'Big data' in mental health research: current status and emerging possibilities

Biosensors for psychiatric biomarkers in mental health monitoring

Psychiatric disorders are associated with serve disturbances in cognition, emotional control, and/or behavior regulation, yet few routine clinical tools are available for the real-time evaluation and early-stage diagnosis of mental health. Abnormal levels of relevant biomarkers may imply biological, neurological, and developmental dysfunctions of psychiatric patients. Exploring biosensors that can provide rapid, in-situ, and real-time monitoring of psychiatric biomarkers is therefore vital for prevention, diagnosis, treatment, and prognosis of mental disorders. Recently, psychiatric biosensors with high sensitivity, selectivity, and reproducibility have been widely developed, which are mainly based on electrochemical and optical sensing technologies. This review presented psychiatric disorders with high morbidity, disability, and mortality, followed by describing pathophysiology in a biomarker-implying manner. The latest biosensors developed for the detection of representative psychiatric biomarkers (e.g., cortisol, dopamine, and serotonin) were comprehensively summarized and compared in their sensitivities, sensing technologies, applicable biological platforms, and integrative readouts. These well-developed biosensors are promising for facilitating the clinical utility and commercialization of point-of-care diagnostics. It is anticipated that mental healthcare could be gradually improved in multiple perspectives, ranging from innovations in psychiatric biosensors in terms of biometric elements, transducing principles, and flexible readouts, to the construction of 'Big-Data' networks utilized for sharing intractable psychiatric indicators and cases.

Borrowing strength from clinical trials in analysing longitudinal data from a treated cohort: investigating the effectiveness of acetylcholinesterase inhibitors in the management of dementia

BACKGROUND

Health care professionals seek information about effectiveness of treatments in patients who would be offered them in routine clinical practice. Electronic medical records (EMRs) and randomized controlled trials (RCTs) can both provide data on treatment effects; however, each data source has limitations when considered in isolation.

METHODS

A novel modelling methodology which incorporates RCT estimates in the analysis of EMR data via informative prior distributions is proposed. A Bayesian mixed modelling approach is used to model outcome trajectories among patients in the EMR dataset receiving the treatment of interest. This model incorporates an estimate of treatment effect based on a meta-analysis of RCTs as an informative prior distribution. This provides a combined estimate of treatment effect based on both data sources.

RESULTS

The superior performance of the novel combined estimator is demonstrated via a simulation study. The new approach is applied to estimate the effectiveness at 12 months after treatment initiation of acetylcholinesterase inhibitors in the management of the cognitive symptoms of dementia in terms of Mini-Mental State Examination scores. This demonstrated that estimates based on either trials data only (1.10, SE = 0.316) or cohort data only (1.56, SE = 0.240) overestimated this compared with the estimate using data from both sources (0.86, SE = 0.327).

CONCLUSIONS

It is possible to combine data from EMRs and RCTs in order to provide better estimates of treatment effectiveness.

General and age-specific fertility rates in non-affective psychosis: population-based analysis of Scottish women

PURPOSE

Women diagnosed with non-affective psychosis have a lower general fertility rate (GFR) and age-specific fertility rate (ASFR) than women in the general population. Contemporary data on GFR in this group remain limited, despite substantive changes in prescribing and management. We calculated contemporary estimates of the GFR and ASFR for women diagnosed with non-affective psychosis compared with the general population of women without this diagnosis.

METHODS

A population-based design combined routinely collected historical maternity and psychiatric data from two representative areas of Scotland. Women were included from the NHS Grampian or Greater Glasgow and Clyde areas and were aged 15-44 between 2005 and 2013 inclusive. The 'exposed' group had a diagnosis of non-affective psychosis (ICD-10 F20-F29) and was compared to the general population of 'unexposed' women in the same geographical areas.

RESULTS

Annual GFR between 2005 and 2013 for women with non-affective psychosis varied from 9.6 to 21.3 live births/1000 women per year in the exposed cohort and 52.7 to 57.8 live births/1000 women per year in the unexposed cohort, a rate ratio (RR) of 0.28 [p < 0.001; 95% CI (0.24, 0.32)]. ASFR for all 5-year age groups was lower in the exposed cohort than amongst unexposed women.

CONCLUSION

We highlight continued low fertility rates in women with a diagnosis of non-affective psychosis, despite widespread availability of prolactin-sparing atypical antipsychotics. Accurate estimation of fertility rates remains crucial in developing needs-matched perinatal care for these women. Methodological improvements using routine datasets to investigate perinatal mental health are also urgently needed.

Impact of pet dog or cat exposure during childhood on mental illness during adolescence: a cohort study

BACKGROUND

In our prior study of 643 children, ages 4-11 years, children with pet dogs had lower anxiety scores than children without pet dogs. This follow-up study examines whether exposure to pet dogs or cats during childhood reduces the risk of adolescent mental health (MH) disorders.

METHODS

Using a retrospective cohort study design, we merged our prior study database with electronic medical record (EMR) data to create an analytic database. Common MH diagnoses (anxiety, depression, ADHD) occurring from the time of prior study enrollment to 10/27/21 were identified using ICD-9 and ICD-10 codes. We used proportional hazards regression to compare time to MH diagnoses, between youths with and without pets. From 4/1/20 to 10/27/21, parents and youth in the prior study were interviewed about the amount of time the youth was exposed to a pet and how attached s/he was to the pet. Exposure included having a pet dog at baseline, cumulative exposure to a pet dog or cat during follow-up, and level of pet attachment. The main outcomes were anxiety diagnosis, any MH diagnosis, and MH diagnosis associated with a psychotropic prescription.

RESULTS

EMR review identified 571 youths with mean age of 14 years (range 11-19), 53% were male, 58% had a pet dog at baseline. During follow-up (mean of 7.8 years), 191 children received a MH diagnosis: 99 were diagnosed with anxiety (52%), 61 with ADHD (32%), 21 with depression (11%), 10 with combined MH diagnoses (5%). After adjusting for significant confounders, having a pet dog at baseline was associated with lower risk of any MH diagnosis (HR = 0.74, p = .04) but not for anxiety or MH diagnosis with a psychotropic prescription. Among the 241 (42%) youths contacted for follow-up, parent-reported cumulative exposure to pet dogs was borderline negatively associated with occurrence of any MH diagnosis (HR = 0.74, p = .06). Cumulative exposure to the most attached pet (dog or cat) was negatively associated with anxiety diagnosis (HR = 0.57, p = .006) and any MH diagnosis (HR = 0.64, p = .013).

CONCLUSION

Cumulative exposure to a highly attached pet dog or cat is associated with reduced risk of adolescent MH disorders.

Digital Transformation in Healthcare 4.0: Critical Factors for Business Intelligence Systems

The health sector is one of the most knowledge-intensive and complicated globally. It has been proven repeatedly that Business Intelligence (BI) systems in the healthcare industry can help hospitals make better decisions. Some studies have looked at the usage of BI in health, but there is still a lack of information on how to develop a BI system successfully. There is a significant research gap in the health sector because these studies do not concentrate on the organizational determinants that impact the development and acceptance of BI systems in different organizations; therefore, the aim of this article is to develop a framework for successful BI system development in the health sector taking into consideration the organizational determinants of BI systems’ acceptance, implementation, and evaluation. The proposed framework classifies the determinants under organizational, process, and strategic aspects as different types to ensure the success of BI system deployment. Concerning practical implications, this paper gives a roadmap for a wide range of healthcare practitioners to ensure the success of BI system development.

Big Data in der Gesundheitsförderung und Prävention

Hintergrund

Die Nutzung großer und vielfältiger Datenmengen (Big Data) kann zur Gewinnung gesundheitsbezogener Erkenntnisse führen. Die Relevanz untermauern aktuelle Erfordernisse, bspw. in Zusammenhang mit der Digitalisierung, der Gesundheitsversorgung in Ausnahmesituationen und der zunehmenden Bedeutung von Personalisierungsprozessen in der Gesundheitsforschung. Das Potenzial von Big Data zur Erforschung vulnerabler Gruppen ist strittig, jedoch vor dem Hintergrund relativ stabiler sozialbedingter gesundheitlicher Ungleichheit besonders relevant.

Ziel der Arbeit

In der Studie wird untersucht, wie Expert*innen im Bereich der Analyse von Gesundheitsdaten das Potenzial von Big Data in der Gesundheitsförderung und Prävention, insbesondere zur Erforschung vulnerabler Gruppen, einschätzen.

Material und Methode

In einer Delphi-Studie wurden Expert*innen in zwei Runden mit einem Onlinefragebogen befragt, um Konsens und Dissens über das Potenzial von Big Data zu identifizieren.

Ergebnisse und Schlussfolgerung

Aus Sicht der Expert*innen birgt Big Data ein Potenzial für die Gesundheitsförderung und Prävention, insbesondere im klinischen Setting und durch die Personalisierung gesundheitsbezogener Maßnahmen. Vor allem Menschen mit seltenen Erkrankungen und ältere Personen könnten durch Big-Data-Analysen profitieren, bspw. durch beschleunigte Diagnoseprozesse oder personalisierte digitale Gesundheitsanwendungen. Uneinig sind sich die Expert*innen über den Umfang, in welchem es Forschungseinrichtungen, Krankenversicherungen oder Unternehmen, erlaubt sein soll, derartige Daten zu nutzen oder zu teilen.

NeuroBlu, an electronic health record (EHR) trusted research environment (TRE) to support mental healthcare analytics with real-world data

PURPOSE

NeuroBlu is a real-world data (RWD) repository that contains deidentified electronic health record (EHR) data from US mental healthcare providers operating the MindLinc EHR system. NeuroBlu enables users to perform statistical analysis through a secure web-based interface. Structured data are available for sociodemographic characteristics, mental health service contacts, hospital admissions, International Classification of Diseases ICD-9/ICD-10 diagnosis, prescribed medications, family history of mental disorders, Clinical Global Impression-Severity and Improvement (CGI-S/CGI-I) and Global Assessment of Functioning (GAF). To further enhance the data set, natural language processing (NLP) tools have been applied to obtain mental state examination (MSE) and social/environmental data. This paper describes the development and implementation of NeuroBlu, the procedures to safeguard data integrity and security and how the data set supports the generation of real-world evidence (RWE) in mental health.

PARTICIPANTS

As of 31 July 2021, 562 940 individuals (48.9% men) were present in the data set with a mean age of 33.4 years (SD: 18.4 years). The most frequently recorded diagnoses were substance use disorders (1 52 790 patients), major depressive disorder (1 29 120 patients) and anxiety disorders (1 03 923 patients). The median duration of follow-up was 7 months (IQR: 1.3 to 24.4 months).

FINDINGS TO DATE

The data set has supported epidemiological studies demonstrating increased risk of psychiatric hospitalisation and reduced antidepressant treatment effectiveness among people with comorbid substance use disorders. It has also been used to develop data visualisation tools to support clinical decision-making, evaluate comparative effectiveness of medications, derive models to predict treatment response and develop NLP applications to obtain clinical information from unstructured EHR data.

FUTURE PLANS

The NeuroBlu data set will be further analysed to better understand factors related to poor clinical outcome, treatment responsiveness and the development of predictive analytic tools that may be incorporated into the source EHR system to support real-time clinical decision-making in the delivery of mental healthcare services.

What is the Current and Future Status of Digital Mental Health Interventions?

The prevalence of mental disorders continues to increase, especially with the advent of the COVID-19 pandemic. Although we have evidence-based psychological treatments to address these conditions, most people encounter some barriers to receiving this help (e.g., stigma, geographical or time limitations). Digital mental health interventions (e.g., Internet-based interventions, smartphone apps, mixed realities -virtual and augmented reality) provide an opportunity to improve accessibility to these treatments. This article summarizes the main contributions of the different types of digital mental health solutions. It analyzes their limitations (e.g., drop-out rates, lack of engagement, lack of personalization, lack of cultural adaptations) and showcases the latest sophisticated and innovative technological advances under the umbrella of precision medicine (e.g., digital phenotyping, chatbots, or conversational agents). Finally, future challenges related to the need for real world implementation of these interventions, the use of predictive methodology, and hybrid models of care in clinical practice, among others, are discussed.

Rising Mental Health Incidence Among Adolescents in Westchester, NY

CONTEXT

Many governments have publicly released healthcare data, which can be mined for insights about disease conditions, and their impact on society.

METHODS

We present a big-data analytics approach to investigate data in the New York Statewide Planning and Research Cooperative System (SPARCS) consisting of 20 million patient records.

FINDINGS

Whereas the age group 30-48 years exhibited an 18% decline in mental health (MH) disorders from 2009 to 2016, the age group 0-17 years showed a 5.4% increase. MH issues amongst the age group 0-17 years comprise a significant expenditure in New York State. Within this age group, we find a higher prevalence of MH disorders in females and minority populations. Westchester County has seen a 32% increase in incidences and a 41% increase in costs.

CONCLUSIONS

Our approach is scalable to data from multiple government agencies and provides an independent perspective on health care issues, which can prove valuable to policy and decision-makers.

iPSYcare: the development of a linked electronic medical records database to study and optimize psychiatric care in Antwerp

Objective

The study of care trajectories of psychiatric patients across hospitals was previously not possible in Belgium as each hospital stores its data autonomously, and government-related registrations do not contain a unique identifier or are incomplete. A new longitudinal database called iPSYcare (Improved Psychiatric Care and Research) was therefore constructed in 2021, and links the electronic medical records of patients in psychiatric units of eight hospitals in the Antwerp Province, Belgium. The database provides a wide range of information on patients, care trajectories and delivered care in the region. In a first phase, the database will only contain information about adult patients who were admitted to a hospital or treated by an outreach team and who gave explicit consent. In the future, the database may be expanded to other regions and additional data on outpatient care may be added.

Results

IPSYcare is a close collaboration between the University of Antwerp and hospitals in the province of Antwerp. This paper describes the development of the database, how privacy and ethical issues will be handled, and how the governance of the database will be organized.

Use of Data to Understand the Social Determinants of Depression in Two Middle-Income Countries: the 3-D Commission

Depression accounts for a large share of the global disease burden, with an estimated 264 million people globally suffering from depression. Despite being one of the most common kinds of mental health (MH) disorders, much about depression remains unknown. There are limited data about depression, in terms of its occurrence, distribution, and wider social determinants. This work examined the use of novel data sources for assessing the scope and social determinants of depression, with a view to informing the reduction of the global burden of depression.This study focused on new and traditional sources of data on depression and its social determinants in two middle-income countries (LMICs), namely, Brazil and India. We identified data sources using a combination of a targeted PubMed search, Google search, expert consultations, and snowball sampling of the relevant literature published between October 2010 and September 2020. Our search focused on data sources on the following HEALTHY subset of determinants: healthcare (H), education (E), access to healthy choices (A), labor/employment (L), transportation (T), housing (H), and income (Y).Despite the emergence of a variety of data sources, their use in the study of depression and its HEALTHY determinants in India and Brazil are still limited. Survey-based data are still the most widely used source. In instances where new data sources are used, the most commonly used data sources include social media (twitter data in particular), geographic information systems/global positioning systems (GIS/GPS), mobile phone, and satellite imagery. Often, the new data sources are used in conjunction with traditional sources of data. In Brazil, the limited use of new data sources to study depression and its HEALTHY determinants may be linked to (a) the government's outsized role in coordinating healthcare delivery and controlling the data system, thus limiting innovation that may be expected from the private sector; (b) the government routinely collecting data on depression and other MH disorders (and therefore, does not see the need for other data sources); and (c) insufficient prioritization of MH as a whole. In India, the limited use of new data sources to study depression and its HEALTHY determinants could be a function of (a) the lack of appropriate regulation and incentives to encourage data sharing by and within the private sector, (b) absence of purposeful data collection at subnational levels, and (c) inadequate prioritization of MH. There is a continuing gap in the collection and analysis of data on depression, possibly reflecting the limited priority accorded to mental health as a whole. The relatively limited use of data to inform our understanding of the HEALTHY determinants of depression suggests a substantial need for support of independent research using new data sources. Finally, there is a need to revisit the universal health coverage (UHC) frameworks, as these frameworks currently do not include depression and other mental health-related indicators so as to enable tracking of progress (or lack thereof) on such indicators.

The Safe pilot study: A prospective naturalistic study with repeated measures design to test the psychosis - violence link in and after discharge from forensic facilities

The research evidence is very strong for high recidivism rates of violence after discharge from forensic facilities. Big data research has found that a substantial proportion of the forensic population with relapse into violence has a psychosis diagnosis and a criminal record. However, more research on the association between psychotic symptoms and violence may inform and enhance risk assessment, prevention, and treatment. We conducted a prospective naturalistic study with a repeated measures design in a sample of 22 psychotic patients during follow-up after discharge from forensic mental health facilities. We had three aims: to test the predictive validity of three psychotic symptom scales for violence, to analyze main and interaction effects between psychotic symptoms and previous criminal conviction, and to explore the feasibility and potential benefit of the repeated measures design for prospective follow-up research. Interpreted within the limitation of the small sample size, the results were promising for all scales, particularly for adjusted effects without interaction. Two scales remained significant when their interaction with criminal conviction was adjusted. This indicates that risk judgment of psychotic patients with criminal conviction can be improved by adding measurement of fluctuations in psychotic symptoms. The repeated measures design was instrumental in this research.

Schizophrenia Related Hospitalizations - a Big Data Analysis of a National Hospitalization Database

Schizophrenia is a mental disorder characterized by long hospitalizations and frequent need for chronic/acute psychiatric care. Hospitalizations represent a valuable quality of care indicator in schizophrenia patients. The aim of this study was to describe a nationwide perspective of schizophrenia related hospitalizations. We performed a retrospective observational study using a nationwide hospitalization database containing all hospitalizations registered in Portuguese public hospitals from 2008 to 2015. Hospitalizations with a primary diagnosis of schizophrenia were selected based on the definition by CCS - Clinical Classification Software diagnostic single-level 659. Schizophrenia subtypes were identified based on International Classification of Diseases version 9, Clinical Modification (ICD-9-CM) codes of diagnosis 295.xx. A total of 25,385 hospitalizations were registered belonging to 14,279 patients. 68.0% of the hospitalizations occurred in male patients and the median length of stay was 18.0 days. In male patients' hospitalizations, the most frequent age group was 31-50 years followed by the age group of 18-30 years (55.9 and 24.0% respectively). For female patients, the most frequent age group was 31-50 years followed by 51-70 years (54.1 and 22.6%, respectively). There were 73 hospitalization with a deadly outcome (0.29%). Paranoid type was the most frequent subtype of schizophrenia (50.5%). The mean hospitalization charges were 3509.7€ per episode, with a total charge of 89.1 M€ in the 8-year period. This is a nationwide study using Big Data analysis giving a broad perspective of schizophrenia hospitalization panorama at a nationwide level. We found differences in hospitalization characteristics according to patients' gender, age and primary diagnosis.

Attitudes towards Risk Prediction in a Help Seeking Population of Early Detection Centers for Mental Disorders-A Qualitative Approach

Big Data approaches raise hope for a paradigm shift towards illness prevention, while others are concerned about discrimination resulting from these approaches. This will become particularly important for people with mental disorders, as research on medical risk profiles and early detection progresses rapidly. This study aimed to explore views and attitudes towards risk prediction in people who, for the first time, sought help at one of three early detection centers for mental disorders in Germany (Cologne, Munich, Dresden). A total of 269 help-seekers answered an open-ended question on the potential use of risk prediction. Attitudes towards risk prediction and motives for its approval or rejection were categorized inductively and analyzed using qualitative content analysis. The anticipated impact on self-determination was a driving decision component, regardless of whether a person would decide for or against risk prediction. Results revealed diverse, sometimes contrasting, motives for both approval and rejection (e.g., the desire to control of one’s life as a reason for and against risk prediction). Knowledge about a higher risk as a potential psychological burden was one of the major reasons against risk prediction. The decision to make use of risk prediction is expected to have far-reaching effects on the quality of life and self-perception of potential users. Healthcare providers should empower those seeking help by carefully considering individual expectations and perceptions of risk prediction.

Creating a Multisite Perinatal Psychiatry Databank: Purpose and Development

Mental health issues during the perinatal period are common; up to 29% of pregnant and 15% of postpartum women meet psychiatric diagnostic criteria. Despite its ubiquity, little is known about the longitudinal trajectories of perinatal psychiatric illness. This paper describes a collaboration among six perinatal mental health services in Quebec, Canada, to create an electronic databank that captures longitudinal patient data over the course of the perinatal period. The collaborating sites met to identify research interests and to select a standardized set of variables to be collected during clinical appointments. Procedures were implemented for creating a databank that serves both research and clinical purposes. The resulting databank allows pregnant and postpartum patients to complete self-report questionnaires on medical and psychosocial variables during their intake appointment in conjunction with their clinicians who fill in relevant medical information. All participants are followed until 6 months postpartum. The databank represents an opportunity to examine illness trajectories and to study rare mental disorders and the relationship between biological and psychosocial variables.

Epidemiology of depressive disorders in people living with hypertension in Africa: a systematic review and meta-analysis

OBJECTIVES

Better knowledge of epidemiology of depressive disorders in people living with hypertension can help to implement pertinent strategies to address its burden. The objective was to estimate the prevalence of depressive disorders and symptoms in people living with hypertension in Africa.

DESIGN

Systematic review and meta-analysis.

DATA SOURCES

PubMed, EMBASE, African Index Medicus, African Journals OnLine were searched up to 31 January 2020, regardless of the language of publication.

ELIGIBILITY CRITERIA

We included studies conducted among adult patients with hypertension (≥18 years) living in Africa and reporting the prevalence of depressive disorders and symptoms.

DATA EXTRACTION AND SYNTHESIS

Two independent investigators selected studies, extracted data and assessed the methodological quality of included studies by using the tools developed by Joanna Briggs Institute. Multivariate random-effects meta-analysis served to pool data by considering the variability between diagnostic tools used to identify patients with depressive disorders or symptoms.

RESULTS

We included 11 studies with 5299 adults with hypertension. Data were collected between 2002 and 2017, from South Africa, Nigeria, Ghana, Ethiopia and Burkina Faso. The mean age varied between 50.3 years and 59.6 years. The proportion of men ranged from 28% to 54%. The adjusted prevalence of depressive disorders taking into account the variance between diagnostic tools was 17.9% (95% CI 13.0% to 23.4%). The prevalence of depressive symptoms and major depressive symptoms was 33.3% (95% CI 9.9% to 61.6%) and 7.8% (95% CI 3.0% to 14.5%), respectively. There was heterogeneity attributable to the diagnostic tools for depressive disorders and symptoms. There was no publication bias.

CONCLUSION

Notwithstanding the representativeness lack of some (sub) regions of Africa, weakening the generalisability of findings to the entire region; depressive disorders and symptoms are prevalent in people living with hypertension in Africa, indicating that strategies from clinicians, researchers and public health makers are needed to reduce its burden in the region.

Big data in severe mental illness: the role of electronic monitoring tools and metabolomics

There is an increasing interest in the development of effective early detection and intervention strategies in severe mental illness (SMI). Ideally, these efforts should lead to the delineation of accurate staging models of SMI enabling personalized interventions. It is plausible that big data approaches will be instrumental in describing the developmental trajectories of SMI by facilitating the incorporation of data from multiple sources, including those pertaining to the biological make-up of affected subjects. In this review, we first aimed to offer a perspective on how big data are helping the delineation of personalized approaches in SMI, and, second, to offer a quantitative synthesis of big data approaches in metabolomics of SMI. We finally described future directions of this research area.

Community-care unit model of residential mental health rehabilitation services in Queensland, Australia: predicting outcomes of consumers 1-year post discharge

AIMS

Community care units (CCUs) are a model of residential psychiatric rehabilitation aiming to improve the independence and community functioning of people with severe and persistent mental illness. This study examined factors predicting improvement in outcomes among CCU consumers.

METHODS

Hierarchical regression using data from a retrospective cohort (N = 501) of all consumers admitted to five CCUs in Queensland, Australia between 2005 and 2014. The primary outcome was changed in mental health and social functioning (Health of the Nation Outcome Scale). Secondary outcomes were disability (Life Skills Profile-16), service use, accommodation instability, and involuntary treatment. Potential predictors covered service, consumer, and treatment characteristics. Group-level and individualised change were assessed between the year pre-admission and post-discharge. Where relevant and available, the reliable and clinically significant (RCS) change was assessed by comparison with a normative sample.

RESULTS

Group-level analyses showed statistically significant improvements in mental health and social functioning, and reductions in psychiatry-related bed-days, emergency department (ED) presentations and involuntary treatment. There were no significant changes in disability or accommodation instability. A total of 54.7% of consumers demonstrated reliable improvement in mental health and social functioning, and 43.0% showed RCS improvement. The majority (60.6%) showed a reliable improvement in psychiatry-related bed-use; a minority demonstrated reliable improvement in ED presentations (12.5%). Significant predictors of improvement included variables related to the CCU care (e.g. episode duration), consumer characteristics (e.g. primary diagnosis) and treatment variables (e.g. psychiatry-related bed-days pre-admission). Higher baseline impairment in mental health and social functioning (β = 1.12) and longer episodes of CCU care (β = 1.03) increased the likelihood of RCS improvement in mental health and social functioning.

CONCLUSIONS

CCU care was followed by reliable improvements in relevant outcomes for many consumers. Consumers with poorer mental health and social functioning, and a longer episode of CCU care were more likely to make RCS improvements in mental health and social functioning.

Case identification of mental health and related problems in children and young people using the New Zealand Integrated Data Infrastructure

Background

In a novel endeavour we aimed to develop a clinically relevant case identification method for use in research about the mental health of children and young people in New Zealand using the Integrated Data Infrastructure (IDI). The IDI is a linked individual-level database containing New Zealand government and survey microdata.

Methods

We drew on diagnostic and pharmaceutical information contained within five secondary care service use and medication dispensing datasets to identify probable cases of mental health and related problems. A systematic classification and refinement of codes, including restrictions by age, was undertaken to assign cases into 13 different mental health problem categories. This process was carried out by a panel of eight specialists covering a diverse range of mental health disciplines (a clinical psychologist, four child and adolescent psychiatrists and three academic researchers in child and adolescent mental health). The case identification method was applied to the New Zealand youth estimated resident population for the 2014/15 fiscal year.

Results

Over 82,000 unique individuals aged 0–24 with at least one specified mental health or related problem were identified using the case identification method for the 2014/15 fiscal year. The most prevalent mental health problem subgroups were emotional problems (31,266 individuals), substance problems (16,314), and disruptive behaviours (13,758). Overall, the pharmaceutical collection was the largest source of case identification data (59,862).

Conclusion

This study demonstrates the value of utilising IDI data for mental health research. Although the method is yet to be fully validated, it moves beyond incidence rates based on single data sources, and provides directions for future use, including further linkage of data to the IDI.

Digital Innovations for Global Mental Health: Opportunities for Data Science, Task Sharing, and Early Intervention

PURPOSE

Globally, individuals living with mental disorders are more likely to have access to a mobile phone than mental health care. In this commentary, we highlight opportunities for expanding access to and use of digital technologies to advance research and intervention in mental health, with emphasis on the potential impact in lower resource settings.

RECENT FINDINGS

Drawing from empirical evidence, largely from higher income settings, we considered three emerging areas where digital technology will potentially play a prominent role: supporting methods in data science to further our understanding of mental health and inform interventions, task sharing for building workforce capacity by training and supervising non-specialist health workers, and facilitating new opportunities for early intervention for young people in lower resource settings. Challenges were identified related to inequities in access, threats of bias in big data analyses, risks to users, and need for user involvement to support engagement and sustained use of digital interventions.

SUMMARY

For digital technology to achieve its potential to transform the ways we detect, treat, and prevent mental disorders, there is a clear need for continued research involving multiple stakeholders, and rigorous studies showing that these technologies can successfully drive measurable improvements in mental health outcomes.

Case-finding for common mental disorders in primary care using routinely collected data: a systematic review

PURPOSE

Case-finding for common mental disorders (CMD) in routine data unobtrusively identifies patients for mental health research. There is absence of a review of studies examining CMD-case-finding accuracy in routine primary care data. CMD-case definitions include diagnostic/prescription codes, signs/symptoms, and free text within electronic health records. This systematic review assesses evidence for case-finding accuracy of CMD-case definitions compared to reference standards.

METHODS

PRISMA-DTA checklist guided review. Eligibility criteria were outlined prior to study search; studies compared CMD-case definitions in routine primary care data to diagnostic interviews, screening instruments, or clinician judgement. Studies were quality assessed using QUADAS-2.

RESULTS

Fourteen studies were included, and most were at high risk of bias. Nine studies examined depressive disorders and seven utilised diagnostic interviews as reference standards. Receiver operating characteristic (ROC) planes illustrated overall variable case-finding accuracy across case definitions, quantified by Youden's index. Forest plots demonstrated most case definitions provide high specificity.

CONCLUSION

Case definitions effectively identify cases in a population with good accuracy and few false positives. For 100 anxiety cases, identified using diagnostic codes, between 12 and 20 will be false positives; 0-47 cases will be missed. Sensitivity is more variable and specificity is higher in depressive cases; for 100 cases identified using diagnostic codes, between 0 and 87 will be false positives; 4-18 cases will be missed. Incorporating context to case definitions may improve overall case-finding accuracy. Further research is required for meta-analysis and robust conclusions.

Indicators of mental disorders in UK Biobank-A comparison of approaches

OBJECTIVES

For many research cohorts, it is not practical to provide a "gold-standard" mental health diagnosis. It is therefore important for mental health research that potential alternative measures for ascertaining mental disorder status are understood.

METHODS

Data from UK Biobank in those participants who had completed the online Mental Health Questionnaire (n = 157,363) were used to compare the classification of mental disorder by four methods: symptom-based outcome (self-complete based on diagnostic interviews), self-reported diagnosis, hospital data linkage, and self-report medication.

RESULTS

Participants self-reporting any psychiatric diagnosis had elevated risk of any symptom-based outcome. Cohen's κ between self-reported diagnosis and symptom-based outcome was 0.46 for depression, 0.28 for bipolar affective disorder, and 0.24 for anxiety. There were small numbers of participants uniquely identified by hospital data linkage and medication.

CONCLUSION

Our results confirm that ascertainment of mental disorder diagnosis in large cohorts such as UK Biobank is complex. There may not be one method of classification that is right for all circumstances, but an informed and transparent use of outcome measure(s) to suit each research question will maximise the potential of UK Biobank and other resources for mental health research.

Narrative Case Notes Have the Potential to Predict Seclusion 3 Days in Advance: A Mixed-Method Analysis

Objectives: Current risk assessment tools can predict problematic behavior and the need for coercive measures, but only with a moderate level of accuracy. The aim of this study was to assess antecedents and triggers of seclusion. Methods: Narrative notes of health care professionals on psychiatric inpatients were analyzed daily starting 3 days prior to seclusion in the case group (n = 26) and compared to a matched control group without seclusion (n = 26) by use of quantitative and qualitative research methods, based on qualitative content analysis. Results: Quantitative measures showed more aggression in the case group with highly significant differences between the groups (p < 0.001) at all measurement times. Seclusion was significantly associated with the total word count of the narrative notes. Subjective emotional expressions by staff were more apparent before seclusion (p = 0.003). Most frequently, subjective expressions regarding "arduous/provocative" (p < 0.001) and "anxious" (p = 0.010) sentiments could be identified in the case group. Description of patients' behavior in the case group included more negatively assessed terms (p = 0.001). Moreover, sleep loss, refusing medication, high contact frequency, demanding behavior and denied requests were present in a significantly higher frequency before seclusion. Expressions like "threatening" (p = 0.001) were found only before seclusion and appeared to have the function of personal risk assessment. The expression "manageable" (p = 0.035) appeared often in difficult situations that could still be handled. Conclusion: Several factors preceding seclusion could be identified. Narrative notes of staff already showed differences 3 days before the escalation. Particularly the word count, the analysis of terms describing patients' behavior, subjective expressions of staff, and terms used as a function of personal risk assessment could help to provide better predictions of aggressive incidents and to prevent coercive measures.

Risk Assessment Tools and Data-Driven Approaches for Predicting and Preventing Suicidal Behavior

Risk assessment of suicidal behavior is a time-consuming but notoriously inaccurate activity for mental health services globally. In the last 50 years a large number of tools have been designed for suicide risk assessment, and tested in a wide variety of populations, but studies show that these tools suffer from low positive predictive values. More recently, advances in research fields such as machine learning and natural language processing applied on large datasets have shown promising results for health care, and may enable an important shift in advancing precision medicine. In this conceptual review, we discuss established risk assessment tools and examples of novel data-driven approaches that have been used for identification of suicidal behavior and risk. We provide a perspective on the strengths and weaknesses of these applications to mental health-related data, and suggest research directions to enable improvement in clinical practice.

Negative Symptoms in Early-Onset Psychosis and Their Association With Antipsychotic Treatment Failure

The prevalence of negative symptoms (NS) at first episode of early-onset psychosis (EOP), and their effect on psychosis prognosis is unclear. In a sample of 638 children with EOP (aged 10-17 y, 51% male), we assessed (1) the prevalence of NS at first presentation to mental health services and (2) whether NS predicted eventual development of multiple treatment failure (MTF) prior to the age of 18 (defined by initiation of a third trial of novel antipsychotic due to prior insufficient response, intolerable adverse-effects or non-adherence). Data were extracted from the electronic health records held by child inpatient and community-based services in South London, United Kingdom. Natural Language Processing tools were used to measure the presence of Marder Factor NS and antipsychotic use. The association between presenting with ≥2 NS and the development of MTF over a 5-year period was modeled using Cox regression. Out of the 638 children, 37.5% showed ≥2 NS at first presentation, and 124 (19.3%) developed MTF prior to the age of 18. The presence of NS at first episode was significantly associated with MTF (adjusted hazard ratio 1.62, 95% CI 1.07-2.46; P = .02) after controlling for a number of potential confounders including psychosis diagnostic classification, positive symptoms, comorbid depression, and family history of psychosis. Other factors associated with MTF included comorbid autism spectrum disorder, older age at first presentation, Black ethnicity, and family history of psychosis. In EOP, NS at first episode are prevalent and may help identify a subset of children at higher risk of responding poorly to antipsychotics.

Psychiatric stressor recognition from clinical notes to reveal association with suicide

Suicide takes the lives of nearly a million people each year and it is a tremendous economic burden globally. One important type of suicide risk factor is psychiatric stress. Prior studies mainly use survey data to investigate the association between suicide and stressors. Very few studies have investigated stressor data in electronic health records, mostly due to the data being recorded in narrative text. This study takes the initiative to automatically extract and classify psychiatric stressors from clinical text using natural language processing-based methods. Suicidal behaviors were also identified by keywords. Then, a statistical association analysis between suicide ideations/attempts and stressors extracted from a clinical corpus is conducted. Experimental results show that our natural language processing method could recognize stressor entities with an F-measure of 89.01 percent. Mentions of suicidal behaviors were identified with an F-measure of 97.3 percent. The top three significant stressors associated with suicide are health, pressure, and death, which are similar to previous studies. This study demonstrates the feasibility of using natural language processing approaches to unlock information from psychiatric notes in electronic health record, to facilitate large-scale studies about associations between suicide and psychiatric stressors.

How has big data contributed to obesity research? A review of the literature

There has been growing interest in the potential of ‘big data’ to enhance our understanding in medicine and public health. Although there is no agreed definition of big data, accepted critical components include greater volume, complexity, coverage and speed of availability. Much of these data are ‘found’ (as opposed to ‘made’), in that they have been collected for non-research purposes, but could include valuable information for research. The aim of this paper is to review the contribution of ‘found’ data to obesity research to date, and describe the benefits and challenges encountered. A narrative review was conducted to identify and collate peer-reviewed research studies. Database searches conducted up to September 2017 found original studies using a variety of data types and sources. These included: retail sales, transport, geospatial, commercial weight management data, social media, and smartphones and wearable technologies. The narrative review highlights the variety of data uses in the literature: describing the built environment, exploring social networks, estimating nutrient purchases or assessing the impact of interventions. The examples demonstrate four significant ways in which ‘found’ data can complement conventional ‘made’ data: firstly, in moving beyond constraints in scope (coverage, size and temporality); secondly, in providing objective, quantitative measures; thirdly, in reaching hard-to-access population groups; and lastly in the potential for evaluating real-world interventions. Alongside these opportunities, ‘found’ data come with distinct challenges, such as: ethical and legal questions around access and ownership; commercial sensitivities; costs; lack of control over data acquisition; validity; representativeness; finding appropriate comparators; and complexities of data processing, management and linkage. Despite widespread recognition of the opportunities, the impact of ‘found’ data on academic obesity research has been limited. The merit of such data lies not in their novelty, but in the benefits they could add over and above, or in combination with, conventionally collected data.

Big data handling mechanisms in the healthcare applications: A comprehensive and systematic literature review

Healthcare provides many services such as diagnosing, treatment, prevention of diseases, illnesses, injuries, and other physical and mental disorders. Large-scale distributed data processing applications in healthcare as a basic concept operates on large amounts of data. Therefore, big data application functions are the main part of healthcare operations, but there was not any comprehensive and systematic survey about studying and evaluating the important techniques in this field. Therefore, this paper aims at providing the comprehensive, detailed, and systematic study of the state-of-the-art mechanisms in the big data related to healthcare applications in five categories, including machine learning, cloud-based, heuristic-based, agent-based, and hybrid mechanisms. Also, this paper displayed a systematic literature review (SLR) of the big data applications in the healthcare literature up to the end of 2016. Initially, 205 papers were identified, but a paper selection process reduced the number of papers to 29 important studies.

Mental Ill-Health and the Epidemiology of Representations

One of major challenges facing contemporary psychiatry is the insufficient grasp of relationship between individual and collective mental pathologies. A long tradition of diagnosing "mental illness" of society-exemplified by Erich Fromm-stands apart from approach of contemporary social psychiatry and is not perceived as relevant for psychiatric discourse. In this Perspective article, I argue that it is possible to uphold the idea of a supra-individual dimension to mental health, while avoiding the obvious pitfalls involved in categorical diagnosing of society as suffering from mental illness. I argue for an extended notion of public mental ill-health, which goes beyond the quantitative understanding of mental health as an aggregate of individual diseased minds captured in statistics, and which can be conceived as a dynamic, emergent property resulting from interactions of individual brains/minds in social space. Such a notion, in turn, presents a challenge of how to account for the interfacing between individual minds/brains and the collective mental phenomena. A suitable theoretical framework is provided by the notion of epidemiology of representations, originally formulated by cognitive anthropologist Dan Sperber. Within this framework, it is possible to highlight the role of public (material) representations in inter-individual transfer of mental representations and mental states. It is a suitable conceptual platform to explain how the troubling experiences with causal or mediating role on mental health, to a significant degree arise through a person's direct interaction with material representations and participation in collective mental states, again generated by material representations.