Citations

Objective: To determine whether a bi-directional relationship exists between depression and HF within a single population of individuals receiving primary care services, using longitudinal electronic health records (EHRs). Methods: This retrospective cohort study utilized EHRs for adults who received primary care services within a large healthcare system in 2006. Validated EHR-based algorithms identified 10,649 people with depression (depression cohort) and 5,911 people with HF (HF cohort) between January 1, 2006 and December 31, 2018. Each person with depression or HF was matched 1:1 with an unaffected referent on age, sex, and outpatient service use. Each cohort (with their matched referents) was followed up electronically to identify newly diagnosed HF (in the depression cohort) and depression (in the HF cohort) that occurred after the index diagnosis of depression or HF, respectively. The risks of these outcomes were compared (vs. referents) using marginal Cox proportional hazard models adjusted for 16 comorbid chronic conditions. Results: 2,024 occurrences of newly diagnosed HF were observed in the depression cohort and 944 occurrences of newly diagnosed depression were observed in the HF cohort over approximately 4–6 years of follow-up. People with depression had significantly increased risk for developing newly diagnosed HF (HR 2.08, 95% CI 1.89–2.28) and people with HF had a significantly increased risk of newly diagnosed depression (HR 1.34, 95% CI 1.17–1.54) after adjusting for all 16 comorbid chronic conditions. Conclusion: These results provide evidence of a bi-directional relationship between depression and HF independently of age, sex, and multimorbidity from chronic illnesses.

AbstractIntroductionWe sought to assess longitudinal electronic health records (EHRs) using machine learning (ML) methods to computationally derive probable Alzheimer's Disease (AD) and related dementia subphenotypes.MethodsA retrospective analysis of EHR data from a cohort of 7587 patients seen at a large, multi‐specialty urban academic medical center in New York was conducted. Subphenotypes were derived using hierarchical clustering from 792 probable AD patients (cases) who had received at least one diagnosis of AD using their clinical data. The other 6795 patients, labeled as controls, were matched on age and gender with the cases and randomly selected in the ratio of 9:1. Prediction models with multiple ML algorithms were trained on this cohort using 5‐fold cross‐validation. XGBoost was used to rank the variable importance.ResultsFour subphenotypes were computationally derived. Subphenotype A (n = 273; 28.2%) had more patients with cardiovascular diseases; subphenotype B (n = 221; 27.9%) had more patients with mental health illnesses, such as depression and anxiety; patients in subphenotype C (n = 183; 23.1%) were overall older (mean (SD) age, 79.5 (5.4) years) and had the most comorbidities including diabetes, cardiovascular diseases, and mental health disorders; and subphenotype D (n = 115; 14.5%) included patients who took anti‐dementia drugs and had sensory problems, such as deafness and hearing impairment.The 0‐year prediction model for AD risk achieved an area under the receiver operating curve (AUC) of 0.764 (SD: 0.02); the 6‐month model, 0.751 (SD: 0.02); the 1‐year model, 0.752 (SD: 0.02); the 2‐year model, 0.749 (SD: 0.03); and the 3‐year model, 0.735 (SD: 0.03), respectively. Based on variable importance, the top‐ranked comorbidities included depression, stroke/transient ischemic attack, hypertension, anxiety, mobility impairments, and atrial fibrillation. The top‐ranked medications included anti‐dementia drugs, antipsychotics, antiepileptics, and antidepressants.ConclusionsFour subphenotypes were computationally derived that correlated with cardiovascular diseases and mental health illnesses. ML algorithms based on patient demographics, diagnosis, and treatment demonstrated promising results in predicting the risk of developing AD at different time points across an individual's lifespan.

Background While there is a long history of measuring death and disability from injuries, modern research methods must account for the wide spectrum of disability that can occur in an injury, and must provide estimates with sufficient demographic, geographical and temporal detail to be useful for policy makers. The Global Burden of Disease (GBD) 2017 study used methods to provide highly detailed estimates of global injury burden that meet these criteria. Methods In this study, we report and discuss the methods used in GBD 2017 for injury morbidity and mortality burden estimation. In summary, these methods included estimating cause-specific mortality for every cause of injury, and then estimating incidence for every cause of injury. Non-fatal disability for each cause is then calculated based on the probabilities of suffering from different types of bodily injury experienced. Results GBD 2017 produced morbidity and mortality estimates for 38 causes of injury. Estimates were produced in terms of incidence, prevalence, years lived with disability, cause-specific mortality, years of life lost and disability-adjusted life-years for a 28-year period for 22 age groups, 195 countries and both sexes. Conclusions GBD 2017 demonstrated a complex and sophisticated series of analytical steps using the largest known database of morbidity and mortality data on injuries. GBD 2017 results should be used to help inform injury prevention policy making and resource allocation. We also identify important avenues for improving injury burden estimation in the future.

AbstractObjectiveTo 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 MethodsUsing 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.ResultsUsing 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.ConclusionComputationally 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.

AbstractIntroductionElectronic health record (EHR)‐driven phenotyping is a critical first step in generating biomedical knowledge from EHR data. Despite recent progress, current phenotyping approaches are manual, time‐consuming, error‐prone, and platform‐specific. This results in duplication of effort and highly variable results across systems and institutions, and is not scalable or portable. In this work, we investigate how the nascent Clinical Quality Language (CQL) can address these issues and enable high‐throughput, cross‐platform phenotyping.MethodsWe selected a clinically validated heart failure (HF) phenotype definition and translated it into CQL, then developed a CQL execution engine to integrate with the Observational Health Data Sciences and Informatics (OHDSI) platform. We executed the phenotype definition at two large academic medical centers, Northwestern Medicine and Weill Cornell Medicine, and conducted results verification (n = 100) to determine precision and recall. We additionally executed the same phenotype definition against two different data platforms, OHDSI and Fast Healthcare Interoperability Resources (FHIR), using the same underlying dataset and compared the results.ResultsCQL is expressive enough to represent the HF phenotype definition, including Boolean and aggregate operators, and temporal relationships between data elements. The language design also enabled the implementation of a custom execution engine with relative ease, and results verification at both sites revealed that precision and recall were both 100%. Cross‐platform execution resulted in identical patient cohorts generated by both data platforms.ConclusionsCQL supports the representation of arbitrarily complex phenotype definitions, and our execution engine implementation demonstrated cross‐platform execution against two widely used clinical data platforms. The language thus has the potential to help address current limitations with portability in EHR‐driven phenotyping and scale in learning health systems.

Background Past research in population health trends has shown that injuries form a substantial burden of population health loss. Regular updates to injury burden assessments are critical. We report Global Burden of Disease (GBD) 2017 Study estimates on morbidity and mortality for all injuries. Methods We reviewed results for injuries from the GBD 2017 study. GBD 2017 measured injury-specific mortality and years of life lost (YLLs) using the Cause of Death Ensemble model. To measure non-fatal injuries, GBD 2017 modelled injury-specific incidence and converted this to prevalence and years lived with disability (YLDs). YLLs and YLDs were summed to calculate disability-adjusted life years (DALYs). Findings In 1990, there were 4 260 493 (4 085 700 to 4 396 138) injury deaths, which increased to 4 484 722 (4 332 010 to 4 585 554) deaths in 2017, while age-standardised mortality decreased from 1079 (1073 to 1086) to 738 (730 to 745) per 100 000. In 1990, there were 354 064 302 (95% uncertainty interval: 338 174 876 to 371 610 802) new cases of injury globally, which increased to 520 710 288 (493 430 247 to 547 988 635) new cases in 2017. During this time, age-standardised incidence decreased non-significantly from 6824 (6534 to 7147) to 6763 (6412 to 7118) per 100 000. Between 1990 and 2017, age-standardised DALYs decreased from 4947 (4655 to 5233) per 100 000 to 3267 (3058 to 3505). Interpretation Injuries are an important cause of health loss globally, though mortality has declined between 1990 and 2017. Future research in injury burden should focus on prevention in high-burden populations, improving data collection and ensuring access to medical care.

Background The incidence of VTE in African-Americans (AAs) is similar to or higher than in Americans of European ancestry. However, the carrier frequencies of inherited thrombophilias common in whites (i.e., Factor V Leiden, Prothrombin G20210A) are very low in AAs, suggesting that other inherited thrombophilias may be associated with VTE in AAs. Objective To identify potentially novel non-hypothesis-driven single nucleotide polymorphisms (SNPs) associated with VTE in AAs. Methods We used the resources of the eMERGE Network to perform a GWAS of VTE in AAs. The eMERGE Network (funded by NHGRI) is comprised of nine sites each with DNA biobanks linked to electronic health records (EHRs). Approximately 39,206 unique DNA samples have been genotyped using either Affymetrix or Illumina genome-wide SNP arrays. Led by the Coordinating Center and eMERGE genomics workgroup, an imputation pipeline was developed for merging genomic data across the different SNP arrays used by the eMERGE sites, to maximize sample size and the power to detect associations. The imputation was performed using the 1000 Genomes Cosmopolitan reference panel which includes 1092 individuals and over 36 million SNPs. From our previously-identified cohort of Olmsted County, MN residents with incident or recurrent VTE, 1996-2005, we derived and validated an EHR-driven phenotype extraction algorithm that leveraged structured data based on ICD-9-CM codes and unstructured data from clinical notes via natural language processing to identify VTE cases and controls with 100% and 94% positive and negative predictive values, respectively. We tested for an association between each SNP and VTE among AAs using unconditional logistic regression, adjusting for age, sex and eMERGE site. Results Among 294 AA VTE cases and 3,661 AA controls (total n=3,955; females, n=2,512), the Factor V Leiden (F5 rs6025) was not analyzed due to a very low minor allele frequency (MAF=0.0036). The prothrombin G20210A (F2 rs1799963) and ABO blood type O (ABO rs8176719) SNPs were not genotyped in any of the arrays and could not be imputed. Among SNPs with an imputation score >0.8, the most significant SNPs associated with VTE were ITPR3 (inositol 1,4,5-triphosphate receptor type 3) rs2229637 (OR=1.65; p=3.61E-07; MAF=0.19) and CLEC7A (C-type lectin domain family 7, member A) rs59819090 (OR=2.16; p=1.06E-06; MAF=0.056). ITPR3 SNPs have been associated with coronary artery aneurysm in Kawasaki disease, type 1 diabetes mellitus and other autoimmune disorders. CLEC7A rs59819090 encodes for a serine82 to leucine nonsynonymous amino acid change in dectin-1. Dectin-1 is a transmembrane innate immune pattern recognition receptor on myeloid cells that upon binding it’s agonist, β-glucans, stimulates cytokine production and the respiratory burst, a prerequisite for formation of neutrophil extracelluar traps (NETs). NETs have been associated with VTE (van Montfoort, et al. Ateriosclero Thromb Vasc Biol 2013;33:147-51). Dectin-1 serine82 is not evolutionarily conserved and the serine82 to leucine amino acid change is predicted to be tolerated, with a moderate effect on protein function. Conclusions ITPR3 rs2229637 and CLAC7A rs59819090 may be associated with VTE in African-Americans. These observations require replication and functional studies toward understanding how they may lie on the causal pathway to VTE. Disclosures: No relevant conflicts of interest to declare.

Abstract Background Morbidity and mortality in tumor lysis syndrome (TLS) remains high despite increasing efforts for prevention, early detection and treatment in recent years. The current risk stratification system and treatment guidelines are largely consensus based without strong evidence. There is paucity of data on the in-hospital mortality and predictors of poor clinical outcome in this population. Methods We used the 2009-2011 Nationwide Inpatient Sample database to identify hospitalizations in patients ≥18 years with a diagnosis of TLS (International Classification of Diseases, 9th Revision, Clinical Modification [ICD-9-CM] code 277.88). Nationwide Inpatient Sample is the largest all-payer publicly available inpatient care database in the US. It contains data from five to eight million hospital stays from about 1,000 hospitals across the country and approximates a 20% sample of all US hospitals. The interval 2009-2011 was selected as ICD-9-CM code 277.88 for TLS was only introduced from the year 2009 onwards. Univariate and multivariate logistic regression were used to determine the independent predictors of in-hospital mortality. Data analysis was done using STATA version 13.0 (College Station, TX). Results Among the 997 admissions (mean age ± SD 67.58±3.33, 62.6 % males, and 80.4 % white) with TLS, in-hospital mortality was 14.44 %. Based on the results of univariate analyses (table 1), we used obesity, coronary artery disease, cardiac dysrhythmias, acute kidney injury and sepsis in the final regression model. We found that cardiac dysrhythmias (OR 4.79; 95% CI, 1.67-13.77; p=0.004) and sepsis (OR 19.70; 95% CI, 5.33-72.78; p<0.001) were independent predictors of increased in-hospital mortality in patients with TLS. Although certain myeloid and lymphoid malignancies are associated with higher risk of tumor lysis syndrome, we did not observe any statistically significant increased risk of mortality with one specific type of malignancy over the other. Similarly, none of the demographic characteristics (age <40 vs ≥40, sex or race), or hospital characteristics (rural vs urban, teaching vs non-teaching, small vs large bed-size or geographic region) predicted increased in-hospital mortality in these patients. Conclusion In this study of large national database, patients with TLS had an in-hospital mortality of 14.44 %. Independent predictors of in-hospital mortality were cardiac dysrhythmias and sepsis. Our findings might help physicians to identify sub-group of patients with TLS that are at increased risk of in-hospital mortality. Table 1 Baseline characteristics of patients with tumor lysis syndrome according to in-hospital survival Characteristics Survivors (n=853) Non-survivors (n=144) p Age( in years) Mean 68.40±3.06 67.58±3.33 0.73 Male sex 62.1 65.8 0.69 Race 0.16 White 81.52 73.54 Black 14.98 22.8 Hispanic 2.27 0 Others/unknown 1.22 3.66 Insurance status 0.41 Medicare 65.55 48.15 Medicaid 5.56 13.34 Private insurance 24.89 32.58 Self-Pay 0.65 2.99 No charge 0.63 0 Other 2.72 2.93 Region 0.13 Northeast 27.59 37.74 Midwest 28.23 12.92 South 31.34 39.9 West 12.84 9.44 Location/teaching status 0.59 Rural 7.87 10.16 Urban nonteaching 26.01 33.52 Urban teaching 66.12 56.32 Bed-size 0.84 Small 11.11 8.91 Medium 21.04 24.55 Large 67.85 66.54 Malignancies ALL 1.19 6.79 0.25 AML 2.34 6.99 0.29 CLL 27.86 16.83 0.18 CML 1.85 3.19 0.68 NHL 36.33 22.37 0.1 HD 0.53 0 0.32 Multiple Myeloma 7.39 6.56 0.82 Other hematologic malignancies 38.67 25.99 0.16 Solid tumors 35.87 26.8 0.32 Co-morbidities Smoking 15.89 21.87 0.46 Obesity 5.16 0 0.005 Dyslipidemia 27.8 16.86 0.18 Hypertension 54.17 49.19 0.63 Diabetes mellitus 27.81 16.65 0.15 PVD 1.71 3.19 0.65 CAD 13.78 3.19 0.02 AKI 60.06 79.54 0.03 CKD 28.81 25.23 0.65 Stroke 0.63 0 0.32 Sepsis 3.58 31.81 0.003 Cardiac dysrhythmias 17.96 43.42 0.04 Acute CHF 11.44 7.08 0.43 AKI=Acute Kidney Injury; ALL=Acute Lymphoblastic Leukemia; AML=Acute Myelogenous Leukemia; CAD=Coronary Artery Disease; CHF=Congestive Heart Failure; CKD=Chronic Kidney Disease; CLL=Chronic Lymphocytic Leukemia; CML=Chronic Myelogenous Leukemia; HD=Hodgkin Disease; NHL=Non Hodgkin Lymphoma; PVD=Peripheral Vascular Disease Disclosures No relevant conflicts of interest to declare.

Abstract Despite considerable progress in prevention and treatment, cancer remains the second leading cause of death in the United States. Cancer researchers around the world are generating massive amounts of clinical and genetic data, although due to its volume, complexity and lack of centralization, much is left unanalyzed. Big Data backed by powerful analytics holds the key to gain important insights from such high volume, variety and velocity data enabling a new understanding of cancer from molecular biology through clinical management. It provides opportunities to ask complex questions and identify novel knowledge from existing data including the study of genetics of an individual's cancer cells, on her response to treatment and sensitivity to side effects. In this session, we will understand the recent developments in Big Data Analytics platforms and how this transformative technology can be harnessed to leverage multidimensional data for developing new preventive measures, diagnostic tools and interventions in cancer research. We will also review success stories from early adopters of Big Data in cancer and highlight the challenges and opportunities for future research and development. Learning objectives: By the end of this session, the participants should be able to: 1.) Understand the fundamental concepts and principles of Big Data technology 2.) Identify community-based open-source Big Data informatics resources and tools 3.) Recognize the importance of Big Data in cancer prevention research Citation Format: Jyotishman Pathak. The era of big data informatics for clinical and translational research in cancer prevention. [abstract]. In: Proceedings of the Twelfth Annual AACR International Conference on Frontiers in Cancer Prevention Research; 2013 Oct 27-30; National Harbor, MD. Philadelphia (PA): AACR; Can Prev Res 2013;6(11 Suppl): Abstract nr ED03-02.