Predicting Suicide Among US Veterans Using Natural Language Processing-enriched Social and Behavioral Determinants of Health.

Despite recognizing the critical association between social and behavioral determinants of health (SBDH) and suicide risk, SBDHs from unstructured electronic health record (EHR) notes for suicide predictive modeling remain underutilized. This study investigates the impact of SBDH, identified from both structured and unstructured data utilizing a natural language processing (NLP) system, on suicide prediction within 7, 30, 90, and 180 days of discharge. Using EHR data of 2,987,006 Veterans between October 1, 2009, and September 30, 2015, from the US Veterans Health Administration (VHA), we designed a case-control study that demonstrates that incorporating structured and NLP-extracted SBDH significantly enhances the performance of three architecturally distinct suicide predictive models - elastic-net logistic regression, random forest (RF), and multilayer perceptron. For example, RF achieved notable improvements in suicide prediction within 180 days of discharge, with an increase in the area under the receiver operating characteristic curve from 83.57-84.25% (95% CI = 0.63%-0.98%, p-val < 0.001) and the area under the precision recall curve from 57.38-59.87% (95% CI = 3.86%-4.82%, p-val < 0.001) after integrating NLP-extracted SBDH. These findings underscore the potential of NLP-extracted SBDH in enhancing suicide prediction across various prediction timeframes, offering valuable insights for healthcare practitioners and policymakers.

TransformEHR: transformer-based encoder-decoder generative model to enhance prediction of disease outcomes using electronic health records.

Deep learning transformer-based models using longitudinal electronic health records (EHRs) have shown a great success in prediction of clinical diseases or outcomes. Pretraining on a large dataset can help such models map the input space better and boost their performance on relevant tasks through finetuning with limited data. In this study, we present TransformEHR, a generative encoder-decoder model with transformer that is pretrained using a new pretraining objective-predicting all diseases and outcomes of a patient at a future visit from previous visits. TransformEHR's encoder-decoder framework, paired with the novel pretraining objective, helps it achieve the new state-of-the-art performance on multiple clinical prediction tasks. Comparing with the previous model, TransformEHR improves area under the precision-recall curve by 2% (p < 0.001) for pancreatic cancer onset and by 24% (p = 0.007) for intentional self-harm in patients with post-traumatic stress disorder. The high performance in predicting intentional self-harm shows the potential of TransformEHR in building effective clinical intervention systems. TransformEHR is also generalizable and can be easily finetuned for clinical prediction tasks with limited data.

Intentional Self-Harm Among US Veterans With Traumatic Brain Injury or Posttraumatic Stress Disorder: Retrospective Cohort Study From 2008 to 2017.

BACKGROUND: Veterans with a history of traumatic brain injury (TBI) and/or posttraumatic stress disorder (PTSD) may be at increased risk of suicide attempts and other forms of intentional self-harm as compared to veterans without TBI or PTSD. OBJECTIVE: Using administrative data from the US Veterans Health Administration (VHA), we studied associations between TBI and PTSD diagnoses, and subsequent diagnoses of intentional self-harm among US veterans who used VHA health care between 2008 and 2017. METHODS: All veterans with encounters or hospitalizations for intentional self-harm were assigned "index dates" corresponding to the date of the first related visit; among those without intentional self-harm, we randomly selected a date from among the veteran's health care encounters to match the distribution of case index dates over the 10-year period. We then examined the prevalence of TBI and PTSD diagnoses within the 5-year period prior to veterans' index dates. TBI, PTSD, and intentional self-harm were identified using International Classification of Diseases diagnosis and external cause of injury codes from inpatient and outpatient VHA encounters. We stratified analyses by veterans' average yearly VHA utilization in the 5-year period before their index date (low, medium, or high). Variations in prevalence and odds of intentional self-harm diagnoses were compared by veterans' prior TBI and PTSD diagnosis status (TBI only, PTSD only, and comorbid TBI/PTSD) for each VHA utilization stratum. Multivariable models adjusted for age, sex, race, ethnicity, marital status, Department of Veterans Affairs service-connection status, and Charlson Comorbidity Index scores. RESULTS: About 6.7 million veterans with at least two VHA visits in the 5-year period before their index dates were included in the analyses; 86,644 had at least one intentional self-harm diagnosis during the study period. During the periods prior to veterans' index dates, 93,866 were diagnosed with TBI only; 892,420 with PTSD only; and 102,549 with comorbid TBI/PTSD. Across all three VHA utilization strata, the prevalence of intentional self-harm diagnoses was higher among veterans diagnosed with TBI, PTSD, or TBI/PTSD than among veterans with neither diagnosis. The observed difference was most pronounced among veterans in the high VHA utilization stratum. The prevalence of intentional self-harm was six times higher among those with comorbid TBI/PTSD (6778/58,295, 11.63%) than among veterans with neither TBI nor PTSD (21,979/1,144,991, 1.92%). Adjusted odds ratios suggested that, after accounting for potential confounders, veterans with TBI, PTSD, or comorbid TBI/PTSD had higher odds of self-harm compared to veterans without these diagnoses. Among veterans with high VHA utilization, those with comorbid TBI/PTSD were 4.26 (95% CI 4.15-4.38) times more likely to receive diagnoses for intentional self-harm than veterans with neither diagnosis. This pattern was similar for veterans with low and medium VHA utilization. CONCLUSIONS: Veterans with TBI and/or PTSD diagnoses, compared to those with neither diagnosis, were substantially more likely to be subsequently diagnosed with intentional self-harm between 2008 and 2017. These associations were most pronounced among veterans who used VHA health care most frequently. These findings suggest a need for suicide prevention efforts targeted at veterans with these diagnoses.

Automated identification of eviction status from electronic health record notes.

OBJECTIVE: Evictions are important social and behavioral determinants of health. Evictions are associated with a cascade of negative events that can lead to unemployment, housing insecurity/homelessness, long-term poverty, and mental health problems. In this study, we developed a natural language processing system to automatically detect eviction status from electronic health record (EHR) notes. MATERIALS AND METHODS: We first defined eviction status (eviction presence and eviction period) and then annotated eviction status in 5000 EHR notes from the Veterans Health Administration (VHA). We developed a novel model, KIRESH, that has shown to substantially outperform other state-of-the-art models such as fine-tuning pretrained language models like BioBERT and Bio_ClinicalBERT. Moreover, we designed a novel prompt to further improve the model performance by using the intrinsic connection between the 2 subtasks of eviction presence and period prediction. Finally, we used the Temperature Scaling-based Calibration on our KIRESH-Prompt method to avoid overconfidence issues arising from the imbalance dataset. RESULTS: KIRESH-Prompt substantially outperformed strong baseline models including fine-tuning the Bio_ClinicalBERT model to achieve 0.74672 MCC, 0.71153 Macro-F1, and 0.83396 Micro-F1 in predicting eviction period and 0.66827 MCC, 0.62734 Macro-F1, and 0.7863 Micro-F1 in predicting eviction presence. We also conducted additional experiments on a benchmark social determinants of health (SBDH) dataset to demonstrate the generalizability of our methods. CONCLUSION AND FUTURE WORK: KIRESH-Prompt has substantially improved eviction status classification. We plan to deploy KIRESH-Prompt to the VHA EHRs as an eviction surveillance system to help address the US Veterans' housing insecurity.

Associations Between Natural Language Processing-Enriched Social Determinants of Health and Suicide Death Among US Veterans.

Importance: Social determinants of health (SDOHs) are known to be associated with increased risk of suicidal behaviors, but few studies use SDOHs from unstructured electronic health record notes. Objective: To investigate associations between veterans' death by suicide and recent SDOHs, identified using structured and unstructured data. Design, Setting, and Participants: This nested case-control study included veterans who received care under the US Veterans Health Administration from October 1, 2010, to September 30, 2015. A natural language processing (NLP) system was developed to extract SDOHs from unstructured clinical notes. Structured data yielded 6 SDOHs (ie, social or familial problems, employment or financial problems, housing instability, legal problems, violence, and nonspecific psychosocial needs), NLP on unstructured data yielded 8 SDOHs (social isolation, job or financial insecurity, housing instability, legal problems, barriers to care, violence, transition of care, and food insecurity), and combining them yielded 9 SDOHs. Data were analyzed in May 2022. Exposures: Occurrence of SDOHs over a maximum span of 2 years compared with no occurrence of SDOH. Main Outcomes and Measures: Cases of suicide death were matched with 4 controls on birth year, cohort entry date, sex, and duration of follow-up. Suicide was ascertained by National Death Index, and patients were followed up for up to 2 years after cohort entry with a study end date of September 30, 2015. Adjusted odds ratios (aORs) and 95% CIs were estimated using conditional logistic regression. Results: Of 6 122 785 veterans, 8821 committed suicide during 23 725 382 person-years of follow-up (incidence rate 37.18 per 100 000 person-years). These 8821 veterans were matched with 35 284 control participants. The cohort was mostly male (42 540 [96.45%]) and White (34 930 [79.20%]), with 6227 (14.12%) Black veterans. The mean (SD) age was 58.64 (17.41) years. Across the 5 common SDOHs, NLP-extracted SDOH, on average, retained 49.92% of structured SDOHs and covered 80.03% of all SDOH occurrences. SDOHs, obtained by structured data and/or NLP, were significantly associated with increased risk of suicide. The 3 SDOHs with the largest effect sizes were legal problems (aOR, 2.66; 95% CI, 2.46-2.89), violence (aOR, 2.12; 95% CI, 1.98-2.27), and nonspecific psychosocial needs (aOR, 2.07; 95% CI, 1.92-2.23), when obtained by combining structured data and NLP. Conclusions and Relevance: In this study, NLP-extracted SDOHs, with and without structured SDOHs, were associated with increased risk of suicide among veterans, suggesting the potential utility of NLP in public health studies.

Tailoring Galactose Oxidase for Self-Powered Benzyl Alcohol Sensing.

Benzyl alcohol (BnOH) is a widely-used preservative in a variety of cosmetics, but the excess addition (≥1.0 %) may cause strong symptoms such as nausea, gastrointestinal irritation, convulsion, even death, making it crucial to monitor and control the addition quantity. Herein, we have developed a test-strip-like BnOH detection method via tailoring a galactose oxidase (GOase) towards BnOH oxidation and preparing a self-powered electrochromic strip for BnOH concentration visualization. A double-substituted GOase variant (Y329S/R330F), on the basis of the reported GOase M1 , has been obtained by semi-rational design with a 24.6-fold improved activity towards BnOH compared to GOase M1 . The GOase Y329S/R330F electrode has a response to BnOH with a linear range of 0.04 to 3.25 mM (R2 =0.9985), a sensitivity of 122.78 µA mM-1 cm-2 , and a detection limit of 0.03 mM (S/N=3). Coupling an electrochromic Prussian blue (PB) cathode helps the successful sensing visualization without any further power supply. The present sensing is more convenient and user-friendly than the generally used gas chromatography (GC) and high performance liquid chromatography (HPLC), and brings a more accessible solution to the field of quality controlling.

Effect of microbubbles on transcranial doppler ultrasound-assisted intracranial recombinant tissue-type plasminogen activator thrombolysis.

OBJECTIVES: The aim of this study was to evaluate the effect of microbubbles on the efficacy of transcranial doppler (TCD) ultrasound-assisted thrombolytic therapy of recombinant tissue-type plasminogen activator (rt-PA). METHODS: Male New Zealand white rabbits (n = 36) were randomly divided into an rt-PA group (n = 18) and an rt-PA plus microbubble group (n = 18). After the cerebral infarction model was constructed with autologous blood clots, rt-PA and rt-PA plus microbubble intervention were performed, respectively. The hemodynamic changes and infarct size of the two groups were recorded. In addition, the ELISA method was used to detect the level of nitric oxide (NO), superoxide dismutase (SOD), and malondialdehyde (MDA) in the brain tissue of the two-group graph model and high-sensitivity C-reactive protein (hs-CRP) in the serum. RESULTS: In the rt-PA group, the recanalization rate was 38.9% and the average infarct size was 11.8%. In the rt-PA plus microbubble group, the recanalization rate was 66.7% and the average infarct size was 8.2%. In addition, the average values for NO, SOD, MDA, and hs-CRP were 16.48 ± 5.39 µmol/L, 730.2 ± 9.86 U/mg, 0.92 ± 0.43 nmol/mg, and 8.56 ± 1.64 mg/L in the rt-PA group, respectively, and the average values were 9.18 ± 3.37 µmol/L, 426.2 ± 6.39 U/mg, 0.73 ± 0.44 nmol/mg, and 5.23 ± 0.94 mg/L in the rt-PA plus microbubble group, respectively. CONCLUSIONS: The addition of microbubbles enhanced the effects of TCD-assisted rrt-PA thrombolysis.

MSCs-Derived miR-150-5p-Expressing Exosomes Promote Skin Wound Healing by Activating PI3K/AKT Pathway through PTEN.

Background and Objectives: The goal of this study was to investigate the mechanism of mesenchymal stem cell (MSC)-derived microRNA (miR)-150-5p-expressing exosomes in promoting skin wound healing through activating PI3K/AKT pathway by PTEN. Methods and Results: Human umbilical cord (HUC)-MSCs were infected with miR-150-5p overexpression and its control lentivirus, and HUC-MSCs-derived exosomes (MSCs-Exos) with stable expression of miR-150-5p were obtained. HaCaT cells were induced by H2O2 to establish a cellular model of skin injury, in which the expression of miR-150-5p and PTEN and the phosphorylation of PI3K and AKT were evaluated. HaCaT cells were transfected with pcDNA3.1-PTEN or pcDNA3.1 and then cultured with normal exosomes or exosomes stably expressing miR-150-5p. Cell proliferation was inspected by CCK-8. Cell migration was detected by scratch test and cell apoptosis by flow cytometry. The starBase tool was used to predict the binding site of miR-150-5p to PTEN. Dual-luciferase reporter assay and RIP assay were applied to assess the interaction between miR-150-5p and PTEN. In H2O2-induced HaCaT cells, the miR-150-5p expression decreased, and PTEN expression increased in a concentration-dependent manner. MSCs-Exos promoted the growth and migration of H2O2-induced HaCaT cells and inhibited their apoptosis. In addition, overexpression of exosomal miR-150-5p enhanced the protective effect of MSCs-Exos on H2O2-induced HaCaT cells; PTEN overexpression in HaCaT cells partially restrained miR-150-5p-mediated inhibition on H2O2-induced injury in HaCaT cells. PTEN was a target gene of miR-150-5p. MiR-150-5p regulated PI3K/AKT pathway through PTEN. Conclusions: MSCs-derived miR-150-5p-expressing exosomes promote skin wound healing by activating PI3K/AKT pathway through PTEN.

Prevalence of Frailty and Associations with Oral Anticoagulant Prescribing in Atrial Fibrillation.

BACKGROUND: Frailty is often cited as a factor influencing oral anticoagulation (OAC) prescription in patients with non-valvular atrial fibrillation (NVAF). We sought to determine the prevalence of frailty and its association with OAC prescription in older veterans with NVAF. METHODS: We used ICD-9 codes in Veterans Affairs (VA) records and Medicare claims data to identify patients with NVAF and CHA2DS2VASC ≥2 receiving care between February 2010 and September 2015. We examined rates of OAC prescription, further stratified by direct oral anticoagulant (DOAC) or vitamin K antagonist (VKA). Participants were characterized into 3 categories: non-frail, pre-frail, and frail based on a validated 30-item EHR-derived frailty index. We examined relations between frailty and OAC receipt; and frailty and type of OAC prescribed in regression models adjusted for factors related to OAC prescription. RESULTS: Of 308,664 veterans with NVAF and a CHA2DS2VASC score ≥2, 121,839 (39%) were prescribed OAC (73% VKA). The mean age was 77.7 (9.6) years; CHA2DS2VASC and ATRIA scores were 4.6 (1.6) and 5.0 (2.9) respectively. Approximately a third (38%) were frail, another third (32%) were pre-frail, and the remainder were not frail. Veterans prescribed OAC were younger, had higher bleeding risk, and were less likely to be frail than participants not receiving OAC (all p's<0.001). After adjustment for factors associated with OAC use, pre-frail (OR: 0.89, 95% CI: 0.87-0.91) and frail (OR: 0.66, 95% CI: 0.64-0.68) veterans were significantly less likely to be prescribed OAC than non-frail veterans. Of those prescribed OAC, pre-frail (OR:1.27, 95% CI: 1.22-1.31) and frail (OR: 1.75, 95% CI: 1.67-1.83) veterans were significantly more likely than non-frail veterans to be prescribed a DOAC than a VKA. CONCLUSIONS: There are high rates of frailty among older veterans with NVAF. Frailty using an EHR-derived index is associated with decreased OAC prescription.

Guideline-discordant dosing of direct-acting oral anticoagulants in the veterans health administration.

BACKGROUND: Clear guidelines exist to guide the dosing of direct-acting oral anticoagulants (DOACs). It is not known how consistently these guidelines are followed in practice. METHODS: We studied patients from the Veterans Health Administration (VA) with non-valvular atrial fibrillation who received DOACs (dabigatran, rivaroxaban, apixaban) between 2010 and 2016. We used patient characteristics (age, creatinine, body mass) to identify which patients met guideline recommendations for low-dose therapy and which for full-dose therapy. We examined how often patient dosing was concordant with these recommendations. We examined variation in guideline-concordant dosing by site of care and over time. We examined patient-level predictors of guideline-concordant dosing using multivariable logistic models. RESULTS: A total of 73,672 patients who were prescribed DOACS were included. Of 5837 patients who were recommended to receive low-dose therapy, 1331 (23%) received full-dose therapy instead. Of 67,935 patients recommended to receive full-dose therapy, 4079 (6%) received low-dose therapy instead. Sites varied widely on guideline discordant dosing; on inappropriate low-dose therapy, sites varied from 0 to 15%, while on inappropriate high-dose therapy, from 0 to 41%. Guideline discordant therapy decreased by about 20% in a relative sense over time, but its absolute numbers grew as DOAC therapy became more common. The most important patient-level predictors of receiving guideline-discordant therapy were older age and creatinine function being near the cutoff value. CONCLUSIONS: A substantial portion of DOAC prescriptions in the VA system are dosed contrary to clinical guidelines. This phenomenon varies widely across sites of care and has persisted over time.

Risk Factors Associated With Nonfatal Opioid Overdose Leading to Intensive Care Unit Admission: A Cross-sectional Study.

BACKGROUND: Opioid overdose (OD) and related deaths have significantly increased in the United States over the last 2 decades. Existing studies have mostly focused on demographic and clinical risk factors in noncritical care settings. Social and behavioral determinants of health (SBDH) are infrequently coded in the electronic health record (EHR) and usually buried in unstructured EHR notes, reflecting possible gaps in clinical care and observational research. Therefore, SBDH often receive less attention despite being important risk factors for OD. Natural language processing (NLP) can alleviate this problem. OBJECTIVE: The objectives of this study were two-fold: First, we examined the usefulness of NLP for SBDH extraction from unstructured EHR text, and second, for intensive care unit (ICU) admissions, we investigated risk factors including SBDH for nonfatal OD. METHODS: We performed a cross-sectional analysis of admission data from the EHR of patients in the ICU of Beth Israel Deaconess Medical Center between 2001 and 2012. We used patient admission data and International Classification of Diseases, Ninth Revision (ICD-9) diagnoses to extract demographics, nonfatal OD, SBDH, and other clinical variables. In addition to obtaining SBDH information from the ICD codes, an NLP model was developed to extract 6 SBDH variables from EHR notes, namely, housing insecurity, unemployment, social isolation, alcohol use, smoking, and illicit drug use. We adopted a sequential forward selection process to select relevant clinical variables. Multivariable logistic regression analysis was used to evaluate the associations with nonfatal OD, and relative risks were quantified as covariate-adjusted odds ratios (aOR). RESULTS: The strongest association with nonfatal OD was found to be drug use disorder (aOR 8.17, 95% CI 5.44-12.27), followed by bipolar disorder (aOR 2.69, 95% CI 1.68-4.29). Among others, major depressive disorder (aOR 2.57, 95% CI 1.12-5.88), being on a Medicaid health insurance program (aOR 2.26, 95% CI 1.43-3.58), history of illicit drug use (aOR 2.09, 95% CI 1.15-3.79), and current use of illicit drugs (aOR 2.06, 95% CI 1.20-3.55) were strongly associated with increased risk of nonfatal OD. Conversely, Blacks (aOR 0.51, 95% CI 0.28-0.94), older age groups (40-64 years: aOR 0.65, 95% CI 0.44-0.96; >64 years: aOR 0.16, 95% CI 0.08-0.34) and those with tobacco use disorder (aOR 0.53, 95% CI 0.32-0.89) or alcohol use disorder (aOR 0.64, 95% CI 0.42-1.00) had decreased risk of nonfatal OD. Moreover, 99.82% of all SBDH information was identified by the NLP model, in contrast to only 0.18% identified by the ICD codes. CONCLUSIONS: This is the first study to analyze the risk factors for nonfatal OD in an ICU setting using NLP-extracted SBDH from EHR notes. We found several risk factors associated with nonfatal OD including SBDH. SBDH are richly described in EHR notes, supporting the importance of integrating NLP-derived SBDH into OD risk assessment. More studies in ICU settings can help health care systems better understand and respond to the opioid epidemic.

Overexpression of RPN2 suppresses radiosensitivity of glioma cells by activating STAT3 signal transduction.

BACKGROUND: Radiation therapy is the primary method of treatment for glioblastoma (GBM). Therefore, the suppression of radioresistance in GBM cells is of enormous significance. Ribophorin II (RPN2), a protein component of an N-oligosaccharyl transferase complex, has been associated with chemotherapy drug resistance in multiple cancers, including GBM. However, it remains unclear whether this also plays a role in radiation therapy resistance in GBM. METHODS: We conducted a bioinformatic analysis of RPN2 expression using the UCSC Cancer Genomics Browser and GEPIA database and performed an immunohistochemical assessment of RPN2 expression in biopsy specimens from 34 GBM patients who had received radiation-based therapy. We also studied the expression and function of RPN2 in radiation-resistant GBM cells. RESULTS: We found that RPN2 expression was upregulated in GBM tumors and correlated with poor survival. The expression of RPN2 was also higher in GBM patients with tumor recurrence, who were classified to be resistant to radiation therapy. In the radiation-resistant GBM cells, the expression of RPN2 was also higher than in the parental cells. Depletion of RPN2 in resistant cells can sensitize these cells to radiation-induced apoptosis, and overexpression of RPN2 had the reverse effect. Myeloid cell leukemia 1 (MCL1) was found to be the downstream target of RPN2, and contributed to radiation resistance in GBM cells. Furthermore, STAT3 was found to be the regulator of MCL1, which can be activated by RPN2 dysregulation. CONCLUSION: Our study has revealed a novel function of RPN2 in radiation-resistant GBM, and has shown that MCL1 depletion or suppression could be a promising method of therapy to overcome the resistance promoted by RPN2 dysregulation.

Fine-Tuning Bidirectional Encoder Representations From Transformers (BERT)-Based Models on Large-Scale Electronic Health Record Notes: An Empirical Study.

BACKGROUND: The bidirectional encoder representations from transformers (BERT) model has achieved great success in many natural language processing (NLP) tasks, such as named entity recognition and question answering. However, little prior work has explored this model to be used for an important task in the biomedical and clinical domains, namely entity normalization. OBJECTIVE: We aim to investigate the effectiveness of BERT-based models for biomedical or clinical entity normalization. In addition, our second objective is to investigate whether the domains of training data influence the performances of BERT-based models as well as the degree of influence. METHODS: Our data was comprised of 1.5 million unlabeled electronic health record (EHR) notes. We first fine-tuned BioBERT on this large collection of unlabeled EHR notes. This generated our BERT-based model trained using 1.5 million electronic health record notes (EhrBERT). We then further fine-tuned EhrBERT, BioBERT, and BERT on three annotated corpora for biomedical and clinical entity normalization: the Medication, Indication, and Adverse Drug Events (MADE) 1.0 corpus, the National Center for Biotechnology Information (NCBI) disease corpus, and the Chemical-Disease Relations (CDR) corpus. We compared our models with two state-of-the-art normalization systems, namely MetaMap and disease name normalization (DNorm). RESULTS: EhrBERT achieved 40.95% F1 in the MADE 1.0 corpus for mapping named entities to the Medical Dictionary for Regulatory Activities and the Systematized Nomenclature of Medicine-Clinical Terms (SNOMED-CT), which have about 380,000 terms. In this corpus, EhrBERT outperformed MetaMap by 2.36% in F1. For the NCBI disease corpus and CDR corpus, EhrBERT also outperformed DNorm by improving the F1 scores from 88.37% and 89.92% to 90.35% and 93.82%, respectively. Compared with BioBERT and BERT, EhrBERT outperformed them on the MADE 1.0 corpus and the CDR corpus. CONCLUSIONS: Our work shows that BERT-based models have achieved state-of-the-art performance for biomedical and clinical entity normalization. BERT-based models can be readily fine-tuned to normalize any kind of named entities.

Anticoagulant Prescribing for Non-Valvular Atrial Fibrillation in the Veterans Health Administration.

Background Direct acting oral anticoagulants (DOACs) theoretically could contribute to addressing underuse of anticoagulation in non-valvular atrial fibrillation (NVAF). Few studies have examined this prospect, however. The potential of DOACs to address underuse of anticoagulation in NVAF could be magnified within a healthcare system that sharply limits patients' exposure to out-of-pocket copayments, such as the Veterans Health Administration (VA). Methods and Results We used a clinical data set of all patients with NVAF treated within VA from 2007 to 2016 (n=987 373). We examined how the proportion of patients receiving any anticoagulation, and which agent was prescribed, changed over time. When first approved for VA use in 2011, DOACs constituted a tiny proportion of all prescriptions for anticoagulants (2%); by 2016, this proportion had increased to 45% of all prescriptions and 67% of new prescriptions. Patient characteristics associated with receiving a DOAC, rather than warfarin, included white race, better kidney function, fewer comorbid conditions overall, and no history of stroke or bleeding. In 2007, before the introduction of DOACs, 56% of VA patients with NVAF were receiving anticoagulation; this dipped to 44% in 2012 just after the introduction of DOACs and had risen back to 51% by 2016. Conclusions These results do not suggest that the availability of DOACs has led to an increased proportion of patients with NVAF receiving anticoagulation, even in the context of a healthcare system that sharply limits patients' exposure to out-of-pocket copayments.

Learning to detect and understand drug discontinuation events from clinical narratives.

OBJECTIVE: Identifying drug discontinuation (DDC) events and understanding their reasons are important for medication management and drug safety surveillance. Structured data resources are often incomplete and lack reason information. In this article, we assessed the ability of natural language processing (NLP) systems to unlock DDC information from clinical narratives automatically. MATERIALS AND METHODS: We collected 1867 de-identified providers' notes from the University of Massachusetts Medical School hospital electronic health record system. Then 2 human experts chart reviewed those clinical notes to annotate DDC events and their reasons. Using the annotated data, we developed and evaluated NLP systems to automatically identify drug discontinuations and reasons at the sentence level using a novel semantic enrichment-based vector representation (SEVR) method for enhanced feature representation. RESULTS: Our SEVR-based NLP system achieved the best performance of 0.785 (AUC-ROC) for detecting discontinuation events and 0.745 (AUC-ROC) for identifying reasons when testing this highly imbalanced data, outperforming 2 state-of-the-art non-SEVR-based models. Compared with a rule-based baseline system for discontinuation detection, our system improved the sensitivity significantly (57.75% vs 18.31%, absolute value) while retaining a high specificity of 99.25%, leading to a significant improvement in AUC-ROC by 32.83% (absolute value). CONCLUSION: Experiments have shown that a high-performance NLP system can be developed to automatically identify DDCs and their reasons from providers' notes. The SEVR model effectively improved the system performance showing better generalization and robustness on unseen test data. Our work is an important step toward identifying reasons for drug discontinuation that will inform drug safety surveillance and pharmacovigilance.

Detecting Hypoglycemia Incidents Reported in Patients' Secure Messages: Using Cost-Sensitive Learning and Oversampling to Reduce Data Imbalance.

BACKGROUND: Improper dosing of medications such as insulin can cause hypoglycemic episodes, which may lead to severe morbidity or even death. Although secure messaging was designed for exchanging nonurgent messages, patients sometimes report hypoglycemia events through secure messaging. Detecting these patient-reported adverse events may help alert clinical teams and enable early corrective actions to improve patient safety. OBJECTIVE: We aimed to develop a natural language processing system, called HypoDetect (Hypoglycemia Detector), to automatically identify hypoglycemia incidents reported in patients' secure messages. METHODS: An expert in public health annotated 3000 secure message threads between patients with diabetes and US Department of Veterans Affairs clinical teams as containing patient-reported hypoglycemia incidents or not. A physician independently annotated 100 threads randomly selected from this dataset to determine interannotator agreement. We used this dataset to develop and evaluate HypoDetect. HypoDetect incorporates 3 machine learning algorithms widely used for text classification: linear support vector machines, random forest, and logistic regression. We explored different learning features, including new knowledge-driven features. Because only 114 (3.80%) messages were annotated as positive, we investigated cost-sensitive learning and oversampling methods to mitigate the challenge of imbalanced data. RESULTS: The interannotator agreement was Cohen kappa=.976. Using cross-validation, logistic regression with cost-sensitive learning achieved the best performance (area under the receiver operating characteristic curve=0.954, sensitivity=0.693, specificity 0.974, F1 score=0.590). Cost-sensitive learning and the ensembled synthetic minority oversampling technique improved the sensitivity of the baseline systems substantially (by 0.123 to 0.728 absolute gains). Our results show that a variety of features contributed to the best performance of HypoDetect. CONCLUSIONS: Despite the challenge of data imbalance, HypoDetect achieved promising results for the task of detecting hypoglycemia incidents from secure messages. The system has a great potential to facilitate early detection and treatment of hypoglycemia.

Detection of Bleeding Events in Electronic Health Record Notes Using Convolutional Neural Network Models Enhanced With Recurrent Neural Network Autoencoders: Deep Learning Approach.

BACKGROUND: Bleeding events are common and critical and may cause significant morbidity and mortality. High incidences of bleeding events are associated with cardiovascular disease in patients on anticoagulant therapy. Prompt and accurate detection of bleeding events is essential to prevent serious consequences. As bleeding events are often described in clinical notes, automatic detection of bleeding events from electronic health record (EHR) notes may improve drug-safety surveillance and pharmacovigilance. OBJECTIVE: We aimed to develop a natural language processing (NLP) system to automatically classify whether an EHR note sentence contains a bleeding event. METHODS: We expert annotated 878 EHR notes (76,577 sentences and 562,630 word-tokens) to identify bleeding events at the sentence level. This annotated corpus was used to train and validate our NLP systems. We developed an innovative hybrid convolutional neural network (CNN) and long short-term memory (LSTM) autoencoder (HCLA) model that integrates a CNN architecture with a bidirectional LSTM (BiLSTM) autoencoder model to leverage large unlabeled EHR data. RESULTS: HCLA achieved the best area under the receiver operating characteristic curve (0.957) and F1 score (0.938) to identify whether a sentence contains a bleeding event, thereby surpassing the strong baseline support vector machines and other CNN and autoencoder models. CONCLUSIONS: By incorporating a supervised CNN model and a pretrained unsupervised BiLSTM autoencoder, the HCLA achieved high performance in detecting bleeding events.

Overview of the First Natural Language Processing Challenge for Extracting Medication, Indication, and Adverse Drug Events from Electronic Health Record Notes (MADE 1.0).

INTRODUCTION: This work describes the Medication and Adverse Drug Events from Electronic Health Records (MADE 1.0) corpus and provides an overview of the MADE 1.0 2018 challenge for extracting medication, indication, and adverse drug events (ADEs) from electronic health record (EHR) notes. OBJECTIVE: The goal of MADE is to provide a set of common evaluation tasks to assess the state of the art for natural language processing (NLP) systems applied to EHRs supporting drug safety surveillance and pharmacovigilance. We also provide benchmarks on the MADE dataset using the system submissions received in the MADE 2018 challenge. METHODS: The MADE 1.0 challenge has released an expert-annotated cohort of medication and ADE information comprising 1089 fully de-identified longitudinal EHR notes from 21 randomly selected patients with cancer at the University of Massachusetts Memorial Hospital. Using this cohort as a benchmark, the MADE 1.0 challenge designed three shared NLP tasks. The named entity recognition (NER) task identifies medications and their attributes (dosage, route, duration, and frequency), indications, ADEs, and severity. The relation identification (RI) task identifies relations between the named entities: medication-indication, medication-ADE, and attribute relations. The third shared task (NER-RI) evaluates NLP models that perform the NER and RI tasks jointly. In total, 11 teams from four countries participated in at least one of the three shared tasks, and 41 system submissions were received in total. RESULTS: The best systems F1 scores for NER, RI, and NER-RI were 0.82, 0.86, and 0.61, respectively. Ensemble classifiers using the team submissions improved the performance further, with an F1 score of 0.85, 0.87, and 0.66 for the three tasks, respectively. CONCLUSION: MADE results show that recent progress in NLP has led to remarkable improvements in NER and RI tasks for the clinical domain. However, some room for improvement remains, particularly in the NER-RI task.

Automatic extraction of quantitative data from ClinicalTrials.gov to conduct meta-analyses.

OBJECTIVES: Systematic reviews and meta-analyses are labor-intensive and time-consuming. Automated extraction of quantitative data from primary studies can accelerate this process. ClinicalTrials.gov, launched in 2000, is the world's largest trial repository of results data from clinical trials; it has been used as a source instead of journal articles. We have developed a Web application called EXACT (EXtracting Accurate efficacy and safety information from ClinicalTrials.gov) that allows users without advanced programming skills to automatically extract data from ClinicalTrials.gov in analysis-ready format. We have also used the automatically extracted data to examine the reproducibility of meta-analyses in three published systematic reviews. STUDY DESIGN AND SETTING: We developed a Python-based software application (EXACT) that automatically extracts data required for meta-analysis from the ClinicalTrials.gov database in a spreadsheet format. We confirmed the accuracy of the extracted data and then used those data to repeat meta-analyses in three published systematic reviews. To ensure that we used the same statistical methods and outcomes as the published systematic reviews, we repeated the meta-analyses using data manually extracted from the relevant journal articles. For the outcomes whose results we were able to reproduce using those journal article data, we examined the usability of ClinicalTrials.gov data. RESULTS: EXACT extracted data at ClincalTrials.gov with 100% accuracy, and it required 60% less time than the usual practice of manually extracting data from journal articles. We found that 87% of the data elements extracted using EXACT matched those extracted manually from the journal articles. We were able to reproduce 24 of 28 outcomes using the journal article data. Of these 24 outcomes, we were able to reproduce 83.3% of the published estimates using data at ClinicalTrials.gov. CONCLUSION: EXACT (http://bio-nlp.org/EXACT) automatically and accurately extracted data elements from ClinicalTrials.gov and thus reduced time in data extraction. The ClinicalTrials.gov data reproduced most meta-analysis results in our study, but this conclusion needs further validation.

Extraction of Information Related to Adverse Drug Events from Electronic Health Record Notes: Design of an End-to-End Model Based on Deep Learning.

BACKGROUND: Pharmacovigilance and drug-safety surveillance are crucial for monitoring adverse drug events (ADEs), but the main ADE-reporting systems such as Food and Drug Administration Adverse Event Reporting System face challenges such as underreporting. Therefore, as complementary surveillance, data on ADEs are extracted from electronic health record (EHR) notes via natural language processing (NLP). As NLP develops, many up-to-date machine-learning techniques are introduced in this field, such as deep learning and multi-task learning (MTL). However, only a few studies have focused on employing such techniques to extract ADEs. OBJECTIVE: We aimed to design a deep learning model for extracting ADEs and related information such as medications and indications. Since extraction of ADE-related information includes two steps-named entity recognition and relation extraction-our second objective was to improve the deep learning model using multi-task learning between the two steps. METHODS: We employed the dataset from the Medication, Indication and Adverse Drug Events (MADE) 1.0 challenge to train and test our models. This dataset consists of 1089 EHR notes of cancer patients and includes 9 entity types such as Medication, Indication, and ADE and 7 types of relations between these entities. To extract information from the dataset, we proposed a deep-learning model that uses a bidirectional long short-term memory (BiLSTM) conditional random field network to recognize entities and a BiLSTM-Attention network to extract relations. To further improve the deep-learning model, we employed three typical MTL methods, namely, hard parameter sharing, parameter regularization, and task relation learning, to build three MTL models, called HardMTL, RegMTL, and LearnMTL, respectively. RESULTS: Since extraction of ADE-related information is a two-step task, the result of the second step (ie, relation extraction) was used to compare all models. We used microaveraged precision, recall, and F1 as evaluation metrics. Our deep learning model achieved state-of-the-art results (F1=65.9%), which is significantly higher than that (F1=61.7%) of the best system in the MADE1.0 challenge. HardMTL further improved the F1 by 0.8%, boosting the F1 to 66.7%, whereas RegMTL and LearnMTL failed to boost the performance. CONCLUSIONS: Deep learning models can significantly improve the performance of ADE-related information extraction. MTL may be effective for named entity recognition and relation extraction, but it depends on the methods, data, and other factors. Our results can facilitate research on ADE detection, NLP, and machine learning.