health or social/criminal justice services compared to those in the White British ethnic group; the effect was most pronounced for those aged 16-17 years. Young people in the Black African ethnic group were also significantly more likely to be referred to inpatient and emergency services compared to those in the White British ethnic group. Black individuals living in south east London, particularly those who identify as Black African, are referred to mental health services via more adverse pathways than White individuals. Our findings suggest that inequalities in referral destination may be perpetuated by inequalities generated at the point of access.

Linkage of primary care prescribing records and pharmacy dispensing Records in the Salford Lung Study: application in asthma

BACKGROUND

Records of medication prescriptions can be used in conjunction with pharmacy dispensing records to investigate the incidence of adherence, which is defined as observing the treatment plans agreed between a patient and their clinician. Using prescribing records alone fails to identify primary non-adherence; medications not being collected from the dispensary. Using dispensing records alone means that cases of conditions that resolve and/or treatments that are discontinued will be unaccounted for. While using a linked prescribing and dispensing dataset to measure medication non-adherence is optimal, this linkage is not routinely conducted. Furthermore, without a unique common event identifier, linkage between these two datasets is not straightforward.

METHODS

We undertook a secondary analysis of the Salford Lung Study dataset. A novel probabilistic record linkage methodology was developed matching asthma medication pharmacy dispensing records and primary care prescribing records, using semantic (meaning) and syntactic (structure) harmonization, domain knowledge integration, and natural language feature extraction. Cox survival analysis was conducted to assess factors associated with the time to medication dispensing after the prescription was written. Finally, we used a simplified record linkage algorithm in which only identical records were matched, for a naïve benchmarking to compare against the results of our proposed methodology.

RESULTS

We matched 83% of pharmacy dispensing records to primary care prescribing records. Missing data were prevalent in the dispensing records which were not matched - approximately 60% for both medication strength and quantity. A naïve benchmarking approach, requiring perfect matching, identified one-quarter as many matching prescribing records as our methodology. Factors associated with delay (or failure) to collect the prescribed medication from a pharmacy included season, quantity of medication prescribed, previous dispensing history and class of medication. Our findings indicate that over 30% of prescriptions issued were not collected from a dispensary (primary non-adherence).

CONCLUSIONS

We have developed a probabilistic record linkage methodology matching a large percentage of pharmacy dispensing records with primary care prescribing records for asthma medications. This will allow researchers to link datasets in order to extract information about asthma medication non-adherence.

Reviewing a Decade of Research Into Suicide and Related Behaviour Using the South London and Maudsley NHS Foundation Trust Clinical Record Interactive Search (CRIS) System

Suicide is a serious public health issue worldwide, yet current clinical methods for assessing a person's risk of taking their own life remain unreliable and new methods for assessing suicide risk are being explored. The widespread adoption of electronic health records (EHRs) has opened up new possibilities for epidemiological studies of suicide and related behaviour amongst those receiving healthcare. These types of records capture valuable information entered by healthcare practitioners at the point of care. However, much recent work has relied heavily on the structured data of EHRs, whilst much of the important information about a patient's care pathway is recorded in the unstructured text of clinical notes. Accessing and structuring text data for use in clinical research, and particularly for suicide and self-harm research, is a significant challenge that is increasingly being addressed using methods from the fields of natural language processing (NLP) and machine learning (ML). In this review, we provide an overview of the range of suicide-related studies that have been carried out using the Clinical Records Interactive Search (CRIS): a database for epidemiological and clinical research that contains de-identified EHRs from the South London and Maudsley NHS Foundation Trust. We highlight the variety of clinical research questions, cohorts and techniques that have been explored for suicide and related behaviour research using CRIS, including the development of NLP and ML approaches. We demonstrate how EHR data provides comprehensive material to study prevalence of suicide and self-harm in clinical populations. Structured data alone is insufficient and NLP methods are needed to more accurately identify relevant information from EHR data. We also show how the text in clinical notes provide signals for ML approaches to suicide risk assessment. We envision increased progress in the decades to come, particularly in externally validating findings across multiple sites and countries, both in terms of clinical evidence and in terms of NLP and machine learning method transferability.

National administrative record linkage between specialist community drug and alcohol treatment data (the National Drug Treatment Monitoring System (NDTMS)) and inpatient hospitalisation data (Hospital Episode Statistics (HES)) in England: design, method and evaluation

OBJECTIVES

The creation and evaluation of a national record linkage between substance misuse treatment, and inpatient hospitalisation data in England.

DESIGN

A deterministic record linkage using personal identifiers to link the National Drug Treatment Monitoring System (NDTMS) curated by Public Health England (PHE), and Hospital Episode Statistics (HES) Admitted Patient Care curated by National Health Service (NHS) Digital.

SETTING AND PARTICIPANTS

Adults accessing substance misuse treatment in England between 1 April 2018 and 31 March 2019 (n=268 251) were linked to inpatient hospitalisation records available since 1 April 1997.

OUTCOME MEASURES

Using a gold-standard subset, linked using NHS number, we report the overall linkage sensitivity and precision. Predictors for linkage error were identified, and inverse probability weighting was used to interrogate any potential impact on the analysis of length of hospital stay.

RESULTS

79.7% (n=213 814) people were linked to at least one HES record, with an estimated overall sensitivity of between 82.5% and 83.3%, and a precision of between 90.3% and 96.4%. Individuals were more likely to link if they were women, white and aged between 46 and 60. Linked individuals were more likely to have an average length of hospital stay ≥5 days if they were men, older, had no fixed residential address or had problematic opioid use. These associations did not change substantially after probability weighting, suggesting they were not affected by bias from linkage error.

CONCLUSIONS

Linkage between substance misuse treatment and hospitalisation records offers a powerful new tool to evaluate the impact of treatment on substance related harm in England. While linkage error can produce misleading results, linkage bias appears to have little effect on the association between substance misuse treatment and length of hospital admission. As subsequent analyses are conducted, potential biases associated with the linkage process should be considered in the interpretation of any findings.

Assessing data linkage quality in cohort studies

Background: Linkage of administrative data sources provides an efficient means of collecting detailed data on how individuals interact with cross-sectoral services, society, and the environment. These data can be used to supplement conventional cohort studies, or to create population-level electronic cohorts generated solely from administrative data. However, errors occurring during linkage (false matches/missed matches) can lead to bias in results from linked data.

Aim: This paper provides guidance on evaluating linkage quality in cohort studies.

Methods: We provide an overview of methods for linkage, describe mechanisms by which linkage error can introduce bias, and draw on real-world examples to demonstrate methods for evaluating linkage quality.

Results: Methods for evaluating linkage quality described in this paper provide guidance on (i) estimating linkage error rates, (ii) understanding the mechanisms by which linkage error might bias results, and (iii) information that should be shared between data providers, linkers and users, so that approaches to handling linkage error in analysis can be implemented.

Conclusion: Linked administrative data can enhance conventional cohorts and offers the ability to answer questions that require large sample sizes or hard-to-reach populations. Care needs to be taken to evaluate linkage quality in order to provide robust results.

Cohort profile: the eLIXIR Partnership-a maternity-child data linkage for life course research in South London, UK

PURPOSE

Linked maternity, neonatal and maternal mental health records were created to support research into the early life origins of physical and mental health, in mothers and children. The Early Life Cross Linkage in Research (eLIXIR) Partnership was developed in 2018, generating a repository of real-time, pseudonymised, structured data derived from the electronic health record systems of two acute and one Mental Health Care National Health Service (NHS) Provider in South London. We present early descriptive data for the linkage database and the robust data security and governance structures, and describe the intended expansion of the database from its original development. Additionally, we report details of the accompanying eLIXIR Research Tissue Bank of maternal and neonatal blood samples.

PARTICIPANTS

Descriptive data were generated from the eLIXIR database from 1 October 2018 to 30 June 2019. Over 17 000 electronic patient records were included.

FINDINGS TO DATE

10 207 women accessed antenatal care from the 2 NHS maternity services, with 8405 deliveries (8772 infants). This diverse, inner-city maternity service population was born in over 170 countries with an ethnic profile of 46.1% white, 19.1% black, 7.0% Asian, 4.1% mixed and 4.1% other. Of the 10 207 women, 11.6% had a clinical record in mental health services with 3.0% being treated during their pregnancy. This first data extract included 947 infants treated in the neonatal intensive care unit, of whom 19.1% were postnatal transfers from external healthcare providers.

FUTURE PLANS

Electronic health records provide potentially transformative information for life course research, integrating physical and mental health disorders and outcomes in routine clinical care. The eLIXIR database will grow by ~14 000 new maternity cases annually, in addition to providing child follow-up data. Additional datasets will supplement the current linkage from other local and national resources, including primary care and hospital inpatient data for mothers and their children.

Guidance for researchers wanting to link NHS data using non-consent approaches: a thematic analysis of feedback from the Health Research Authority Confidentiality Advisory Group

INTRODUCTION

The use of linked data and non-consent methodologies is a rapidly growing area of health research due to the increasing detail, availability and scope of routinely collected electronic health records data. However, gaining the necessary legal and governance approvals to undertake data linkage is a complex process in England.

OBJECTIVES

We reflect on our own experience of establishing lawful basis for data linkage through Section 251 approval, with the intention to build a knowledgebase of practical advice for future applicants.

METHODS

Thematic analysis was conducted on a corpus of Section 251 feedback reports from the NHS Health Research Authority Confidentiality Advisory Group.

RESULTS

Four themes emerged from the feedback. These were: (a) Patient and Public Involvement, (b) Establishing Rationale, (c) Data maintenance and contingency, and the need to gain (d) Further Permissions from external authorities prior to full approval.

CONCLUSIONS

Securing Section 251 approval poses ethical, practical and governance challenges. However, through a comprehensive, planned approach Section 251 approval is possible, enabling researchers to unlock the potential of linked data for the purposes of health research.

Impact of matching error on linked mortality outcome in a data linkage of secondary mental health data with Hospital Episode Statistics (HES) and mortality records in South East London: a cross-sectional study

OBJECTIVES

Linkage of electronic health records (EHRs) to Hospital Episode Statistics (HES)-Office for National Statistics (ONS) mortality data has provided compelling evidence for lower life expectancy in people with severe mental illness. However, linkage error may underestimate these estimates. Using a clinical sample (n=265 300) of individuals accessing mental health services, we examined potential biases introduced through missed matching and examined the impact on the association between clinical disorders and mortality.

SETTING

The South London and Maudsley NHS Foundation Trust (SLaM) is a secondary mental healthcare provider in London. A deidentified version of SLaM's EHR was available via the Clinical Record Interactive Search system linked to HES-ONS mortality records.

PARTICIPANTS

Records from SLaM for patients active between January 2006 and December 2016.

OUTCOME MEASURES

Two sources of death data were available for SLaM participants: accurate and contemporaneous date of death via local batch tracing (gold standard) and date of death via linked HES-ONS mortality data. The effect of linkage error on mortality estimates was evaluated by comparing sociodemographic and clinical risk factor analyses using gold standard death data against HES-ONS mortality records.

RESULTS

Of the total sample, 93.74% were successfully matched to HES-ONS records. We found a number of statistically significant administrative, sociodemographic and clinical differences between matched and unmatched records. Of note, schizophrenia diagnosis showed a significant association with higher mortality using gold standard data (OR 1.08; 95% CI 1.01 to 1.15; p=0.02) but not in HES-ONS data (OR 1.05; 95% CI 0.98 to 1.13; p=0.16). Otherwise, little change was found in the strength of associated risk factors and mortality after accounting for missed matching bias.

CONCLUSIONS

Despite significant clinical and sociodemographic differences between matched and unmatched records, changes in mortality estimates were minimal. However, researchers and policy analysts using HES-ONS linked resources should be aware that administrative linkage processes can introduce error.

The Development of the Military Service Identification Tool: Identifying Military Veterans in a Clinical Research Database Using Natural Language Processing and Machine Learning

Background

Electronic health care records (EHRs) are a rich source of health-related information, with potential for secondary research use. In the United Kingdom, there is no national marker for identifying those who have previously served in the Armed Forces, making analysis of the health and well-being of veterans using EHRs difficult.

Objective

This study aimed to develop a tool to identify veterans from free-text clinical documents recorded in a psychiatric EHR database.

Methods

Veterans were manually identified using the South London and Maudsley (SLaM) Biomedical Research Centre Clinical Record Interactive Search—a database holding secondary mental health care electronic records for the SLaM National Health Service Foundation Trust. An iterative approach was taken; first, a structured query language (SQL) method was developed, which was then refined using natural language processing and machine learning to create the Military Service Identification Tool (MSIT) to identify if a patient was a civilian or veteran. Performance, defined as correct classification of veterans compared with incorrect classification, was measured using positive predictive value, negative predictive value, sensitivity, F1 score, and accuracy (otherwise termed Youden Index).

Results

A gold standard dataset of 6672 free-text clinical documents was manually annotated by human coders. Of these documents, 66.00% (4470/6672) were then used to train the SQL and MSIT approaches and 34.00% (2202/6672) were used for testing the approaches. To develop the MSIT, an iterative 2-stage approach was undertaken. In the first stage, an SQL method was developed to identify veterans using a keyword rule–based approach. This approach obtained an accuracy of 0.93 in correctly predicting civilians and veterans, a positive predictive value of 0.81, a sensitivity of 0.75, and a negative predictive value of 0.95. This method informed the second stage, which was the development of the MSIT using machine learning, which, when tested, obtained an accuracy of 0.97, a positive predictive value of 0.90, a sensitivity of 0.91, and a negative predictive value of 0.98.

Conclusions

The MSIT has the potential to be used in identifying veterans in the United Kingdom from free-text clinical documents, providing new and unique insights into the health and well-being of this population and their use of mental health care services.

Maximizing the use of social and behavioural information from secondary care mental health electronic health records

PURPOSE

The contribution of social and behavioural factors in the development of mental health conditions and treatment effectiveness is widely supported, yet there are weak population level data sources on social and behavioural determinants of mental health. Enriching these data gaps will be crucial to accelerating precision medicine. Some have suggested the broader use of electronic health records (EHR) as a source of non-clinical determinants, although social and behavioural information are not systematically collected metrics in EHRs, internationally.

OBJECTIVE

In this commentary, we highlight the nature and quality of key available structured and unstructured social and behavioural data using a case example of value counts from secondary mental health data available in the UK from the UK Clinical Record Interactive Search (CRIS) database; highlight the methodological challenges in the use of such data; and possible solutions and opportunities involving the use of natural language processing (NLP) of unstructured EHR text.

CONCLUSIONS

Most structured non-clinical data fields within secondary care mental health EHR data have too much missing data for adequate use. The utility of other non-clinical fields reported semi-consistently (e.g., ethnicity and marital status) is entirely dependent on treating them appropriately in analyses, quantifying the many reasons behind missingness in consideration of selection biases. Advancements in NLP offer new opportunities in the exploitation of unstructured text from secondary care EHR data particularly given that clinical notes and attachments are available in large volumes of patients and are more routinely completed by clinicians. Tackling ways to re-use, harmonize, and improve our existing and future secondary care mental health data, leveraging advanced analytics such as NLP is worth the effort in an attempt to fill the data gap on social and behavioural contributors to mental health conditions and will be necessary to fulfill all of the domains needed to inform personalized interventions.

Investigating Bullying as a Predictor of Suicidality in a Clinical Sample of Adolescents with Autism Spectrum Disorder

For typically developing adolescents, being bullied is associated with increased risk of suicidality. Although adolescents with autism spectrum disorder (ASD) are at increased risk of both bullying and suicidality, there is very little research that examines the extent to which an experience of being bullied may increase suicidality within this specific population. To address this, we conducted a retrospective cohort study to investigate the longitudinal association between experiencing bullying and suicidality in a clinical population of 680 adolescents with ASD. Electronic health records of adolescents (13–17 years), using mental health services in South London, with a diagnosis of ASD were analyzed. Natural language processing was employed to identify mentions of bullying and suicidality in the free text fields of adolescents' clinical records. Cox regression analysis was employed to investigate the longitudinal relationship between bullying and suicidality outcomes. Reported experience of bullying in the first month of clinical contact was associated with an increased risk suicidality over the follow‐up period (hazard ratio = 1.82; 95% confidence interval = 1.28–2.59). In addition, female gender, psychosis, affective disorder diagnoses, and higher intellectual ability were all associated with suicidality at follow‐up. This study is the first to demonstrate the strength of longitudinal associations between bullying and suicidality in a clinical population of adolescents with ASD, using automated approaches to detect key life events within clinical records. Our findings provide support for identifying and dealing with bullying in schools, and for antibullying strategy's incorporation into wider suicide prevention programs for young people with ASD. Autism Res 2020, 13: 988‐997. © 2020 The Authors. Autism Research published by International Society for Autism Research published by Wiley Periodicals, Inc.

Five models for child and adolescent data linkage in the UK: a review of existing and proposed methods

Over the last decade dramatic advances have been made in both the technology and data available to better understand the multifactorial influences on child and adolescent health and development. This paper seeks to clarify methods that can be used to link information from health, education, social care and research datasets. Linking these different types of data can facilitate epidemiological research that investigates mental health from the population to the patient; enabling advanced analytics to better identify, conceptualise and address child and adolescent needs. The majority of adolescent mental health research is not able to maximise the full potential of data linkage, primarily due to four key challenges: confidentiality, sampling, matching and scalability. By presenting five existing and proposed models for linking adolescent data in relation to these challenges, this paper aims to facilitate the clinical benefits that will be derived from effective integration of available data in understanding, preventing and treating mental disorders.

Co-occurring obsessive-compulsive disorder and autism spectrum disorder in young people: prevalence, clinical characteristics and outcomes

Obsessive-compulsive disorder (OCD) and autism spectrum disorders (ASD) commonly co-occur and are considered challenging to manage when they co-occur in youth. However, clinical characteristics and prognosis of this group remain poorly understood. This study examined the prevalence, clinical correlates and outcomes of paediatric OCD co-occurring with ASD (OCD + ASD) in a large clinical cohort. Data were extracted from electronic clinical records of young people aged 4-17 years who had attended a mental health trust in South London, United Kingdom. We identified young people with diagnoses of OCD + ASD (n = 335), OCD without ASD (n = 1010), and ASD without OCD (n = 6577). 25% of youth with OCD had a diagnosis of ASD, while 5% of those with ASD had a diagnosis of OCD. At diagnosis, youth with OCD + ASD had lower psychosocial functioning scores on the clinician-rated Child Global Assessment Scale (CGAS) compared to those with either OCD or ASD. Youth with OCD + ASD were equally likely to receive CBT compared to those with OCD but were more likely to be prescribed medication and use services for longer than either comparison group. Youth with OCD + ASD showed significant improvements in functioning (CGAS scores) after service utilisation but their gains were smaller than those with OCD. OCD + ASD commonly co-occur, conferring substantial impairment, although OCD may be underdiagnosed in youth with ASD. Young people with co-occurring OCD + ASD can make significant improvements in functioning with routine clinical care but are likely to remain more impaired than typically developing youth with OCD, indicating a need for longer-term support for these young people.

Diagnostic trajectories in child and adolescent mental health services: exploring the prevalence and patterns of diagnostic adjustments in an electronic mental health case register

Community-based epidemiological studies show transitions between psychiatric disorders are common during child development. However, little research has explored the prevalence or patterns of the diagnostic adjustments that occur in child and adolescent mental health services (CAMHS). Understanding diagnostic trajectories is necessary to inform theory development in developmental psychopathology and clinical judgements regarding risk and prognosis. In this study, data from CAMHS clinical records were extracted from a British mental health case register (N = 12,543). Analysis calculated the proportion of children whose clinical records showed a longitudinal diagnostic adjustment (i.e. addition of a subsequent diagnosis of a different diagnostic class, at > 30 days' distance from their first diagnosis). Regression analyses investigated typical diagnostic sequences and their relationships with socio-demographic variables, service use and standardised measures of mental health. Analysis found that 19.3% of CAMHS attendees had undergone a longitudinal diagnostic adjustment. Ethnicity, diagnostic class and symptom profiles significantly influenced the likelihood of a diagnostic adjustment. Affective and anxiety/stress-related disorders longitudinally predicted each other, as did hyperkinetic and conduct disorders, and hyperkinetic and pervasive developmental disorders. Results suggest that approximately one in five young service users have their original psychiatric diagnosis revised or supplemented during their time in CAMHS. By revealing the most common diagnostic sequences, this study enables policy makers to anticipate future service needs and clinicians to make informed projections about their patients' likely trajectories. Further research is required to understand how young people experience diagnostic adjustments and their psychological and pragmatic implications.

Identifying Veterans Using Electronic Health Records in the United Kingdom: A Feasibility Study

There is a lack of quantitative evidence concerning UK (United Kingdom) Armed Forces (AF) veterans who access secondary mental health care services-specialist care often delivered in high intensity therapeutic clinics or hospitals-for their mental health difficulties. The current study aimed to investigate the utility and feasibility of identifying veterans accessing secondary mental health care services using National Health Service (NHS) electronic health records (EHRs) in the UK. Veterans were manually identified using the Clinical Record Interactive Search (CRIS) system-a database holding secondary mental health care EHRs for an NHS Trust in the UK. We systematically and manually searched CRIS for veterans, by applying a military-related key word search strategy to the free-text clinical notes completed by clinicians. Relevant data on veterans' socio-demographic characteristics, mental disorder diagnoses and treatment pathways through care were extracted for analysis. This study showed that it is feasible, although time consuming, to identify veterans through CRIS. Using the military-related key word search strategy identified 1600 potential veteran records. Following manual review, 693 (43.3%) of these records were verified as "probable" veterans and used for analysis. They had a median age of 74 years (interquartile range (IQR): 53-86); the majority were male (90.8%) and lived alone (38.0%). The most common mental diagnoses overall were depressive disorders (22.9%), followed by alcohol use disorders (10.5%). Differences in care pathways were observed between pre and post national service (NS) era veterans. This feasibility study represents a first step in showing that it is possible to identify veterans through free-text clinical notes. It is also the first to compare veterans from pre and post NS era.

The association between depression and later educational attainment in children and adolescents: a systematic review protocol

INTRODUCTION

Depression represents a major public health concern for children and adolescents, and is thought to negatively impact subsequent educational attainment. However, the extent to which depression and educational attainment are directly associated, and whether other factors play a role, is uncertain. Therefore, we aim to systematically review the literature to provide an up-to-date estimate on the strength of this association, and to summarise potential mediators and moderators on the pathway between the two.

METHODS AND ANALYSIS

To identify relevant studies, we will systematically search Embase, PsycINFO, PubMed, Education Resources Information Centre and British Education Index, manually search reference lists and contact experts in the field. Studies will be included if they investigate and report on the association between major depression diagnosis or depressive symptoms in children and adolescents aged 4-18 years (exposure) and later educational attainment (outcome). Two independent reviewers will screen titles, abstracts and full texts according to eligibility criteria, perform data extraction and assess study quality according to a modified version of the Newcastle-Ottawa Scale. If sufficiently homogeneous studies are identified, summary effect estimates will be pooled in meta-analysis, with further tests for study heterogeneity, publication bias and the effects of moderators using meta-regression.

ETHICS AND DISSEMINATION

Because this review will make use of already published data, ethical approval will not be sought. The review will be submitted for publication in a peer-reviewed journal, presented at practitioner-facing conferences, and a lay summary will be written for non-scientific audiences such as parents, young people and teachers. The work will inform upcoming investigations on the association between child and adolescent mental health and educational attainment.

PROSPERO REGISTRATION NUMBER

CRD42019123068.

Data Resource: the National Pupil Database (NPD)

Introduction

The National Pupil Database (NPD) is a record-level administrative data resource curated by the UK government’s Department for Education that is used for funding purposes, school performance tables, policy making, and research.

Processes

Data are sourced from schools, exam awarding bodies, and local authorities who collect data on an on-going basis and submit to the Department for Education either termly or yearly.

Data contents

NPD contains child-level and school-level data on all pupils in state schools in England (6.6 million in the 2016/17 academic year). The primary module is the census, which has information on characteristics and school enrolment. Other modules include alternative provision, exam attainment, absence and exclusions. Data from children’s social care are also available on children referred for support and those who become looked after. Children’s records are linkable across different modules and across time using a nationally unique, anonymised child-level identifier. Linkage to external datasets has also been accomplished using child-level identifiers.

Conclusions

The NPD is an especially valuable data resource for researchers interested in the educational experience and outcomes of children and young people in England. Although limited by the fact that children in private schools or who are home schooled are not included, it provides a near-complete picture of school trajectories and outcomes for the majority of children. Linkage to other datasets can enhance analyses and provide answers to questions that would otherwise be costly, time consuming and difficult to find