Clinical Record Interactive Search Research Articles

32. Demographic and Clinical Predictors of Duration of Untreated Psychosis

Background: A shorter duration of untreated psychosis (DUP) is associated with better outcomes following first-episode psychosis (FEP; Power et al., 2007; Birchwood et al., 2013; McGorry et al., 2008). However, the evidence on social and clinical factors that may predict to DUP is inconsistent. Aims: To investigate the association between DUP and social and clinical factors. Methods: A retrospective incidence study design was employed, using the Biomedical Research Centre (BRC) clinical record interactive search (CRIS) system. In brief, CRIS is a regional case register based in South London containing a large data set of anonymous clinical data of over 250 000 patient records derived from the South London and Maudsley NHS Foundation Trust (SLaM) electronic health record system. All patients presenting to SLaM adult mental health services for the first time with a psychotic disorder between May 2010 and April 2012 and in the catchment area served by SLaM were screened for inclusion. Data relating to DUP, sociodemographic characteristics, mode of contact, and source of referral were collated from clinical records. Individuals were included as cases if they were: resident in the London boroughs of Lambeth or Southwark (served by SLaM); aged 18–64 years (inclusive), experienced psychotic symptoms of at least one day duration, and were making their first contact for psychosis with mental health services. Results: A total of 558 individuals with first-episode psychosis were identified. Individuals who were unemployed (68.5%) experienced longer DUP (median = 119; interquartile range= 28–492) compared to people who were employed (19.6%; median =45; interquartile range= 6–426, P < .001). Living alone (29.7%) was also associated with longer DUP (median = 116; interquartile range = 30–531) compared to living with family/relative (median = 90; interquartile range = 14–370, P = .04). Further, an insidious mode of onset of psychosis (37.5%) was associated with longer DUP (median = 608; interquartile range = 360–1918) compared to acute onset (20.7%; median = 4; interquartile range = 2–6.5, P < .001). Similarly, individuals accessing care via accident and emergency departments (AE interquartile range = 6–295) compared with those referred by their general practitioner (GP 35.1%; Median=184.5; interquartile range = 46.5–821.5, P < .001). Conclusion: Findings from this sample of FEP patients suggest that indicators of social isolation were associated with DUP. Clinically, pathways into care were also strongly associated with DUP prior to help seeking. Our results are consistent with previous findings.

Linking health and education data to plan and evaluate services for children

Linkage of routinely collected data from public services has the potential to improve how local health, education and social care are delivered to children. All mental health services, hospital-based child...

Natural language processing to extract symptoms of severe mental illness from clinical text: the Clinical Record Interactive Search Comprehensive Data Extraction (CRIS-CODE) project

ObjectivesWe sought to use natural language processing to develop a suite of language models to capture key symptoms of severe mental illness (SMI) from clinical text, to facilitate the secondary...

Admission to acute mental health services after contact with crisis resolution and home treatment teams: an investigation in two large mental health-care providers

Crisis resolution and home treatment teams (CRTs) offer an alternative to hospital admission for patients undergoing mental health crises in the UK. Few studies have been done to examine predictors of relapse and readmission after contact with CRTs. We used the Clinical Record Interactive Search to identify all patients receiving care from CRTs in two National Health Service (NHS) mental health trusts in London: Camden and Islington NHS Foundation Trust and South London and Maudsley NHS Foundation Trust. We used Cox regression models to examine rates and predictors of admission to acute mental health services within 1 year of contact with CRTs. Sex, age, ethnicity, marital status, social deprivation, severity of psychopathology, duration of index CRT episode, first contact with services, and diagnosis were extracted and examined as predictors of admission. Between Jan 1, 2008, and Aug 31, 2014, 17 666 patients were treated by CRTs-8759 patients in the Camden and Islington trust and 8907 patients in the South London and Maudsley trust. 53·9 patients per 100 person-years (95% CI 52·1-55·8) in Camden and Islington and 51·3 patients per 100 person-years (95% CI 49·6-53·1) in South London and Maudsley were admitted to acute services within 1 year of seeing the CRT. In both cohorts, non-affective psychotic disorders (adjusted hazard ratio [HR] 1·25, 95% CI 1·09-1·44 in Camden and Islington; 1·27, 1·17-1·38 in South London and Maudsley) and age older than 65 years (1·18, 1·01-1·37 in Camden and Islington; 1·32, 1·12-1·56 in South London and Maudsley) were associated with increased risk of admission, whereas first contact with services (0·57, 0·52-0·62 in Camden and Islington; 0·69, 0·63-0·75 in South London and Maudsley), anxiety disorders (0·81, 0·69-0·96 in Camden and Islington; 0·77, 0·67-0·87 in South London and Maudsley), and longer index CRT episodes (adjusted HR per day 0·996, 0·994-0·998 in Camden and Islington; 0·989, 0·987-0·991 in South London and Maudsley) were associated with reduced risk of admission. Past use of mental health services and a diagnosis of non-affective psychosis, which are markers of severity of mental illness, and older age, which is a marker of chronicity, are all risk factors for future relapse after interactions with CRTs. These findings might help clinicians and policy makers to offer more targeted and cost-effective services to reduce relapse rates. None.

Individual and area-level risk factors for suicidal ideation and attempt in people with severe depression

IntroductionPrevious research has identified several risk factors that are strongly associated with suicidal behavior in patients with severe depression. However, the effects of area-level characteristics on suicidal ideation and attempt in this population remain unclear. MethodsThe Clinical Record Interactive Search (CRIS) database was used to identify 2587 patients with severe depression who received secondary mental health services from the Camden & Islington NHS Foundation Trust. Stepwise multivariable logistic regression models were used to examine associations between socio-demographic characteristics, clinical variables, area-level measures, and suicidal ideation and attempt as separate outcomes. ResultsBoth suicidal ideation and attempts were common among this cohort of severely depressed individuals (70.5% and 37.7%, respectively). While several individual socio-demographic and clinical characteristics were associated with both outcomes, particularly past psychiatric admission (suicidal ideation: adjusted OR=2.86, 95% CI: 2.26–3.62; suicide attempt: adjusted OR=4.00, 95% CI: 3.30–4.89), neither social deprivation nor ethnic density (measured at the area-level) was associated with risk for either outcome. LimitationsData were not collected specifically for research purposes and hence information on some potential confounders was not available. Additionally, information was restricted to individuals who accessed secondary mental health services in a defined catchment area and period. The study therefore does not take into account individuals who did not access mental health services. ConclusionsThe variation in risk for suicidal ideation and attempt among severely depressed individuals is explained by differences in individual socio-demographic and clinical characteristics, most notably past psychiatric admission and substance misuse, and not by area-level measures.

ELECTRONIC RECORDS FOR DEMENTIA RESEARCH: DO BEHAVIOURAL DISTURBANCES, ANTIHYPERTENSIVES AND ANTIDEPRESSANTS INFLUENCE DECLINE TRAJECTORIES?

Background: Understanding the drivers of heterogeneous progression in dementia has huge implications for recruitment to clinical trails and care planning. One way to investigate the factors contributing to this heterogeneity is to stratify subjects by similar patterns of change and compare the characteristics of the sub-populations identified. We applied this approach to routinely collected electronic health records from the South London and Maudesly NHS Foundation Trust (SLAM). Methods: This retrospective study includes 3441 patients with at least three MMSE scores recorded and available for research through the SLAM clinical record interactive search system (CRIS). A Latent Class Growth Analysis was used to identify key trajectories of decline. Information on age, gender, ethnicity, qualification levels, cohabiting status, retirement status, Health of Nation Outcome Scales (HoNOS) and medications were obtained to describe characteristics of the identified trajectory sub-populations in a multivariable multinomial regression analysis. Results: We identified six trajectories of cognitive decline. Four of these trajectories differed in initial MMSE score, and showed increased rate of decline with lower initial MMSE. Two trajectories had very similar initial MMSE scores but differed in the rate of decline. Exploring these trajectories further, the severity of cognitive problems at baseline and prescription of Donepezil and Amlodipine was significantly higher in the slower declining trajectory. In the faster declining trajectory, severity of behavioral problems and prescription of sertraline was significantly higher. Conclusions: Most of the trajectories differed by initial MMSE and thus likely represented patients at different disease stages, however differences in behavioral problems, antihypertensive, antidepressant and dementia medication prescription may be indicators of future rate of cognitive decline. Further information is required on depression and hypertensive comorbidities to explore these findings in greater detail.

Cohort profile of the South London and Maudsley NHS Foundation Trust Biomedical Research Centre (SLaM BRC) Case Register: current status and recent enhancement of an Electronic Mental Health Record-derived data resource

PurposeThe South London and Maudsley National Health Service (NHS) Foundation Trust Biomedical Research Centre (SLaM BRC) Case Register and its Clinical Record Interactive Search (CRIS) application were developed in 2008,...

Antipsychotic polypharmacy and augmentation strategies prior to clozapine initiation: a historical cohort study of 310 adults with treatment-resistant schizophrenic disorders.

Antipsychotic polypharmacy (APP) is commonly used in schizophrenia despite a lack of robust evidence for efficacy, as well as evidence of increased rates of adverse drug reactions and mortality. We sought to examine APP and the use of other adjunctive medications in patients with treatment-resistant schizophrenic disorders (ICD-10 diagnoses F20-F29) immediately prior to clozapine initiation, and to investigate clinical and sociodemographic factors associated with APP use in this setting. Analysis of case notes from 310 patients receiving their first course of clozapine at the South London and Maudsley NHS Trust (SLaM) was undertaken using the Clinical Record Interactive Search (CRIS) case register. Medication taken immediately prior to clozapine initiation was recorded, and global clinical severity was assessed at time points throughout the year prior to medication assessment using the Clinical Global Impression - Severity scale (CGI-S). Logistic regression was used to investigate factors associated with APP. The point prevalence of APP prior to clozapine initiation was 13.6% (n=42), with 32.6% of subjects prescribed adjuvant psychotropic medications. APP was associated with increasing number of adjuvant medications (odds ratio (OR) 1.97, 95% confidence interval (CI) 1.27-3.06), concurrent depot antipsychotic prescription (OR 2.64, CI 1.24-5.62), concurrent antidepressant prescription (OR 4.40, CI 1.82-10.63) and a CGI-S over the previous year within the two middle quartiles (Quartile 2 vs 1 OR 6.19, CI 1.81-21.10; Quartile 3 vs 1 OR 4.45, CI 1.29-15.37; Quartile 4 vs 1 OR 1.88, CI 0.45-7.13). APP and augmentation of antipsychotics with antidepressants, mood stabilizers and benzodiazepines are being employed in treatment-resistant schizophrenia prior to clozapine. The conservative APP rate observed may have been influenced by an initiative within SLaM that reduced APP rates during the study window. Efforts to reduce the use of poorly evidenced prescribing should focus on adjuvant medications as well as APP.

Mortality of people with chronic fatigue syndrome: a retrospective cohort study in England and Wales from the South London and Maudsley NHS Foundation Trust Biomedical Research Centre (SLaM BRC) Clinical Record Interactive Search (CRIS) Register.

Mortality associated with chronic fatigue syndrome is uncertain. We investigated mortality in individuals diagnosed with chronic fatigue syndrome in secondary and tertiary care using data from the South London and Maudsley NHS Foundation Trust Biomedical Research Centre (SLaM BRC) Clinical Record Interactive Search (CRIS) register. We calculated standardised mortality ratios (SMRs) for all-cause, suicide-specific, and cancer-specific mortality for a 7-year observation period using the number of deaths observed in SLaM records compared with age-specific and sex-specific mortality statistics for England and Wales. Study participants were included if they had had contact with the chronic fatigue service (referral, discharge, or case note entry) and received a diagnosis of chronic fatigue syndrome. We identified 2147 cases of chronic fatigue syndrome from CRIS and 17 deaths from Jan 1, 2007, to Dec 31, 2013. 1533 patients were women of whom 11 died, and 614 were men of whom six died. There was no significant difference in age-standardised and sex-standardised mortality ratios (SMRs) for all-cause mortality (SMR 1·14, 95% CI 0·65-1·85; p=0·67) or cancer-specific mortality (1·39, 0·60-2·73; p=0·45) in patients with chronic fatigue syndrome when compared with the general population in England and Wales. This remained the case when deaths from suicide were removed from the analysis. There was a significant increase in suicide-specific mortality (SMR 6·85, 95% CI 2·22-15·98; p=0·002). We did not note increased all-cause mortality in people with chronic fatigue syndrome, but our findings show a substantial increase in mortality from suicide. This highlights the need for clinicians to be aware of the increased risk of completed suicide and to assess suicidality adequately in patients with chronic fatigue syndrome. National Institute for Health Research (NIHR) Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King's College London.

Clinical outcomes and mortality associated with weekend admission to psychiatric hospital: implications for provision of 7 day mental health care in the National Health Service

BackgroundPrevious studies indicate that risk of mortality is higher for patients admitted to acute hospitals at the weekend. Thus, there is a drive to move towards provision of a 7 day health-care service in the UK. However, less is known about variations in clinical outcomes among patients admitted to psychiatric hospitals. We sought to investigate variations in clinical outcomes for patients admitted to a psychiatric hospital at the weekend. MethodsThe study was conducted with anonymised electronic health record data from the South London and Maudsley NHS Foundation Trust. Data were obtained from all consecutive admissions to psychiatric hospitals within the Trust between April 1, 2006, and March 31, 2015, with the Clinical Record Interactive Search tool. The association of weekend or bank holiday admission with duration of admission, risk of readmission, and inpatient mortality were investigated by multivariable regression analyses with age, sex, and ethnicity as covariates. Secondary analyses were performed to investigate the distribution of admissions, discharges, inpatient mortality, episodes of seclusion, and violent incidents on different days of the week. Findings7303 (16·1%) of 45 264 admissions commenced during a weekend or bank holiday. Patients who were aged between 26 and 35 years, female, or from a minority ethnic group were more likely to be admitted at the weekend. Patients admitted at the weekend were more likely to present via acute hospital services, other psychiatric hospitals, or the criminal justice system than to be admitted directly from their own home. Weekend admission was associated with a shorter duration of admission (β coefficient −21·1 days, 95% CI −24·6 to −17·6; p<0·001) and increased risk of readmission in the 12 months after index admission (incidence rate ratio 1·13, 1·08–1·18; p<0·001) but no significant difference in inpatient mortality (odds ratio 0·79, 0·51–1·23) compared with weekday admission. Fewer episodes of seclusion occurred at the weekend (χ2=15·8, p=0·02) but there was no significant variation in deaths during hospital admission or violent incidents on different days of the week. InterpretationWe show that patients admitted at the weekend had shorter admissions and increased risk of readmission but no significant difference in risk of inpatient mortality compared with those admitted during the week. This finding contrasts with previous studies in acute hospitals where weekend admission is typically associated with increased likelihood of inpatient mortality, and highlights a need to consider variations in demand when implementing health-care services across a 7 day week. FundingNational Institute for Health Research, Medical Research Council.

Identification of Adverse Drug Events from Free Text Electronic Patient Records and Information in a Large Mental Health Case Register.

ObjectivesElectronic healthcare records (EHRs) are a rich source of information, with huge potential for secondary research use. The aim of this study was to develop an application to identify instances of Adverse Drug Events (ADEs) from free text psychiatric EHRs.MethodsWe used the GATE Natural Language Processing (NLP) software to mine instances of ADEs from free text content within the Clinical Record Interactive Search (CRIS) system, a de-identified psychiatric case register developed at the South London and Maudsley NHS Foundation Trust, UK. The tool was built around a set of four movement disorders (extrapyramidal side effects [EPSEs]) related to antipsychotic therapy and rules were then generalised such that the tool could be applied to additional ADEs. We report the frequencies of recorded EPSEs in patients diagnosed with a Severe Mental Illness (SMI) and then report performance in identifying eight other unrelated ADEs.ResultsThe tool identified EPSEs with >0.85 precision and >0.86 recall during testing. Akathisia was found to be the most prevalent EPSE overall and occurred in the Asian ethnic group with a frequency of 8.13%. The tool performed well when applied to most of the non-EPSEs but least well when applied to rare conditions such as myocarditis, a condition that appears frequently in the text as a side effect warning to patients.ConclusionsThe developed tool allows us to accurately identify instances of a potential ADE from psychiatric EHRs. As such, we were able to study the prevalence of ADEs within subgroups of patients stratified by SMI diagnosis, gender, age and ethnicity. In addition we demonstrated the generalisability of the application to other ADE types by producing a high precision rate on a non-EPSE related set of ADE containing documents.AvailabilityThe application can be found at http://git.brc.iop.kcl.ac.uk/rmallah/dystoniaml.

Using routine clinical and administrative data to produce a dataset of attendances at Emergency Departments following self-harm.

BackgroundSelf-harm is a significant public health concern in the UK. This is reflected in the recent addition to the English Public Health Outcomes Framework of rates of attendance at Emergency Departments (EDs) following self-harm. However there is currently no source of data to measure this outcome. Routinely available data for inpatient admissions following self-harm miss the majority of cases presenting to services.We aimed to investigate (i) if a dataset of ED presentations could be produced using a combination of routinely collected clinical and administrative data and (ii) to validate this dataset against another one produced using methods similar to those used in previous studies.MethodsUsing the Clinical Record Interactive Search system, the electronic health records (EHRs) used in four EDs were linked to Hospital Episode Statistics to create a dataset of attendances following self-harm. This dataset was compared with an audit dataset of ED attendances created by manual searching of ED records. The proportion of total cases detected by each dataset was compared.ResultsThere were 1932 attendances detected by the EHR dataset and 1906 by the audit. The EHR and audit datasets detected 77 % and 76 % of all attendances respectively and both detected 82 % of individual patients. There were no differences in terms of age, sex, ethnicity or marital status between those detected and those missed using the EHR method. Both datasets revealed more than double the number of self-harm incidents than could be identified from inpatient admission records.ConclusionsIt was possible to use routinely collected EHR data to create a dataset of attendances at EDs following self-harm. The dataset detected the same proportion of attendances and individuals as the audit dataset, proved more comprehensive than the use of inpatient admission records, and did not show a systematic bias in those cases it missed.

Developing Electronic Health Records Databases for Mental Health Research

Mental health records are increasingly kept in fully electronic format. This offers potentially transformative research opportunities because of both the very large samples and the depth of the data contained. However, there are clearly substantial challenges in ensuring both adequate anonymity and robust governance for these sensitive data. In addition, the depth of information is often not exploited because most of this is contained in text rather than structured fields. Natural language processing (NLP) offers a potential solution to this, but NLP application in health records data is in its infancy. The Clinical Record Interactive Search (CRIS) application will be discussed, which was developed at the Maudsley Hospital in south London and which offers a robust and patient-led de-identification and governance model for records-based research. NLP applications have now been developed which comprehensively profile a range of symptoms, interventions and outcomes in routine mental healthcare, providing both breadth (250,000 cases) and depth of information, and supporting both novel research output and decision support for clinical services.

Cannabis use and treatment resistance in first episode psychosis: a natural language processing study

BackgroundCannabis is frequently used among individuals with first episode psychosis and is associated with poor clinical outcomes. However, little is known about the effect of cannabis use on the response to antipsychotic medications and how use could affect outcomes. Using natural language processing on clinical data from a large electronic case register, we sought to investigate whether resistance to antipsychotic treatment mediated poor clinical outcomes associated with cannabis use. MethodsData were obtained from 2026 people with first episode psychosis in south London, UK. Cannabis use documented in free text clinical records was identified with natural language processing. Data for age, sex, ethnicity, marital status, psychotic disorder diagnosis, subsequent hospital admission, and number of unique antipsychotic medications prescribed were obtained using the Clinical Record Interactive Search instrument. The association of these variables with cannabis use was analysed with multivariable regression and mediation analysis. Findings939 people (46·3%) with first episode psychosis were using cannabis at first presentation. Cannabis use was most strongly associated with being 16–25 years old, male, and single, and was also associated with an increase in number of hospital admissions (incidence rate ratio 1·50, 95% CI 1·25–1·80), compulsory hospital admission (odds ratio 1·55, 1·16–2·08), and number of days spent in hospital (β coefficient 35·1 days, 12·1–58·1) over 5 years' follow-up. An increase in number of unique antipsychotic medications mediated an increase in number of hospital admissions (natural indirect effect 1·11, 1·04–1·17; total effect 1·41, 1·22–1·64), compulsory hospital admission (1·27, 1·10–1·45; 1·71, 1·05–2·78), and number of days spent in hospital (16·1, 6·7–25·5; 19·9, 2·5–37·3). InterpretationWe showed that a substantial number of people with first episode psychosis used cannabis and that its use was associated with increased likelihood of hospital admission and number of days spent in hospital. These associations were partly mediated by an increase in number of unique antipsychotic medications prescribed. These findings suggest that cannabis might reduce response to conventional antipsychotic treatment and highlight the importance of strategies to reduce its use. FundingNational Institute for Health Research, UK Medical Research Council.

Predictors of mortality for people aged over 65 years receiving mental health care for delirium in a South London Mental Health Trust, UK: a retrospective survival analysis.

Delirium is a common phenomenon in older people. Using a large mental health care data resource, we investigated mortality rates and predictors of mortality following delirium in older people. The South London and Maudsley NHS Foundation Trust (SLAM) Clinical Record Interactive Search (CRIS) was used to retrieve anonymised data on patients known to mental health services who were over 65 years of age and received a diagnosis of delirium during a 3-year period. Age-standardised and gender-standardised mortality rates (SMRs) were calculated, and predictors of survival were investigated considering demographic factors, health status rated on the Health of the National Outcome Scale (HoNOS), cognitive function and previous or contemporaneous diagnosed dementia. In 974 patients with delirium, 1- and 3-year mortality rates were 37.2 and 54.9% respectively, representing an SMR of 4.7 overall (95% CI: 4.3-5.1). SMR was 5.2 (95% CI: 4.6-5.7) for patients with delirium without prior dementia; SMR was 4.1 (95% CI: 3.6-4.7) for patients with dementia preceding delirium and 2.2 (2.0-2.5) excluding deaths within 6 months of the delirium diagnosis. Significant predictors of mortality in fully adjusted models were older age, male gender, white (compared with non-white) ethnicity, and HoNOS subscales measuring physical ill-health and functional impairment. No mortality associations were found with cognitive function, dementia, or psychological symptoms. In people with delirium diagnosed by mental health services, mortality risk was high and predicted by demographic and physical health status rather than by cognitive function or psychological profile.

Associations between risk of mortality and atypical antipsychotic use in vascular dementia: a clinical cohort study

People with vascular dementia (VaD) are frequently prescribed atypical antipsychotics to treat behavioural and psychological symptoms, but there is an alarming lack of evidence regarding their safety or efficacy in VaD. This study sought to identify the mortality risk associated with the most commonly prescribed atypical antipsychotics in people with VaD compared with people not exposed to these drugs. A clinical cohort study of 1531 people with VaD performed using anonymised versions of full electronic health records from the Clinical Record Interactive Search application at the South London and Maudsley NHS Foundation Trust. Patients were identified from 2007 to 2010, of whom 337 were exposed to quetiapine, risperidone or olanzapine. The main outcome measure was mortality. Patients exposed to atypical antipsychotics were not at increased risk of mortality [hazard ratio (HR) 1.05, 95% confidence interval (CI): 0.87-1.26]. Exposure to risperidone did not result in an increased risk of mortality (HR = 0.85; 95% CI: 0.59-1.24), and patients exposed to quetiapine had a non-significant numerical increase in mortality risk (HR = 1.14; 95% CI: 0.93-1.39; p-value = 0.20) compared with untreated patients. Too few patients were exposed to olanzapine alone to provide reliable results. The absence of a significant increase in mortality risk associated with atypical antipsychotics in people with VaD indicates that a clinical trial of antipsychotics focussing on the treatment of aggression and agitation in this patient group will be justified and feasible following further consideration of possible confounders, which will be critical to determine the role of antipsychotics in treatment of VaD.

Development and evaluation of a de-identification procedure for a case register sourced from mental health electronic records

BackgroundElectronic health records (EHRs) provide enormous potential for health research but also present data governance challenges. Ensuring de-identification is a pre-requisite for use of EHR data without prior consent. The South London and Maudsley NHS Trust (SLaM), one of the largest secondary mental healthcare providers in Europe, has developed, from its EHRs, a de-identified psychiatric case register, the Clinical Record Interactive Search (CRIS), for secondary research.MethodsWe describe development, implementation and evaluation of a bespoke de-identification algorithm used to create the register. It is designed to create dictionaries using patient identifiers (PIs) entered into dedicated source fields and then identify, match and mask them (with ZZZZZ) when they appear in medical texts. We deemed this approach would be effective, given high coverage of PI in the dedicated fields and the effectiveness of the masking combined with elements of a security model. We conducted two separate performance tests i) to test performance of the algorithm in masking individual true PIs entered in dedicated fields and then found in text (using 500 patient notes) and ii) to compare the performance of the CRIS pattern matching algorithm with a machine learning algorithm, called the MITRE Identification Scrubber Toolkit – MIST (using 70 patient notes – 50 notes to train, 20 notes to test on). We also report any incidences of potential breaches, defined by occurrences of 3 or more true or apparent PIs in the same patient’s notes (and in an additional set of longitudinal notes for 50 patients); and we consider the possibility of inferring information despite de-identification.ResultsTrue PIs were masked with 98.8% precision and 97.6% recall. As anticipated, potential PIs did appear, owing to misspellings entered within the EHRs. We found one potential breach. In a separate performance test, with a different set of notes, CRIS yielded 100% precision and 88.5% recall, while MIST yielded a 95.1% and 78.1%, respectively. We discuss how we overcome the realistic possibility – albeit of low probability – of potential breaches through implementation of the security model.ConclusionCRIS is a de-identified psychiatric database sourced from EHRs, which protects patient anonymity and maximises data available for research. CRIS demonstrates the advantage of combining an effective de-identification algorithm with a carefully designed security model. The paper advances much needed discussion of EHR de-identification – particularly in relation to criteria to assess de-identification, and considering the contexts of de-identified research databases when assessing the risk of breaches of confidential patient information.

Use of electronic health records in identifying drug and alcohol misuse among psychiatric in-patients

Aims and methodTo assess the usefulness of the electronic patient record, we used the search engine Clinical Record Interactive Search (CRIS) to scan all acute admissions during 2008 for possible substance use disorders. In addition, screening interviews were undertaken with 75 in-patients, and documentation in their files was compared with results of screening interviews.ResultsOf 839 acute admissions during 2008, 47% of males and 29% of females had reference to a substance misuse problem in their file. Documentation was unsystematic and inconsistent and mostly occurred in progress notes rather than in structured questionnaires. Screening interviews and manual review of files of 75 current in-patients confirmed that substance use disorders were common, but poorly documented.Clinical implicationsThe study highlights the power of search engines in scanning electronic clinical records, but also identified the limitations of unsystematic documentation in research and practice. Mental health staff were reluctant to diagnose or rate severity of substance misuse problems.

HORMONES OF THE ANTERIOR LOBE OF THE PITUITARY GLAND

<h3>Abstract</h3> <h3>Objectives</h3> The recent COVID-19 pandemic has disrupted mental healthcare delivery, with many services shifting from in- person to remote patient contact. We investigated the impact of the pandemic on the use of remote consultation and on the prescribing of psychiatric medications. <h3>Design and setting</h3> The Clinical Record Interactive Search tool (CRIS) was used to examine de-identified electronic health records (EHRs) of people receiving mental healthcare from the South London and Maudsley (SLaM) NHS Foundation Trust. Data from the period before and after the onset of the pandemic were analysed using linear regression, and visualised using locally estimated scatterplot smoothing (LOESS). <h3>Participants</h3> All patients receiving care from SLaM between 7^th January 2019 and 20^th September 2020 (around 37,500 patients per week). <h3>Outcome measures</h3> The number of clinical contacts (in-person, remote or non-attended) with mental healthcare professionals per week Prescribing of antipsychotic and mood stabiliser medications per week <h3>Results</h3> Following the onset of the pandemic, the frequency of in-person contacts was significantly reduced compared to that in the previous year (β coefficient: −5829.6 contacts, 95% CI −6919.5 to −4739.6, p&lt;0.001), while the frequency of remote contacts significantly increased (β coefficient: 3338.5 contacts, 95% CI 3074.4 to 3602.7, p&lt;0.001). Rates of remote consultation were lower in older adults than in working age adults, children and adolescents. Despite this change in the type of patient contact, antipsychotic and mood stabiliser prescribing remained at similar levels. <h3>Conclusions</h3> The COVID-19 pandemic has been associated with a marked increase in remote consultation, particularly among younger patients. However, there was no evidence that this has led to changes in psychiatric prescribing. Nevertheless, further work is needed to ensure that older patients are able to access mental healthcare remotely.