Toward personalizing treatment for depression: predicting diagnosis and severity

Optimising the use of electronic medical records for large scale research in psychiatry

The explosion and abundance of digital data could facilitate large-scale research for psychiatry and mental health. Research using so-called "real world data"-such as electronic medical/health records-can be resource-efficient, facilitate rapid hypothesis generation and testing, complement existing evidence (e.g. from trials and evidence-synthesis) and may enable a route to translate evidence into clinically effective, outcomes-driven care for patient populations that may be under-represented. However, the interpretation and processing of real-world data sources is complex because the clinically important 'signal' is often contained in both structured and unstructured (narrative or "free-text") data. Techniques for extracting meaningful information (signal) from unstructured text exist and have advanced the re-use of routinely collected clinical data, but these techniques require cautious evaluation. In this paper, we survey the opportunities, risks and progress made in the use of electronic medical record (real-world) data for psychiatric research.

Prediction and diagnosis of depression using machine learning with electronic health records data: a systematic review

BACKGROUND

Depression is one of the most significant health conditions in personal, social, and economic impact. The aim of this review is to summarize existing literature in which machine learning methods have been used in combination with Electronic Health Records for prediction of depression.

METHODS

Systematic literature searches were conducted within arXiv, PubMed, PsycINFO, Science Direct, SCOPUS and Web of Science electronic databases. Searches were restricted to information published after 2010 (from 1st January 2011 onwards) and were updated prior to the final synthesis of data (27th January 2022).

RESULTS

Following the PRISMA process, the initial 744 studies were reduced to 19 eligible for detailed evaluation. Data extraction identified machine learning methods used, types of predictors used, the definition of depression, classification performance achieved, sample size, and benchmarks used. Area Under the Curve (AUC) values more than 0.9 were claimed, though the average was around 0.8. Regression methods proved as effective as more developed machine learning techniques.

LIMITATIONS

The categorization, definition, and identification of the numbers of predictors used within models was sometimes difficult to establish, Studies were largely Western Educated Industrialised, Rich, Democratic (WEIRD) in demography.

CONCLUSION

This review supports the potential use of machine learning techniques with Electronic Health Records for the prediction of depression. All the selected studies used clinically based, though sometimes broad, definitions of depression as their classification criteria. The reported performance of the studies was comparable to or even better than that found in primary care. There are concerns with generalizability and interpretability.

Using Machine Learning to Predict Antidepressant Treatment Outcome From Electronic Health Records

Objective

To evaluate if a machine learning approach can accurately predict antidepressant treatment outcome using electronic health records (EHRs) from patients with depression.

Method

This study examined 808 patients with depression at a New York City-based outpatient mental health clinic between June 13, 2016 and June 22, 2020. Antidepressant treatment outcome was defined based on trend in depression symptom severity over time and was categorized as either "Recovering" or "Worsening" (i.e., non-Recovering), measured by the slope of individual-level Patient Health Questionnaire-9 (PHQ-9) score trajectory spanning 6 months following treatment initiation. A patient was designated as "Recovering" if the slope is less than 0 and as "Worsening" if the slope was no less than 0. Multiple machine learning (ML) models including L2 norm regularized Logistic Regression, Naive Bayes, Random Forest, and Gradient Boosting Decision Tree (GBDT) were used to predict treatment outcome based on additional data from EHRs, including demographics and diagnoses. Shapley Additive Explanations were applied to identify the most important predictors.

Results

The GBDT achieved the best results of predicting "Recovering" (AUC: 0.7654 ± 0.0227; precision: 0.6002 ± 0.0215; recall: 0.5131 ± 0.0336). When excluding patients with low PHQ-9 scores (<10) at baseline, the results of predicting "Recovering" (AUC: 0.7254 ± 0.0218; precision: 0.5392 ± 0.0437; recall: 0.4431 ± 0.0513) were obtained. Prior diagnosis of anxiety, psychotherapy, recurrent depression, and baseline depression symptom severity were strong predictors.

Conclusions

The results demonstrate the potential utility of using ML in longitudinal EHRs to predict antidepressant treatment outcome. Our predictive tool holds the promise to accelerate personalized medical management in patients with psychiatric illnesses.

Adherence to augmentation therapy for the treatment of major depressive disorder

BACKGROUND

Inadequate response to antidepressant medication is common. Often, adjunctive pharmacotherapy or psychotherapy is recommended.

OBJECTIVE

To measure adherence to adjunctive pharmacotherapy and psychotherapy among individuals with major depressive disorder (MDD).

METHODS

Retrospective cohort study of individuals with MDD on antidepressant monotherapy who added adjunctive pharmacotherapy and/or psychotherapy. Medication adherence was measured by proportion of days covered (PDC) with optimal adherence defined as PDC≥0.80 and psychotherapy adherence defined by count of visits (optimal 8+ visits). Factors associated with optimal adherence were assessed by logistic regression.

RESULTS

Among 218,192 individuals with adjunctive therapy, 185,349 added pharmacotherapy and 32,843 added psychotherapy. In the subsequent 12 months, 36.2% and 54.9% achieved optimal adherence to adjunctive pharmacotherapy and psychotherapy, respectively. Adherence to adjunctive pharmacotherapy was associated with adding psychotherapy, index antidepressant adherence, medical comorbidities, and MDD severity codes. Adherence to adjunctive psychotherapy was associated with adding another medication, previous psychiatry visit and psychiatric comorbidities.

CONCLUSION

Adjunctive psychotherapy appears under-utilized and adherence to adjunctive therapy was low. Low adherence to adjunctive therapy reinforces challenges in managing MDD. That a second adjunctive therapy enhanced adherence to the initial adjunctive therapy indicates an opportunity to explore alternative adjunctive therapies.

Natural Language Processing and Graph Theory: Making Sense of Imaging Records in a Novel Representation Frame

BACKGROUND

A concise visualization framework of related reports would increase readability and improve patient management. To this end, temporal referrals to prior comparative exams are an essential connection to previous exams in written reports. Due to unstructured narrative texts' variable structure and content, their extraction is hampered by poor computer readability. Natural language processing (NLP) permits the extraction of structured information from unstructured texts automatically and can serve as an essential input for such a novel visualization framework.

OBJECTIVE

This study proposes and evaluates an NLP-based algorithm capable of extracting the temporal referrals in written radiology reports, applies it to all the radiology reports generated for 10 years, introduces a graphical representation of imaging reports, and investigates its benefits for clinical and research purposes.

METHODS

In this single-center, university hospital, retrospective study, we developed a convolutional neural network capable of extracting the date of referrals from imaging reports. The model's performance was assessed by calculating precision, recall, and F1-score using an independent test set of 149 reports. Next, the algorithm was applied to our department's radiology reports generated from 2011 to 2021. Finally, the reports and their metadata were represented in a modulable graph.

RESULTS

For extracting the date of referrals, the named-entity recognition (NER) model had a high precision of 0.93, a recall of 0.95, and an F1-score of 0.94. A total of 1,684,635 reports were included in the analysis. Temporal reference was mentioned in 53.3% (656,852/1,684,635), explicitly stated as not available in 21.0% (258,386/1,684,635), and omitted in 25.7% (317,059/1,684,635) of the reports. Imaging records can be visualized in a directed and modulable graph, in which the referring links represent the connecting arrows.

CONCLUSIONS

Automatically extracting the date of referrals from unstructured radiology reports using deep learning NLP algorithms is feasible. Graphs refined the selection of distinct pathology pathways, facilitated the revelation of missing comparisons, and enabled the query of specific referring exam sequences. Further work is needed to evaluate its benefits in clinics, research, and resource planning.

A Framework for Automating Psychiatric Distress Screening in Ophthalmology Clinics Using an EHR-Derived AI Algorithm

Purpose

In patients with ophthalmic disorders, psychosocial risk factors play an important role in morbidity and mortality. Proper and early psychiatric screening can result in prompt intervention and mitigate its impact. Because screening is resource intensive, we developed a framework for automating screening using an electronic health record (EHR)-derived artificial intelligence (AI) algorithm.

Methods

Subjects came from the Duke Ophthalmic Registry, a retrospective EHR database for the Duke Eye Center. Inclusion criteria included at least two encounters and a minimum of 1 year of follow-up. Presence of distress was defined at the encounter level using a computable phenotype. Risk factors included available EHR history. At each encounter, risk factors were used to discriminate psychiatric status. Model performance was evaluated using area under the receiver operating characteristic (ROC) curve and area under the precision-recall curve (PR AUC). Variable importance was presented using odds ratios (ORs).

Results

Our cohort included 358,135 encounters from 40,326 patients with an average of nine encounters per patient over 4 years. The ROC and PR AUC were 0.91 and 0.55, respectively. Of the top 25 predictors, the majority were related to existing distress, but some indicated stressful conditions, including chemotherapy (OR = 1.36), esophageal disorders (OR = 1.31), central pain syndrome (OR = 1.25), and headaches (OR = 1.24).

Conclusions

Psychiatric distress in ophthalmology patients can be monitored passively using an AI algorithm trained on existing EHR data.

Translational Relevance

When paired with an effective referral and treatment program, such algorithms may improve health outcomes in ophthalmology.

Computational Methods in Psychotherapy: A Scoping Review

BACKGROUND

The study of complex systems, such as the psychotherapeutic encounter, transcends the mechanistic and reductionist methods for describing linear processes and needs suitable approaches to describe probabilistic and scarcely predictable phenomena.

OBJECTIVE

The present study undertakes a scoping review of research on the computational methods in psychotherapy to gather new developments in this field and to better understand the phenomena occurring in psychotherapeutic interactions as well as in human interaction more generally.

DESIGN

Online databases were used to identify papers published 2011-2022, from which we selected 18 publications from different resources, selected according to criteria established in advance and described in the text. A flow chart and a summary table of the articles consulted have been created.

RESULTS

The majority of publications (44.4%) reported combined computational and experimental approaches, so we grouped the studies according to the types of computational methods used. All but one of the studies collected measured data. All the studies confirmed the usefulness of predictive and learning models in the study of complex variables such as those belonging to psychological, psychopathological and psychotherapeutic processes.

CONCLUSIONS

Research on computational methods will benefit from a careful selection of reference methods and standards. Therefore, this review represents an attempt to systematise the empirical literature on the applications of computational methods in psychotherapy research in order to offer clinicians an overview of the usefulness of these methods and the possibilities of their use in the various fields of application, highlighting their clinical implications, and ultimately attempting to identify potential opportunities for further research.

Bidirectional Representation Learning From Transformers Using Multimodal Electronic Health Record Data to Predict Depression

Advancements in machine learning algorithms have had a beneficial impact on representation learning, classification, and prediction models built using electronic health record (EHR) data. Effort has been put both on increasing models' overall performance as well as improving their interpretability, particularly regarding the decision-making process. In this study, we present a temporal deep learning model to perform bidirectional representation learning on EHR sequences with a transformer architecture to predict future diagnosis of depression. This model is able to aggregate five heterogenous and high-dimensional data sources from the EHR and process them in a temporal manner for chronic disease prediction at various prediction windows. We applied the current trend of pretraining and fine-tuning on EHR data to outperform the current state-of-the-art in chronic disease prediction, and to demonstrate the underlying relation between EHR codes in the sequence. The model generated the highest increases of precision-recall area under the curve (PRAUC) from 0.70 to 0.76 in depression prediction compared to the best baseline model. Furthermore, the self-attention weights in each sequence quantitatively demonstrated the inner relationship between various codes, which improved the model's interpretability. These results demonstrate the model's ability to utilize heterogeneous EHR data to predict depression while achieving high accuracy and interpretability, which may facilitate constructing clinical decision support systems in the future for chronic disease screening and early detection.

Depression and Coping Styles of College Students in China During COVID-19 Pandemic: A Systemic Review and Meta-Analysis

Background: The rapid spread and uncertain outcome of the 2019 novel coronavirus disease (COVID-19) around the world have caused worry, fear, and stress among the general population. Nevertheless, the prevalence of depression among college students in China during lockdown, following the COVID-19 pandemic, and their coping strategies have not been quantitatively assessed. Objective: We aimed to evaluate the prevalence of depression among college students in China during the lockdown due to the COVID-19 pandemic and assess their coping strategies. Methods: Systematic review and meta-analysis were conducted to assess the prevalence of depression among college students in China and their coping strategies. Results: The results indicated that, during lockdown in the COVID-19 pandemic, the prevalence rates of college students in China suffering from mild, moderate, and severe depression were 25% (95% CI = 17-33%), 7% (95% CI = 2-14%), and 2% (95% CI = 1-5%), respectively. Besides, the proportion of college students who use WeChat and Weibo to acquire COVID-19 knowledge was 39% (95% CI = 13-68%), whereas the proportion of college students using mental health application services (APPs) to deal with depression was 59% (95% CI = 41-73%). Conclusions: The prevalence of depression among college students in China was high during the lockdown in the COVID-19 pandemic. Thus, considering the adverse outcomes of depression, it is imperative to screen college students in China for depression during the CIVID-19 pandemic and provide them with necessary psychological interventions to control and prevent depression. Social media platforms, such as WeChat and Weibo, and mental health APPs could provide an opportunity for psychological health information dissemination for college students. However, their effectiveness in reducing depression will have to be assessed.

The promise of machine learning in predicting treatment outcomes in psychiatry

For many years, psychiatrists have tried to understand factors involved in response to medications or psychotherapies, in order to personalize their treatment choices. There is now a broad and growing interest in the idea that we can develop models to personalize treatment decisions using new statistical approaches from the field of machine learning and applying them to larger volumes of data. In this pursuit, there has been a paradigm shift away from experimental studies to confirm or refute specific hypotheses towards a focus on the overall explanatory power of a predictive model when tested on new, unseen datasets. In this paper, we review key studies using machine learning to predict treatment outcomes in psychiatry, ranging from medications and psychotherapies to digital interventions and neurobiological treatments. Next, we focus on some new sources of data that are being used for the development of predictive models based on machine learning, such as electronic health records, smartphone and social media data, and on the potential utility of data from genetics, electrophysiology, neuroimaging and cognitive testing. Finally, we discuss how far the field has come towards implementing prediction tools in real-world clinical practice. Relatively few retrospective studies to-date include appropriate external validation procedures, and there are even fewer prospective studies testing the clinical feasibility and effectiveness of predictive models. Applications of machine learning in psychiatry face some of the same ethical challenges posed by these techniques in other areas of medicine or computer science, which we discuss here. In short, machine learning is a nascent but important approach to improve the effectiveness of mental health care, and several prospective clinical studies suggest that it may be working already.

HCET: Hierarchical Clinical Embedding With Topic Modeling on Electronic Health Records for Predicting Future Depression

Recent developments in machine learning algorithms have enabled models to exhibit impressive performance in healthcare tasks using electronic health record (EHR) data. However, the heterogeneous nature and sparsity of EHR data remains challenging. In this work, we present a model that utilizes heterogeneous data and addresses sparsity by representing diagnoses, procedures, and medication codes with temporal Hierarchical Clinical Embeddings combined with Topic modeling (HCET) on clinical notes. HCET aggregates various categories of EHR data and learns inherent structure based on hospital visits for an individual patient. We demonstrate the potential of the approach in the task of predicting depression at various time points prior to a clinical diagnosis. We found that HCET outperformed all baseline methods with a highest improvement of 0.07 in precision-recall area under the curve (PRAUC). Furthermore, applying attention weights across EHR data modalities significantly improved the performance as well as the model's interpretability by revealing the relative weight for each data modality. Our results demonstrate the model's ability to utilize heterogeneous EHR information to predict depression, which may have future implications for screening and early detection.

Depression screening using mobile phone usage metadata: a machine learning approach

OBJECTIVE

Depression is currently the second most significant contributor to non-fatal disease burdens globally. While it is treatable, depression remains undiagnosed in many cases. As mobile phones have now become an integral part of daily life, this study examines the possibility of screening for depressive symptoms continuously based on patients' mobile usage patterns.

MATERIALS AND METHODS

412 research participants reported a range of their mobile usage statistics. Beck Depression Inventory-2nd ed (BDI-II) was used to measure the severity of depression among participants. A wide array of machine learning classification algorithms was trained to detect participants with depression symptoms (ie, BDI-II score ≥ 14). The relative importance of individual variables was additionally quantified.

RESULTS

Participants with depression were found to have fewer saved contacts on their devices, spend more time on their mobile devices to make and receive fewer and shorter calls, and send more text messages than participants without depression. The best model was a random forest classifier, which had an out-of-sample balanced accuracy of 0.768. The balanced accuracy increased to 0.811 when participants' age and gender were included.

DISCUSSIONS/CONCLUSION

The significant predictive power of mobile usage attributes implies that, by collecting mobile usage statistics, mental health mobile applications can continuously screen for depressive symptoms for initial diagnosis or for monitoring the progress of ongoing treatments. Moreover, the input variables used in this study were aggregated mobile usage metadata attributes, which has low privacy sensitivity making it more likely for patients to grant required application permissions.

Identification of major depressive disorder disease-related genes and functional pathways based on system dynamic changes of network connectivity

BACKGROUND

Major depressive disorder (MDD) is a leading psychiatric disorder that involves complex abnormal biological functions and neural networks. This study aimed to compare the changes in the network connectivity of different brain tissues under different pathological conditions, analyzed the biological pathways and genes that are significantly related to disease progression, and further predicted the potential therapeutic drug targets.

METHODS

Expression of differentially expressed genes (DEGs) were analyzed with postmortem cingulate cortex (ACC) and prefrontal cortex (PFC) mRNA expression profile datasets downloaded from the Gene Expression Omnibus (GEO) database, including 76 MDD patients and 76 healthy subjects in ACC and 63 MDD patients and 63 healthy subjects in PFC. The co-expression network construction was based on system network analysis. The function of the genes was annotated by Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis. Human Protein Reference Database (HPRD, http://www.hprd.org/ ) was used for gene interaction relationship mapping.

RESULTS

We filtered 586 DEGs in ACC and 616 DEGs in PFC for further analysis. By constructing the co-expression network, we found that the gene connectivity was significantly reduced under disease conditions (P = 0.04 in PFC and P = 1.227e-09 in ACC). Crosstalk analysis showed that CD19, PTDSS2 and NDST2 were significantly differentially expressed in ACC and PFC of MDD patients. Among them, CD19 and PTDSS2 have been targeted by several drugs in the Drugbank database. KEGG pathway analysis demonstrated that the function of CD19 and PTDSS2 were enriched with the pathway of Glycerophospholipid metabolism and T cell receptor signaling pathway.

CONCLUSION

Co-expression network and tissue comparing analysis can identify signaling pathways and cross talk genes related to MDD, which may provide novel insight for understanding the molecular mechanisms of MDD.

Predictive modeling of depression and anxiety using electronic health records and a novel machine learning approach with artificial intelligence

Generalized anxiety disorder (GAD) and major depressive disorder (MDD) are highly prevalent and impairing problems, but frequently go undetected, leading to substantial treatment delays. Electronic health records (EHRs) collect a great deal of biometric markers and patient characteristics that could foster the detection of GAD and MDD in primary care settings. We approached the problem of predicting MDD and GAD using a novel machine learning pipeline to re-analyze data from an observational study. The pipeline constitutes an ensemble of algorithmically distinct machine learning methods, including deep learning. A sample of 4,184 undergraduate students completed the study, undergoing a general health screening and completing a psychiatric assessment for MDD and GAD. After explicitly excluding all psychiatric information, 59 biomedical and demographic features from the general health survey in addition to a set of engineered features were used for model training. We assessed the model's performance on a held-out test set and found an AUC of 0.73 (sensitivity: 0.66, specificity: 0.7) and 0.67 (sensitivity: 0.55, specificity: 0.7) for GAD, and MDD, respectively. Additionally, we used advanced techniques (SHAP values) to illuminate which features had the greatest impact on prediction for each disease. The top predictive features for MDD were being satisfied with living conditions and having public health insurance. The top predictive features for GAD were vaccinations being up to date and marijuana use. Our results indicate moderate predictive performance for the application of machine learning methods in detection of GAD and MDD based on EHR data. By identifying important predictors of GAD and MDD, these results may be used in future research to aid in the early detection of MDD and GAD.

Dissecting diagnostic heterogeneity in depression by integrating neuroimaging and genetics

Depression is a heterogeneous and etiologically complex psychiatric syndrome, not a unitary disease entity, encompassing a broad spectrum of psychopathology arising from distinct pathophysiological mechanisms. Motivated by a need to advance our understanding of these mechanisms and develop new treatment strategies, there is a renewed interest in investigating the neurobiological basis of heterogeneity in depression and rethinking our approach to diagnosis for research purposes. Large-scale genome-wide association studies have now identified multiple genetic risk variants implicating excitatory neurotransmission and synapse function and underscoring a highly polygenic inheritance pattern that may be another important contributor to heterogeneity in depression. Here, we review various sources of phenotypic heterogeneity and approaches to defining and studying depression subtypes, including symptom-based subtypes and biology-based approaches to decomposing the depression syndrome. We review "dimensional," "categorical," and "hybrid" approaches to parsing phenotypic heterogeneity in depression and defining subtypes using functional neuroimaging. Next, we review recent progress in neuroimaging genetics (correlating neuroimaging patterns of brain function with genetic data) and its potential utility for generating testable hypotheses concerning molecular and circuit-level mechanisms. We discuss how genetic variants and transcriptomic profiles may confer risk for depression by modulating brain structure and function. We conclude by highlighting several promising areas for future research into the neurobiological underpinnings of heterogeneity, including efforts to understand sexually dimorphic mechanisms, the longitudinal dynamics of depressive episodes, and strategies for developing personalized treatments and facilitating clinical decision-making.

Using machine learning for predicting cervical cancer from Swedish electronic health records by mining hierarchical representations

Electronic health records (EHRs) contain rich documentation regarding disease symptoms and progression, but EHR data is challenging to use for diagnosis prediction due to its high dimensionality, relative scarcity, and substantial level of noise. We investigated how to best represent EHR data for predicting cervical cancer, a serious disease where early detection is beneficial for the outcome of treatment. A case group of 1321 patients with cervical cancer were matched to ten times as many controls, and for both groups several types of events were extracted from their EHRs. These events included clinical codes, lab results, and contents of free text notes retrieved using a LSTM neural network. Clinical events are described with great variation in EHR texts, leading to a very large feature space. Therefore, an event hierarchy inferred from the textual events was created to represent the clinical texts. Overall, the events extracted from free text notes contributed the most to the final prediction, and the hierarchy of textual events further improved performance. Four classifiers were evaluated for predicting a future cancer diagnosis where Random Forest achieved the best results with an AUC of 0.70 from a year before diagnosis up to 0.97 one day before diagnosis. We conclude that our approach is sound and had excellent discrimination at diagnosis, but only modest discrimination capacity before this point. Since our study objective was earlier disease prediction than such, we propose further work should consider extending patient histories through e.g. the integration of primary health records preceding referral to hospital.

Artificial Intelligence and Suicide Prevention: A Systematic Review of Machine Learning Investigations

Suicide is a leading cause of death that defies prediction and challenges prevention efforts worldwide. Artificial intelligence (AI) and machine learning (ML) have emerged as a means of investigating large datasets to enhance risk detection. A systematic review of ML investigations evaluating suicidal behaviors was conducted using PubMed/MEDLINE, PsychInfo, Web-of-Science, and EMBASE, employing search strings and MeSH terms relevant to suicide and AI. Databases were supplemented by hand-search techniques and Google Scholar. Inclusion criteria: (1) journal article, available in English, (2) original investigation, (3) employment of AI/ML, (4) evaluation of a suicide risk outcome. N = 594 records were identified based on abstract search, and 25 hand-searched reports. N = 461 reports remained after duplicates were removed, n = 316 were excluded after abstract screening. Of n = 149 full-text articles assessed for eligibility, n = 87 were included for quantitative synthesis, grouped according to suicide behavior outcome. Reports varied widely in methodology and outcomes. Results suggest high levels of risk classification accuracy (>90%) and Area Under the Curve (AUC) in the prediction of suicidal behaviors. We report key findings and central limitations in the use of AI/ML frameworks to guide additional research, which hold the potential to impact suicide on broad scale.

Subphenotyping depression using machine learning and electronic health records

Objective

To identify depression subphenotypes from Electronic Health Records (EHRs) using machine learning methods, and analyze their characteristics with respect to patient demographics, comorbidities, and medications.

Materials and Methods

Using EHRs from the INSIGHT Clinical Research Network (CRN) database, multiple machine learning (ML) algorithms were applied to analyze 11 275 patients with depression to discern depression subphenotypes with distinct characteristics.

Results

Using the computational approaches, we derived three depression subphenotypes: Phenotype_A (n = 2791; 31.35%) included patients who were the oldest (mean (SD) age, 72.55 (14.93) years), had the most comorbidities, and took the most medications. The most common comorbidities in this cluster of patients were hyperlipidemia, hypertension, and diabetes. Phenotype_B (mean (SD) age, 68.44 (19.09) years) was the largest cluster (n = 4687; 52.65%), and included patients suffering from moderate loss of body function. Asthma, fibromyalgia, and Chronic Pain and Fatigue (CPF) were common comorbidities in this subphenotype. Phenotype_C (n = 1452; 16.31%) included patients who were younger (mean (SD) age, 63.47 (18.81) years), had the fewest comorbidities, and took fewer medications. Anxiety and tobacco use were common comorbidities in this subphenotype.

Conclusion

Computationally deriving depression subtypes can provide meaningful insights and improve understanding of depression as a heterogeneous disorder. Further investigation is needed to assess the utility of these derived phenotypes to inform clinical trial design and interpretation in routine patient care.

The Predictive Validity of Machine Learning Models in the Classification and Treatment of Major Depressive Disorder: State of the Art and Future Directions

Major depressive disorder imposes a substantial disease burden worldwide, ranking as the third leading contributor to global disability. In spite of its ubiquity, classifying and treating depression has proven troublesome. One argument put forward to explain this predicament is the heterogeneity of patients diagnosed with the disorder. Recently, many areas of daily life have witnessed the surge of machine learning techniques, computational approaches to elucidate complex patterns in large datasets, which can be employed to make predictions and detect relevant clusters. Due to the multidimensionality at play in the pathogenesis of depression, it is suggested that machine learning could contribute to improving classification and treatment. In this paper, we investigated literature focusing on the use of machine learning models on datasets with clinical variables of patients diagnosed with depression to predict treatment outcomes or find more homogeneous subgroups. Identified studies based on best practices in the field are evaluated. We found 16 studies predicting outcomes (such as remission) and identifying clusters in patients with depression. The identified studies are mostly still in proof-of-concept phase, with small datasets, lack of external validation, and providing single performance metrics. Larger datasets, and models with similar variables present across these datasets, are needed to develop accurate and generalizable models. We hypothesize that harnessing natural language processing to obtain data 'hidden' in clinical texts might prove useful in improving prediction models. Besides, researchers will need to focus on the conditions to feasibly implement these models to support psychiatrists and patients in their decision-making in practice. Only then we can enter the realm of precision psychiatry.

Mobile and wearable technology for monitoring depressive symptoms in children and adolescents: A scoping review

BACKGROUND

There has been rapid growth of mobile and wearable tools that may help to overcome challenges in the diagnosis and prediction of Major Depressive Disorder in children and adolescents, tasks that rely on clinical reporting that is inherently based on retrospective recall of symptoms and associated features. This article reviews more objective ways of measuring and monitoring mood within this population.

METHODS

A scoping review of peer-reviewed studies examined published research that employs mobile and wearable tools to characterize depression in children and/or adolescents. Our search strategy included the following terms: (1) monitoring or prediction (2) depression (3) mobile apps or wearables and (4) children and youth (including adolescents), and was applied to five databases.

RESULTS

Our search produced 829 citations (2008- Feb 2019), of which 30 (journal articles, conference papers and abstracts) were included in the analysis, and 2 reviews included in our discussion. The majority of the evidence involved smartphone apps, with very few studies using actigraphy. Mobile and wearables captured a variety of data including unobtrusive passive analytics, movement and light data, plus physical and mental health data, including depressive symptom monitoring. Most studies also examined feasibility.

LIMITATIONS

This review was limited to published research in the English language. The review criteria excluded any apps that were mainly treatment focused, therefore there was not much of a focus on clinical outcomes.

CONCLUSIONS

This scoping review yielded a variety of studies with heterogeneous populations, research methods and study objectives, which limited our ability to address our research objectives cohesively. Certain mobile technologies, however, have demonstrated feasibility for tracking depression that could inform models for predicting relapse.

BioHackathon 2015: Semantics of data for life sciences and reproducible research

We report on the activities of the 2015 edition of the BioHackathon, an annual event that brings together researchers and developers from around the world to develop tools and technologies that promote the reusability of biological data. We discuss issues surrounding the representation, publication, integration, mining and reuse of biological data and metadata across a wide range of biomedical data types of relevance for the life sciences, including chemistry, genotypes and phenotypes, orthology and phylogeny, proteomics, genomics, glycomics, and metabolomics. We describe our progress to address ongoing challenges to the reusability and reproducibility of research results, and identify outstanding issues that continue to impede the progress of bioinformatics research. We share our perspective on the state of the art, continued challenges, and goals for future research and development for the life sciences Semantic Web.

Defining Major Depressive Disorder Cohorts Using the EHR: Multiple Phenotypes Based on ICD-9 Codes and Medication Orders

Background

Major Depressive Disorder (MDD) is one of the most common mental illnesses and a leading cause of disability worldwide. Electronic Health Records (EHR) allow researchers to conduct unprecedented large-scale observational studies investigating MDD, its disease development and its interaction with other health outcomes. While there exist methods to classify patients as clear cases or controls, given specific data requirements, there are presently no simple, generalizable, and validated methods to classify an entire patient population into varying groups of depression likelihood and severity.

Methods

We have tested a simple, pragmatic electronic phenotype algorithm that classifies patients into one of five mutually exclusive, ordinal groups, varying in depression phenotype. Using data from an integrated health system on 278,026 patients from a 10-year study period we have tested the convergent validity of these constructs using measures of external validation, including patterns of psychiatric prescriptions, symptom severity, indicators of suicidality, comorbidity, mortality, health care utilization, and polygenic risk scores for MDD.

Results

We found consistent patterns of increasing morbidity and/or adverse outcomes across the five groups, providing evidence for convergent validity.

Limitations

The study population is from a single rural integrated health system which is predominantly white, possibly limiting its generalizability.

Conclusion

Our study provides initial evidence that a simple algorithm, generalizable to most EHR data sets, provides categories with meaningful face and convergent validity that can be used for stratification of an entire patient population.

Use of Patient-Reported Data to Match Depression Screening Intervals With Depression Risk Profiles in Primary Care Patients With Diabetes: Development and Validation of Prediction Models for Major Depression

Background

Clinical guidelines recommend screening for depression in the general adult population but recognizes that the optimum interval for screening is unknown. Ideal screening intervals should match the patient risk profiles.

Objective

This study describes a predictive analytics approach for mining clinical and patient-reported data from a large clinical study for the identification of primary care patients at high risk for depression to match depression screening intervals with patient risk profiles.

Methods

This paper analyzed data from a large safety-net primary care study for diabetes and depression. A regression-based data mining technique was used to examine 53 demographics, clinical variables, and patient-reported variables to develop three prediction models for major depression at 6, 12, and 18 months from baseline. Predictors with the strongest predictive power that require low information collection efforts were selected to develop the prediction models. Predictive accuracy was measured by the area under the receiver operating curve (AUROC) and was evaluated by 10-fold cross-validation. The effectiveness of the prediction algorithms in supporting clinical decision making for six “typical” types of patients was demonstrated.

Results

The analysis included 923 patients who were nondepressed at the study baseline. Five patient-reported variables were selected in the prediction models to predict major depression at 6, 12, and 18 months: (1) Patient Health Questionnaire 2-item score; (2) the Sheehan Disability Scale; (3) previous problems with depression; (4) the diabetes symptoms scale; and (5) emotional burden of diabetes. All three depression prediction models had an AUROC>0.80, comparable with published depression prediction studies. Among the 6 “typical” types of patients, the algorithms suggest that patients who reported impaired daily functioning by health status are at an elevated risk for depression in all three periods.

Conclusions

This study demonstrated that leveraging patient-reported data and prediction models can help improve identification of high-risk patients and clinical decisions about the depression screening interval for diabetes patients. Implementation of this approach can be coupled with application of modern technologies such as telehealth and mobile health assessment for collecting patient-reported data to improve privacy, reducing stigma and costs, and promoting a personalized depression screening that matches screening intervals with patient risk profiles.

A machine learning ensemble to predict treatment outcomes following an Internet intervention for depression

BACKGROUND

Some Internet interventions are regarded as effective treatments for adult depression, but less is known about who responds to this form of treatment.

METHOD

An elastic net and random forest were trained to predict depression symptoms and related disability after an 8-week course of an Internet intervention, Deprexis, involving adults (N = 283) from across the USA. Candidate predictors included psychopathology, demographics, treatment expectancies, treatment usage, and environmental context obtained from population databases. Model performance was evaluated using predictive R2$\lpar R_{{\rm pred}}^2\rpar\comma $ the expected variance explained in a new sample, estimated by 10 repetitions of 10-fold cross-validation.

RESULTS

An ensemble model was created by averaging the predictions of the elastic net and random forest. Model performance was compared with a benchmark linear autoregressive model that predicted each outcome using only its baseline. The ensemble predicted more variance in post-treatment depression (8.0% gain, 95% CI 0.8-15; total $R_{{\rm pred}}^2 \; $= 0.25), disability (5.0% gain, 95% CI -0.3 to 10; total $R_{{\rm pred}}^2 \; $= 0.25), and well-being (11.6% gain, 95% CI 4.9-19; total $R_{{\rm pred}}^2 \; $= 0.29) than the benchmark model. Important predictors included comorbid psychopathology, particularly total psychopathology and dysthymia, low symptom-related disability, treatment credibility, lower access to therapists, and time spent using certain Deprexis modules.

CONCLUSION

A number of variables predict symptom improvement following an Internet intervention, but each of these variables makes relatively small contributions. Machine learning ensembles may be a promising statistical approach for identifying the cumulative contribution of many weak predictors to psychosocial depression treatment response.

Use of machine learning techniques in the development and refinement of a predictive model for early diagnosis of ankylosing spondylitis

OBJECTIVE

To develop a predictive mathematical model for the early identification of ankylosing spondylitis (AS) based on the medical and pharmacy claims history of patients with and without AS.

METHODS

This retrospective study used claims data from Truven databases from January 2006 to September 2015 (Segment 1) and October 2015 to February 2018 (Segment 2). Machine learning identified features differentiating patients with AS from matched controls; selected features were used as inputs in developing Model A/B to identify patients likely to have AS. Model A/B was trained and developed in Segment 1, and patients predicted to have AS in Segment 1 were followed up in Segment 2 to evaluate the predictive capability of Model A/B.

RESULTS

Of 228,471 patients in Segment 1 without any history of AS, Model A/B predicted 1923 patients to have AS. Ultimately, 1242 patients received an AS diagnosis in Segment 2; 120 of these were correctly predicted by Model A/B, yielding a positive predictive value (PPV) of 6.24%. The diagnostic accuracy of Model A/B compared favorably with that of a clinical model (PPV, 1.29%) that predicted AS based on spondyloarthritis features described in the Assessment of SpondyloArthritis international Society classification criteria. A simplified linear regression model created to test the operability of Model A/B yielded a lower PPV (2.55%).

CONCLUSIONS

Model A/B performed better than a clinically based model in predicting a diagnosis of AS among patients in a large claims database; its use may contribute to early recognition of AS and a timely diagnosis.

Synthesizing electronic health records using improved generative adversarial networks

Objective

The aim of this study was to generate synthetic electronic health records (EHRs). The generated EHR data will be more realistic than those generated using the existing medical Generative Adversarial Network (medGAN) method.

Materials and Methods

We modified medGAN to obtain two synthetic data generation models-designated as medical Wasserstein GAN with gradient penalty (medWGAN) and medical boundary-seeking GAN (medBGAN)-and compared the results obtained using the three models. We used 2 databases: MIMIC-III and National Health Insurance Research Database (NHIRD), Taiwan. First, we trained the models and generated synthetic EHRs by using these three 3 models. We then analyzed and compared the models' performance by using a few statistical methods (Kolmogorov-Smirnov test, dimension-wise probability for binary data, and dimension-wise average count for count data) and 2 machine learning tasks (association rule mining and prediction).

Results

We conducted a comprehensive analysis and found our models were adequately efficient for generating synthetic EHR data. The proposed models outperformed medGAN in all cases, and among the 3 models, boundary-seeking GAN (medBGAN) performed the best.

Discussion

To generate realistic synthetic EHR data, the proposed models will be effective in the medical industry and related research from the viewpoint of providing better services. Moreover, they will eliminate barriers including limited access to EHR data and thus accelerate research on medical informatics.

Conclusion

The proposed models can adequately learn the data distribution of real EHRs and efficiently generate realistic synthetic EHRs. The results show the superiority of our models over the existing model.

How data science can advance mental health research

Accessibility of powerful computers and availability of so-called big data from a variety of sources means that data science approaches are becoming pervasive. However, their application in mental health research is often considered to be at an earlier stage than in other areas despite the complexity of mental health and illness making such a sophisticated approach particularly suitable. In this Perspective, we discuss current and potential applications of data science in mental health research using the UK Clinical Research Collaboration classification: underpinning research; aetiology; detection and diagnosis; treatment development; treatment evaluation; disease management; and health services research. We demonstrate that data science is already being widely applied in mental health research, but there is much more to be done now and in the future. The possibilities for data science in mental health research are substantial.

Chronic Kidney Disease stratification using office visit records: Handling data imbalance via hierarchical meta-classification

Background

Chronic Kidney Disease (CKD) is one of several conditions that affect a growing percentage of the US population; the disease is accompanied by multiple co-morbidities, and is hard to diagnose in-and-of itself. In its advanced forms it carries severe outcomes and can lead to death. It is thus important to detect the disease as early as possible, which can help devise effective intervention and treatment plan.

Here we investigate ways to utilize information available in electronic health records (EHRs) from regular office visits of more than 13,000 patients, in order to distinguish among several stages of the disease. While clinical data stored in EHRs provide valuable information for risk-stratification, one of the major challenges in using them arises from data imbalance. That is, records associated with a more severe condition are typically under-represented compared to those associated with a milder manifestation of the disease. To address imbalance, we propose and develop a sampling-based ensemble approach, hierarchical meta-classification, aiming to stratify CKD patients into severity stages, using simple quantitative non-text features gathered from standard office visit records.

Methods

The proposed hierarchical meta-classification method frames the multiclass classification task as a hierarchy of two subtasks. The first is binary classification, separating records associated with the majority class from those associated with all minority classes combined, using meta-classification. The second subtask separates the records assigned to the combined minority classes into the individual constituent classes.

Results

The proposed method identifies a significant proportion of patients suffering from the more advanced stages of the condition, while also correctly identifying most of the less severe cases, maintaining high sensitivity, specificity and F-measure (≥ 93%). Our results show that the high level of performance attained by our method is preserved even when the size of the training set is significantly reduced, demonstrating the stability and generalizability of our approach.

Conclusion

We present a new approach to perform classification while addressing data imbalance, which is inherent in the biomedical domain. Our model effectively identifies severity stages of CKD patients, using information readily available in office visit records within the realistic context of high data imbalance.

Feature Ranking in Predictive Models for Hospital-Acquired Acute Kidney Injury

Acute Kidney Injury (AKI) is a common complication encountered among hospitalized patients, imposing significantly increased cost, morbidity, and mortality. Early prediction of AKI has profound clinical implications because currently no treatment exists for AKI once it develops. Feature selection (FS) is an essential process for building accurate and interpretable prediction models, but to our best knowledge no study has investigated the robustness and applicability of such selection process for AKI. In this study, we compared eight widely-applied FS methods for AKI prediction using nine-years of electronic medical records (EMR) and examined heterogeneity in feature rankings produced by the methods. FS methods were compared in terms of stability with respect to data sampling variation, similarity between selection results, and AKI prediction performance. Prediction accuracy did not intrinsically guarantee the feature ranking stability. Across different FS methods, the prediction performance did not change significantly, while the importance rankings of features were quite different. A positive correlation was observed between the complexity of suitable FS method and sample size. This study provides several practical implications, including recognizing the importance of feature stability as it is desirable for model reproducibility, identifying important AKI risk factors for further investigation, and facilitating early prediction of AKI.

Association networks in a matched case-control design - Co-occurrence patterns of preexisting chronic medical conditions in patients with major depression versus their matched controls

OBJECTIVE

We present a method for comparing association networks in a matched case-control design, which provides a high-level comparison of co-occurrence patterns of features after adjusting for confounding factors. We demonstrate this approach by examining the differential distribution of chronic medical conditions in patients with major depressive disorder (MDD) compared to the distribution of these conditions in their matched controls.

MATERIALS AND METHODS

Newly diagnosed MDD patients were matched to controls based on their demographic characteristics, socioeconomic status, place of residence, and healthcare service utilization in the Korean National Health Insurance Service's National Sample Cohort. Differences in the networks of chronic medical conditions in newly diagnosed MDD cases treated with antidepressants, and their matched controls, were prioritized with a permutation test accounting for the false discovery rate. Sensitivity analyses for the associations between prioritized pairs of chronic medical conditions and new MDD diagnosis were performed with regression modeling.

RESULTS

By comparing the association networks of chronic medical conditions in newly diagnosed depression patients and their matched controls, five pairs of such conditions were prioritized among 105 possible pairs after controlling the false discovery rate at 5%. In sensitivity analyses using regression modeling, four out of the five prioritized pairs were statistically significant for the interaction terms.

CONCLUSION

Association networks in a matched case-control design can provide a high-level comparison of comorbid features after adjusting for confounding factors, thereby supplementing traditional clinical study approaches. We demonstrate the differential co-occurrence pattern of chronic medical conditions in patients with MDD and prioritize the chronic conditions that have statistically significant interactions in regression models for depression.

The use of electronic health records for psychiatric phenotyping and genomics

The widespread adoption of electronic health record (EHRs) in healthcare systems has created a vast and continuously growing resource of clinical data and provides new opportunities for population-based research. In particular, the linking of EHRs to biospecimens and genomic data in biobanks may help address what has become a rate-limiting study for genetic research: the need for large sample sizes. The principal roadblock to capitalizing on these resources is the need to establish the validity of phenotypes extracted from the EHR. For psychiatric genetic research, this represents a particular challenge given that diagnosis is based on patient reports and clinician observations that may not be well-captured in billing codes or narrative records. This review addresses the opportunities and pitfalls in EHR-based phenotyping with a focus on their application to psychiatric genetic research. A growing number of studies have demonstrated that diagnostic algorithms with high positive predictive value can be derived from EHRs, especially when structured data are supplemented by text mining approaches. Such algorithms enable semi-automated phenotyping for large-scale case-control studies. In addition, the scale and scope of EHR databases have been used successfully to identify phenotypic subgroups and derive algorithms for longitudinal risk prediction. EHR-based genomics are particularly well-suited to rapid look-up replication of putative risk genes, studies of pleiotropy (phenomewide association studies or PheWAS), investigations of genetic networks and overlap across the phenome, and pharmacogenomic research. EHR phenotyping has been relatively under-utilized in psychiatric genomic research but may become a key component of efforts to advance precision psychiatry.

Using insurance claims to predict and improve hospitalizations and biologics use in members with inflammatory bowel diseases

OBJECTIVE

Inflammatory Bowel Disease (IBD) is an inflammatory disorder of the gastrointestinal tract that can necessitate hospitalization and the use of expensive biologics. Models predicting these interventions may improve patient quality of life and reduce expenditures.

MATERIALS AND METHODS

We used insurance claims from 2011 to 2013 to predict IBD-related hospitalizations and the initiation of biologics. We derived and optimized our model from a 2011 training set of 7771 members, predicting their outcomes the following year. The best-performing model was then applied to a 2012 validation set of 7450 members to predict their outcomes in 2013.

RESULTS

Our models predicted both IBD-related hospitalizations and the initiation of biologics, with average positive predictive values of 17% and 11%, respectively - each a 200% improvement over chance. Further, when we used topic modeling to identify four member subpopulations, the positive predictive value of predicting hospitalization increased to 20%.

DISCUSSION

We show that our hospitalization model, in concert with a mildly-effective interventional treatment plan for members identified as high-risk, may both improve patient outcomes and reduce insurance expenditures.

CONCLUSION

The success of our approach provides a roadmap for how claims data can complement traditional medical decision making with personalized, data-driven predictive medicine.

Data-based Decision Rules to Personalize Depression Follow-up

Depression is a common mental illness with complex and heterogeneous progression dynamics. Risk grouping of depression treatment population based on their longitudinal patterns has the potential to enable cost-effective monitoring policy design. This paper establishes a rule-based method to identify a set of risk predictive patterns from person-level longitudinal disease measurements by integrating the data transformation, rule discovery and rule evaluation. We further extend the identified rules to create rule-based monitoring strategies to adaptively monitor individuals with different disease severities. We applied the rule-based method on an electronic health record (EHR) dataset of depression treatment population containing person-level longitudinal Patient Health Questionnaire (PHQ)-9 scores for assessing depression severity. 12 risk predictive rules are identified, and the rule-based prognostic model based on identified rules enables more accurate prediction of disease severity than other prognostic models including RuleFit, logistic regression and Support Vector Machine. Two rule-based monitoring strategies outperform the latest PHQ-9 based monitoring strategy by providing higher sensitivity and specificity. The rule-based method can lead to a better understanding of disease dynamics, achieving more accurate prognostics of disease progressions, personalizing follow-up intervals, and designing cost-effective monitoring of patients in clinical practice.

Clinical Informatics Researcher's Desiderata for the Data Content of the Next Generation Electronic Health Record

OBJECTIVE

Clinical informatics researchers depend on the availability of high-quality data from the electronic health record (EHR) to design and implement new methods and systems for clinical practice and research. However, these data are frequently unavailable or present in a format that requires substantial revision. This article reports the results of a review of informatics literature published from 2010 to 2016 that addresses these issues by identifying categories of data content that might be included or revised in the EHR.

MATERIALS AND METHODS

We used an iterative review process on 1,215 biomedical informatics research articles. We placed them into generic categories, reviewed and refined the categories, and then assigned additional articles, for a total of three iterations.

RESULTS

Our process identified eight categories of data content issues: Adverse Events, Clinician Cognitive Processes, Data Standards Creation and Data Communication, Genomics, Medication List Data Capture, Patient Preferences, Patient-reported Data, and Phenotyping.

DISCUSSION

These categories summarize discussions in biomedical informatics literature that concern data content issues restricting clinical informatics research. These barriers to research result from data that are either absent from the EHR or are inadequate (e.g., in narrative text form) for the downstream applications of the data. In light of these categories, we discuss changes to EHR data storage that should be considered in the redesign of EHRs, to promote continued innovation in clinical informatics.

CONCLUSION

Based on published literature of clinical informaticians' reuse of EHR data, we characterize eight types of data content that, if included in the next generation of EHRs, would find immediate application in advanced informatics tools and techniques.

Health Informatics via Machine Learning for the Clinical Management of Patients

OBJECTIVES

To review how health informatics systems based on machine learning methods have impacted the clinical management of patients, by affecting clinical practice.

METHODS

We reviewed literature from 2010-2015 from databases such as Pubmed, IEEE xplore, and INSPEC, in which methods based on machine learning are likely to be reported. We bring together a broad body of literature, aiming to identify those leading examples of health informatics that have advanced the methodology of machine learning. While individual methods may have further examples that might be added, we have chosen some of the most representative, informative exemplars in each case.

RESULTS

Our survey highlights that, while much research is taking place in this high-profile field, examples of those that affect the clinical management of patients are seldom found. We show that substantial progress is being made in terms of methodology, often by data scientists working in close collaboration with clinical groups.

CONCLUSIONS

Health informatics systems based on machine learning are in their infancy and the translation of such systems into clinical management has yet to be performed at scale.

Natural language processing systems for capturing and standardizing unstructured clinical information: A systematic review

We followed a systematic approach based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses to identify existing clinical natural language processing (NLP) systems that generate structured information from unstructured free text. Seven literature databases were searched with a query combining the concepts of natural language processing and structured data capture. Two reviewers screened all records for relevance during two screening phases, and information about clinical NLP systems was collected from the final set of papers. A total of 7149 records (after removing duplicates) were retrieved and screened, and 86 were determined to fit the review criteria. These papers contained information about 71 different clinical NLP systems, which were then analyzed. The NLP systems address a wide variety of important clinical and research tasks. Certain tasks are well addressed by the existing systems, while others remain as open challenges that only a small number of systems attempt, such as extraction of temporal information or normalization of concepts to standard terminologies. This review has identified many NLP systems capable of processing clinical free text and generating structured output, and the information collected and evaluated here will be important for prioritizing development of new approaches for clinical NLP.

Predicting Falls in People Aged 65 Years and Older from Insurance Claims

BACKGROUND

Accidental falls among people aged 65 years and older caused approximately 2,700,000 injuries, 27,000 deaths, and cost more than 34 billion dollars in the US annually in recent years. Here, we derive and validate a predictive model for falls based on a retrospective cohort of those 65 years and older.

METHODS

Insurance claims from a 1-year observational period were used to predict a fall-related claim in the following 2 years. The predictive model takes into account a person's age, sex, prescriptions, and diagnoses. Through random assignment, half of the people had their claims used to derive the model, while the remaining people had their claims used to validate the model.

RESULTS

Of 120,881 individuals with Aetna health insurance coverage, 12,431 (10.3%) members fell. During validation, people were risk stratified across 20 levels, where those in the highest risk stratum had 10.5 times the risk as those in the lowest stratum (33.1% vs 3.1%).

CONCLUSIONS

Using only insurance claims, individuals in this large cohort at high risk of falls could be readily identified up to 2 years in advance. Although external validation is needed, the findings support the use of the model to better target interventions.

Daehr

Electronic health records (EHR) provide a rich source of temporal data that present a unique opportunity to characterize disease patterns and risk of imminent disease. While many data-mining tools have been adopted for EHR-based disease early detection, linear discriminant analysis (LDA) is one of the most commonly used statistical methods. However, it is difficult to train an accurate LDA model for early disease diagnosis when too few patients are known to have the target disease. Furthermore, EHR data are heterogeneous with significant noise. In such cases, the covariance matrices used in LDA are usually singular and estimated with a large variance.

This article presents Daehr , an extension of the LDA framework using electronic health record data to address these issues. Beyond existing LDA analyzers, we propose Daehr to (1) eliminate the data noise caused by the manual encoding of EHR data and (2) lower the variance of parameter (covariance matrices) estimation for LDA models when only a few patients’ EHR are available for training. To achieve these two goals, we designed an iterative algorithm to improve the covariance matrix estimation with embedded data-noise/parameter-variance reduction for LDA. We evaluated Daehr extensively using the College Health Surveillance Network, a large, real-world EHR dataset. Specifically, our experiments compared the performance of LDA to three baselines (i.e., LDA and its derivatives) in identifying college students at high risk for mental health disorders from 23 U.S. universities. Experimental results demonstrate Daehr significantly outperforms the three baselines by achieving 1.4%--19.4% higher accuracy and a 7.5%--43.5% higher F1-score.

Improving risk prediction for depression via Elastic Net regression - Results from Korea National Health Insurance Services Data

Depression, despite its high prevalence, remains severely under-diagnosed across the healthcare system. This demands the development of data-driven approaches that can help screen patients who are at a high risk of depression. In this work, we develop depression risk prediction models that incorporate disease co-morbidities using logistic regression with Elastic Net. Using data from the one million twelve-year longitudinal cohort from Korean National Health Insurance Services (KNHIS), our model achieved an Area Under the Curve (AUC) of the Receiver Operating Characteristic (ROC) of 0.7818, compared to a traditional logistic regression model without co-morbidity analysis (AUC of 0.6992). We also showed co-morbidity adjusted Odds Ratios (ORs), which may be more accurate independent estimate of each predictor variable. In conclusion, inclusion of co-morbidity analysis improved the performance of depression risk prediction models.

Novel Use of Natural Language Processing (NLP) to Predict Suicidal Ideation and Psychiatric Symptoms in a Text-Based Mental Health Intervention in Madrid

Natural language processing (NLP) and machine learning were used to predict suicidal ideation and heightened psychiatric symptoms among adults recently discharged from psychiatric inpatient or emergency room settings in Madrid, Spain. Participants responded to structured mental and physical health instruments at multiple follow-up points. Outcome variables of interest were suicidal ideation and psychiatric symptoms (GHQ-12). Predictor variables included structured items (e.g., relating to sleep and well-being) and responses to one unstructured question, “how do you feel today?” We compared NLP-based models using the unstructured question with logistic regression prediction models using structured data. The PPV, sensitivity, and specificity for NLP-based models of suicidal ideation were 0.61, 0.56, and 0.57, respectively, compared to 0.73, 0.76, and 0.62 of structured data-based models. The PPV, sensitivity, and specificity for NLP-based models of heightened psychiatric symptoms (GHQ-12 ≥ 4) were 0.56, 0.59, and 0.60, respectively, compared to 0.79, 0.79, and 0.85 in structured models. NLP-based models were able to generate relatively high predictive values based solely on responses to a simple general mood question. These models have promise for rapidly identifying persons at risk of suicide or psychological distress and could provide a low-cost screening alternative in settings where lengthy structured item surveys are not feasible.

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

PURPOSE

'Big data' are accumulating in a multitude of domains and offer novel opportunities for research. The role of these resources in mental health investigations remains relatively unexplored, although a number of datasets are in use and supporting a range of projects. We sought to review big data resources and their use in mental health research to characterise applications to date and consider directions for innovation in future.

METHODS

A narrative review.

RESULTS

Clear disparities were evident in geographic regions covered and in the disorders and interventions receiving most attention.

DISCUSSION

We discuss the strengths and weaknesses of the use of different types of data and the challenges of big data in general. Current research output from big data is still predominantly determined by the information and resources available and there is a need to reverse the situation so that big data platforms are more driven by the needs of clinical services and service users.

The Drug Data to Knowledge Pipeline: Large-Scale Claims Data Classification for Pharmacologic Insight

In biomedical informatics, assigning drug codes to categories is a common step in the analysis pipeline. Unfortunately, incomplete mappings are the norm rather than the exception with coverage values less than 85% not uncommon. Here, we perform this linking task on a nationwide insurance claims database with over 13 million members who were dispensed, according to National Drug Codes (NDCs), over 50,000 unique product forms of medication. The chosen approach employs Cerner Multum's VantageRx and the U.S. National Library of Medicine's RxMix. As a result, 94.0% of the NDCs were successfully mapped to categories used by common drug terminologies, e.g., Anatomical Therapeutic Chemical (ATC). Implemented as an SQL database and scripts, the approach is generic and can be setup for a new data set in a few hours. Thus, the method is a viable option for large-scale drug classification.

Using LASSO Regression to Predict Rheumatoid Arthritis Treatment Efficacy

Rheumatoid arthritis (RA) accounts for one-fifth of the deaths due to arthritis, the leading cause of disability in the United States. Finding effective treatments for managing arthritis symptoms are a major challenge, since the mechanisms of autoimmune disorders are not fully understood and disease presentation differs for each patient. The American College of Rheumatology clinical guidelines for treatment consider the severity of the disease when deciding treatment, but do not include any prediction of drug efficacy. Using Electronic Health Records and Biomedical Linked Open Data (LOD), we demonstrate a method to classify patient outcomes using LASSO penalized regression. We show how Linked Data improves prediction and provides insight into how drug treatment regimes have different treatment outcome. Applying classifiers like this to decision support in clinical applications could decrease time to successful disease management, lessening a physical and financial burden on patients individually and the healthcare system as a whole.

Predicting mortality over different time horizons: which data elements are needed?

OBJECTIVE

Electronic health records (EHRs) are a resource for "big data" analytics, containing a variety of data elements. We investigate how different categories of information contribute to prediction of mortality over different time horizons among patients undergoing hemodialysis treatment.

MATERIAL AND METHODS

We derived prediction models for mortality over 7 time horizons using EHR data on older patients from a national chain of dialysis clinics linked with administrative data using LASSO (least absolute shrinkage and selection operator) regression. We assessed how different categories of information relate to risk assessment and compared discrete models to time-to-event models.

RESULTS

The best predictors used all the available data (c-statistic ranged from 0.72-0.76), with stronger models in the near term. While different variable groups showed different utility, exclusion of any particular group did not lead to a meaningfully different risk assessment. Discrete time models performed better than time-to-event models.

CONCLUSIONS

Different variable groups were predictive over different time horizons, with vital signs most predictive for near-term mortality and demographic and comorbidities more important in long-term mortality.

Social Media, Big Data, and Mental Health: Current Advances and Ethical Implications

Mental health (including substance abuse) is the fifth greatest contributor to the global burden of disease, with an economic cost estimated to be US $2.5 trillion in 2010, and expected to double by 2030. Developing information systems to support and strengthen population-level mental health monitoring forms a core part of the World Health Organization's Comprehensive Action Plan 2013-2020. In this paper, we review recent work that utilizes social media "big data" in conjunction with associated technologies like natural language processing and machine learning to address pressing problems in population-level mental health surveillance and research, focusing both on technological advances and core ethical challenges.