Dataset Information

Using Neural Networks with Routine Health Records to Identify Suicide Risk: Feasibility Study.

ABSTRACT:

Background

Each year, approximately 800,000 people die by suicide worldwide, accounting for 1-2 in every 100 deaths. It is always a tragic event with a huge impact on family, friends, the community and health professionals. Unfortunately, suicide prevention and the development of risk assessment tools have been hindered by the complexity of the underlying mechanisms and the dynamic nature of a person's motivation and intent. Many of those who die by suicide had contact with health services in the preceding year but identifying those most at risk remains a challenge.

Objective

To explore the feasibility of using artificial neural networks with routinely collected electronic health records to support the identification of those at high risk of suicide when in contact with health services.

Methods

Using the Secure Anonymised Information Linkage Databank UK, we extracted the data of those who died by suicide between 2001 and 2015 and paired controls. Looking at primary (general practice) and secondary (hospital admissions) electronic health records, we built a binary feature vector coding the presence of risk factors at different times prior to death. Risk factors included: general practice contact and hospital admission; diagnosis of mental health issues; injury and poisoning; substance misuse; maltreatment; sleep disorders; and the prescription of opiates and psychotropics. Basic artificial neural networks were trained to differentiate between the suicide cases and paired controls. We interpreted the output score as the estimated suicide risk. System performance was assessed with 10x10-fold repeated cross-validation, and its behavior was studied by representing the distribution of estimated risk across the cases and controls, and the distribution of factors across estimated risks.

Results

We extracted a total of 2604 suicide cases and 20 paired controls per case. Our best system attained a mean error rate of 26.78% (SD 1.46; 64.57% of sensitivity and 81.86% of specificity). While the distribution of controls was concentrated around estimated risks < 0.5, cases were almost uniformly distributed between 0 and 1. Prescription of psychotropics, depression and anxiety, and self-harm increased the estimated risk by ~0.4. At least 95% of those presenting these factors were identified as suicide cases.

Conclusions

Despite the simplicity of the implemented system, the proposed methodology obtained an accuracy like other published methods based on specialized questionnaire generated data. Most of the errors came from the heterogeneity of patterns shown by suicide cases, some of which were identical to those of the paired controls. Prescription of psychotropics, depression and anxiety, and self-harm were strongly linked with higher estimated risk scores, followed by hospital admission and long-term drug and alcohol misuse. Other risk factors like sleep disorders and maltreatment had more complex effects.

SUBMITTER: DelPozo-Banos M

PROVIDER: S-EPMC6035342 | biostudies-literature | 2018 Jun

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Using Neural Networks with Routine Health Records to Identify Suicide Risk: Feasibility Study.

DelPozo-Banos Marcos M John Ann A Petkov Nicolai N Berridge Damon Mark DM Southern Kate K LLoyd Keith K Jones Caroline C Spencer Sarah S Travieso Carlos Manuel CM

JMIR mental health 20180622 2

<h4>Background</h4>Each year, approximately 800,000 people die by suicide worldwide, accounting for 1-2 in every 100 deaths. It is always a tragic event with a huge impact on family, friends, the community and health professionals. Unfortunately, suicide prevention and the development of risk assessment tools have been hindered by the complexity of the underlying mechanisms and the dynamic nature of a person's motivation and intent. Many of those who die by suicide had contact with health servic ...[more]

PMID: 29934287

Similar Datasets

Project description:OBJECTIVE:The authors sought to develop and validate models using electronic health records to predict suicide attempt and suicide death following an outpatient visit. METHOD:Across seven health systems, 2,960,929 patients age 13 or older (mean age, 46 years; 62% female) made 10,275,853 specialty mental health visits and 9,685,206 primary care visits with mental health diagnoses between Jan. 1, 2009, and June 30, 2015. Health system records and state death certificate data identified suicide attempts (N=24,133) and suicide deaths (N=1,240) over 90 days following each visit. Potential predictors included 313 demographic and clinical characteristics extracted from records for up to 5 years before each visit: prior suicide attempts, mental health and substance use diagnoses, medical diagnoses, psychiatric medications dispensed, inpatient or emergency department care, and routinely administered depression questionnaires. Logistic regression models predicting suicide attempt and death were developed using penalized LASSO (least absolute shrinkage and selection operator) variable selection in a random sample of 65% of the visits and validated in the remaining 35%. RESULTS:Mental health specialty visits with risk scores in the top 5% accounted for 43% of subsequent suicide attempts and 48% of suicide deaths. Of patients scoring in the top 5%, 5.4% attempted suicide and 0.26% died by suicide within 90 days. C-statistics (equivalent to area under the curve) for prediction of suicide attempt and suicide death were 0.851 (95% CI=0.848, 0.853) and 0.861 (95% CI=0.848, 0.875), respectively. Primary care visits with scores in the top 5% accounted for 48% of subsequent suicide attempts and 43% of suicide deaths. C-statistics for prediction of suicide attempt and suicide death were 0.853 (95% CI=0.849, 0.857) and 0.833 (95% CI=0.813, 0.853), respectively. CONCLUSIONS:Prediction models incorporating both health record data and responses to self-report questionnaires substantially outperform existing suicide risk prediction tools.

Project description:BackgroundHospital antimicrobial stewardship (AMS) programmes are multidisciplinary initiatives to optimize antimicrobial use. Most hospitals depend on time-consuming manual audits to monitor clinicians' prescribing. But much of the information needed could be sourced from electronic health records (EHRs).ObjectivesTo develop an informatics methodology to analyse characteristics of hospital AMS practice using routine electronic prescribing and laboratory records.MethodsFeasibility study using electronic prescribing, laboratory and clinical coding records from adult patients admitted to six specialities at Queen Elizabeth Hospital, Birmingham, UK (September 2017-August 2018). The study involved: (i) a review of AMS standards of care; (ii) their translation into concepts measurable from commonly available EHRs; and (iii) a pilot application in an EHR cohort study (n = 61679 admissions).ResultsWe developed data modelling methods to characterize antimicrobial use (antimicrobial therapy episode linkage methods, therapy table, therapy changes). Prescriptions were linked into antimicrobial therapy episodes (mean 2.4 prescriptions/episode; mean length of therapy 5.8 days), enabling several actionable findings. For example, 22% of therapy episodes for low-severity community-acquired pneumonia were congruent with prescribing guidelines, with a tendency to use broader-spectrum antibiotics. Analysis of therapy changes revealed IV to oral therapy switching was delayed by an average 3.6 days (95% CI: 3.4-3.7). Microbial cultures were performed prior to treatment initiation in just 22% of antibacterial prescriptions. The proposed methods enabled fine-grained monitoring of AMS practice down to specialities, wards and individual clinical teams by case mix, enabling more meaningful peer comparison.ConclusionsIt is feasible to use hospital EHRs to construct rapid, meaningful measures of prescribing quality with potential to support quality improvement interventions (audit/feedback to prescribers), engagement with front-line clinicians on optimizing prescribing, and AMS impact evaluation studies.

Project description:BackgroundMental health supported housing services are a key component in the rehabilitation of people with severe and complex needs. They are implemented widely in the UK and other deinstitutionalised countries but there have been few empirical studies of their effectiveness due to the logistic challenges and costs of standard research methods. The Clinical Record Interactive Search (CRIS) tool, developed to de-identify and interrogate routinely recorded electronic health records, may provide an alternative to evaluate supported housing services.MethodsThe feasibility of using the Camden and Islington NHS Foundation Trust CRIS database to identify a sample of users of mental health supported accommodation services. Two approaches to data interrogation and case identification were compared; using structured fields indicating individual's accommodation status, and iterative development of free text searches of clinical notes referencing supported housing. The data used were recorded over a 10-year-period (01-January-2008 to 31-December-2017).ResultsBoth approaches were carried out by one full-time researcher over four weeks (150 hours). Two structured fields indicating accommodation status were found, 2,140 individuals had a value in at least one of the fields representative of supported accommodation. The free text search of clinical notes returned 21,103 records pertaining to 1,105 individuals. A manual review of 10% of the notes indicated an estimated 733 of these individuals had used a supported housing service, a positive predictive value of 66.4%. Over two-thirds of the individuals returned in the free text search (768/1,105, 69.5%) were identified via the structured fields approach. Although the estimated positive predictive value was relatively high, a substantial proportion of the individuals appearing only in the free text search (337/1,105, 30.5%) are likely to be false positives.ConclusionsIt is feasible and requires minimal resources to use de-identified electronic health record search tools to identify large samples of users of mental health supported housing using structured and free text fields. Further work is needed to establish the availability and completion of variables relevant to specific clinical research questions in order to fully assess the utility of electronic health records in evaluating the effectiveness of these services.

Dataset Information

Using Neural Networks with Routine Health Records to Identify Suicide Risk: Feasibility Study.

Background

Objective

Methods

Results

Conclusions

Publications

Using Neural Networks with Routine Health Records to Identify Suicide Risk: Feasibility Study.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets