A machine learning approach to determine resilience utilizing wearable device data: analysis of an observational cohort.

Objective: To assess whether an individual's degree of psychological resilience can be determined from physiological metrics passively collected from a wearable device. Materials and Methods: Data were analyzed in this secondary analysis of the Warrior Watch Study dataset, a prospective cohort of healthcare workers enrolled across 7 hospitals in New York City. Subjects wore an Apple Watch for the duration of their participation. Surveys were collected measuring resilience, optimism, and emotional support at baseline. Results: We evaluated data from 329 subjects (mean age 37.4 years, 37.1% male). Across all testing sets, gradient-boosting machines (GBM) and extreme gradient-boosting models performed best for high- versus low-resilience prediction, stratified on a median Connor-Davidson Resilience Scale-2 score of 6 (interquartile range = 5-7), with an AUC of 0.60. When predicting resilience as a continuous variable, multivariate linear models had a correlation of 0.24 (P = .029) and RMSE of 1.37 in the testing data. A positive psychological construct, comprised of resilience, optimism, and emotional support was also evaluated. The oblique random forest method performed best in estimating high- versus low-composite scores stratified on a median of 32.5, with an AUC of 0.65, a sensitivity of 0.60, and a specificity of 0.70. Discussion: In a post hoc analysis, machine learning models applied to physiological metrics collected from wearable devices had some predictive ability in identifying resilience states and a positive psychological construct. Conclusions: These findings support the further assessment of psychological characteristics from passively collected wearable data in dedicated studies.

Implementation and evaluation of personal genetic testing as part of genomics analysis courses in German universities.

PURPOSE: Due to the increasing application of genome analysis and interpretation in medical disciplines, professionals require adequate education. Here, we present the implementation of personal genotyping as an educational tool in two genomics courses targeting Digital Health students at the Hasso Plattner Institute (HPI) and medical students at the Technical University of Munich (TUM). METHODS: We compared and evaluated the courses and the students' perceptions on the course setup using questionnaires. RESULTS: During the course, students changed their attitudes towards genotyping (HPI: 79% [15 of 19], TUM: 47% [25 of 53]). Predominantly, students became more critical of personal genotyping (HPI: 73% [11 of 15], TUM: 72% [18 of 25]) and most students stated that genetic analyses should not be allowed without genetic counseling (HPI: 79% [15 of 19], TUM: 70% [37 of 53]). Students found the personal genotyping component useful (HPI: 89% [17 of 19], TUM: 92% [49 of 53]) and recommended its inclusion in future courses (HPI: 95% [18 of 19], TUM: 98% [52 of 53]). CONCLUSION: Students perceived the personal genotyping component as valuable in the described genomics courses. The implementation described here can serve as an example for future courses in Europe.

Second-Line Biologic Therapy Following Tumor Necrosis Factor Antagonist Failure: A Real-World Propensity Score-Weighted Analysis.

BACKGROUND& AIMS: Tumor necrosis factor (TNF) antagonists often are used as first-line medications to treat moderate to severe inflammatory bowel disease (IBD), but many patients do not achieve or maintain response. Our aim was to compare the effectiveness of second-line treatments (ustekinumab, vedolizumab, or a second TNF antagonist) after TNF antagonist exposure in patients with Crohn's disease (CD) and ulcerative colitis (UC) from 2 electronic health records-based cohorts. METHODS: We identified patients with prior TNF antagonist exposure who switched to a different biologic in the Mount Sinai Health System (MSHS) electronic health records (CD, n = 527; UC, n = 165) and the Study of a Prospective Adult Research Cohort (SPARC) from the Inflammatory Bowel Disease Plexus Program of the Crohn's & Colitis Foundation (CD, n = 412; UC, n = 129). Treatment failure was defined as the composite of any IBD-related surgery, IBD-related hospitalization, new prescription of oral/intravenous corticosteroids, or need to switch to a third biologic agent. Time-to-event analysis was conducted with inverse probability of treatment-weighted data. RESULTS: Overall, treatment failure occurred in 85% of MSHS and 72% of SPARC CD patients. In SPARC, the likelihood of treatment failure was significantly lower with ustekinumab compared with vedolizumab as second-line treatment (adjusted hazard ratio, 0.66; 95% CI, 0.54-0.82; P < .001), a trend confirmed in MSHS (adjusted hazard ratio, 0.89; 95% CI, 0.77-1.04; P = .15). In both cohorts, the superiority of ustekinumab compared with vedolizumab was shown when considering treatment failure as prescription of steroids or a third biologic agent. In UC, no differences between second-line treatment groups were identified. CONCLUSIONS: In 2 independent real-world cohort settings, second-line therapy in CD with ustekinumab after TNF antagonist treatment failure was associated with a lower likelihood of treatment failure than second-line vedolizumab.

Prescriber Adoption of SLCO1B1 Genotype-Guided Simvastatin Clinical Decision Support in a Clinical Pharmacogenetics Program.

Pharmacogenetic implementation programs are increasingly feasible due to the availability of clinical guidelines for implementation research. The utilization of these resources has been reported with selected drug-gene pairs; however, little is known about how prescribers respond to pharmacogenetic recommendations for statin therapy. We prospectively assessed prescriber interaction with point-of-care clinical decision support (CDS) to guide simvastatin therapy for a diverse cohort of primary care patients enrolled in a clinical pharmacogenetics program. Of the 1,639 preemptively genotyped patients, 298 (18.2%) had an intermediate function (IF) OATP1B1 phenotype and 25 (1.53%) had a poor function (PF) phenotype, predicted by a common single nucleotide variant in the SLCO1B1 gene (c.521T>C; rs4149056). Clinicians were presented with CDS when simvastatin was prescribed for patients with IF or PF through the electronic health record. Importantly, 64.2% of the CDS deployed at the point-of-care was accepted by the prescribers and resulted in prescription changes. Statin intensity was found to significantly influence prescriber adoption of the pharmacogenetic-guided CDS, whereas patient gender or race, prescriber type, or pharmacogenetic training status did not significantly influence adoption. This study demonstrates that primary care providers readily adopt pharmacogenetic information to guide statin therapy for the majority of patients with preemptive genotype data.

Mendelian randomization indicates causal effects of estradiol levels on kidney function in males.

Context: Chronic kidney disease (CKD) is a public health burden worldwide. Epidemiological studies observed an association between sex hormones, including estradiol, and kidney function. Objective: We conducted a Mendelian randomization (MR) study to assess a possible causal effect of estradiol levels on kidney function in males and females. Design: We performed a bidirectional two-sample MR using published genetic associations of serum levels of estradiol in men (n = 206,927) and women (n = 229,966), and of kidney traits represented by estimated glomerular filtration rate (eGFR, n = 567,460), urine albumin-to-creatinine ratio (UACR, n = 547,361), and CKD (n = 41,395 cases and n = 439,303 controls) using data obtained from the CKDGen Consortium. Additionally, we conducted a genome-wide association study using UK Biobank cohort study data (n = 11,798 men and n = 6,835 women) to identify novel genetic associations with levels of estradiol, and then used these variants as instruments in a one-sample MR. Results: The two-sample MR indicated that genetically predicted estradiol levels are significantly associated with eGFR in men (beta = 0.077; p = 5.2E-05). We identified a single locus at chromosome 14 associated with estradiol levels in men being significant in the one-sample MR on eGFR (beta = 0.199; p = 0.017). We revealed significant results with eGFR in postmenopausal women and with UACR in premenopausal women, which did not reach statistical significance in the sensitivity MR analyses. No causal effect of eGFR or UACR on estradiol levels was found. Conclusions: We conclude that serum estradiol levels may have a causal effect on kidney function. Our MR results provide starting points for studies to develop therapeutic strategies to reduce kidney disease.

Predicting hypertension onset from longitudinal electronic health records with deep learning.

Objective: Hypertension has long been recognized as one of the most important predisposing factors for cardiovascular diseases and mortality. In recent years, machine learning methods have shown potential in diagnostic and predictive approaches in chronic diseases. Electronic health records (EHRs) have emerged as a reliable source of longitudinal data. The aim of this study is to predict the onset of hypertension using modern deep learning (DL) architectures, specifically long short-term memory (LSTM) networks, and longitudinal EHRs. Materials and Methods: We compare this approach to the best performing models reported from previous works, particularly XGboost, applied to aggregated features. Our work is based on data from 233 895 adult patients from a large health system in the United States. We divided our population into 2 distinct longitudinal datasets based on the diagnosis date. To ensure generalization to unseen data, we trained our models on the first dataset (dataset A "train and validation") using cross-validation, and then applied the models to a second dataset (dataset B "test") to assess their performance. We also experimented with 2 different time-windows before the onset of hypertension and evaluated the impact on model performance. Results: With the LSTM network, we were able to achieve an area under the receiver operating characteristic curve value of 0.98 in the "train and validation" dataset A and 0.94 in the "test" dataset B for a prediction time window of 1 year. Lipid disorders, type 2 diabetes, and renal disorders are found to be associated with incident hypertension. Conclusion: These findings show that DL models based on temporal EHR data can improve the identification of patients at high risk of hypertension and corresponding driving factors. In the long term, this work may support identifying individuals who are at high risk for developing hypertension and facilitate earlier intervention to prevent the future development of hypertension.

Quantifying the phenome-wide disease burden of obesity using electronic health records and genomics.

OBJECTIVE: High BMI is associated with many comorbidities and mortality. This study aimed to elucidate the overall clinical risk of obesity using a genome- and phenome-wide approach. METHODS: This study performed a phenome-wide association study of BMI using a clinical cohort of 736,726 adults. This was followed by genetic association studies using two separate cohorts: one consisting of 65,174 adults in the Electronic Medical Records and Genomics (eMERGE) Network and another with 405,432 participants in the UK Biobank. RESULTS: Class 3 obesity was associated with 433 phenotypes, representing 59.3% of all billing codes in individuals with severe obesity. A genome-wide polygenic risk score for BMI, accounting for 7.5% of variance in BMI, was associated with 296 clinical diseases, including strong associations with type 2 diabetes, sleep apnea, hypertension, and chronic liver disease. In all three cohorts, 199 phenotypes were associated with class 3 obesity and polygenic risk for obesity, including novel associations such as increased risk of renal failure, venous insufficiency, and gastroesophageal reflux. CONCLUSIONS: This combined genomic and phenomic systematic approach demonstrated that obesity has a strong genetic predisposition and is associated with a considerable burden of disease across all disease classes.

Insights From a Large-Scale Whole-Genome Sequencing Study of Systolic Blood Pressure, Diastolic Blood Pressure, and Hypertension.

BACKGROUND: The availability of whole-genome sequencing data in large studies has enabled the assessment of coding and noncoding variants across the allele frequency spectrum for their associations with blood pressure. METHODS: We conducted a multiancestry whole-genome sequencing analysis of blood pressure among 51 456 Trans-Omics for Precision Medicine and Centers for Common Disease Genomics program participants (stage-1). Stage-2 analyses leveraged array data from UK Biobank (N=383 145), Million Veteran Program (N=318 891), and Reasons for Geographic and Racial Differences in Stroke (N=10 643) participants, along with whole-exome sequencing data from UK Biobank (N=199 631) participants. RESULTS: Two blood pressure signals achieved genome-wide significance in meta-analyses of stage-1 and stage-2 single variant findings (P<5×10-8). Among them, a rare intergenic variant at novel locus, LOC100506274, was associated with lower systolic blood pressure in stage-1 (beta [SE]=-32.6 [6.0]; P=4.99×10-8) but not stage-2 analysis (P=0.11). Furthermore, a novel common variant at the known INSR locus was suggestively associated with diastolic blood pressure in stage-1 (beta [SE]=-0.36 [0.07]; P=4.18×10-7) and attained genome-wide significance in stage-2 (beta [SE]=-0.29 [0.03]; P=7.28×10-23). Nineteen additional signals suggestively associated with blood pressure in meta-analysis of single and aggregate rare variant findings (P<1×10-6 and P<1×10-4, respectively). DISCUSSION: We report one promising but unconfirmed rare variant for blood pressure and, more importantly, contribute insights for future blood pressure sequencing studies. Our findings suggest promise of aggregate analyses to complement single variant analysis strategies and the need for larger, diverse samples, and family studies to enable robust rare variant identification.

Evaluation of a machine learning approach utilizing wearable data for prediction of SARS-CoV-2 infection in healthcare workers.

Objective: To determine whether a machine learning model can detect SARS-CoV-2 infection from physiological metrics collected from wearable devices. Materials and Methods: Health care workers from 7 hospitals were enrolled and prospectively followed in a multicenter observational study. Subjects downloaded a custom smart phone app and wore Apple Watches for the duration of the study period. Daily surveys related to symptoms and the diagnosis of Coronavirus Disease 2019 were answered in the app. Results: We enrolled 407 participants with 49 (12%) having a positive nasal SARS-CoV-2 polymerase chain reaction test during follow-up. We examined 5 machine-learning approaches and found that gradient-boosting machines (GBM) had the most favorable validation performance. Across all testing sets, our GBM model predicted SARS-CoV-2 infection with an average area under the receiver operating characteristic (auROC) = 86.4% (confidence interval [CI] 84-89%). The model was calibrated to value sensitivity over specificity, achieving an average sensitivity of 82% (CI ±âˆ¼4%) and specificity of 77% (CI ±âˆ¼1%). The most important predictors included parameters describing the circadian heart rate variability mean (MESOR) and peak-timing (acrophase), and age. Discussion: We show that a tree-based ML algorithm applied to physiological metrics passively collected from a wearable device can identify and predict SARS-CoV-2 infection. Conclusion: Applying machine learning models to the passively collected physiological metrics from wearable devices may improve SARS-CoV-2 screening methods and infection tracking.

Genetic loci and prioritization of genes for kidney function decline derived from a meta-analysis of 62 longitudinal genome-wide association studies.

Estimated glomerular filtration rate (eGFR) reflects kidney function. Progressive eGFR-decline can lead to kidney failure, necessitating dialysis or transplantation. Hundreds of loci from genome-wide association studies (GWAS) for eGFR help explain population cross section variability. Since the contribution of these or other loci to eGFR-decline remains largely unknown, we derived GWAS for annual eGFR-decline and meta-analyzed 62 longitudinal studies with eGFR assessed twice over time in all 343,339 individuals and in high-risk groups. We also explored different covariate adjustment. Twelve genome-wide significant independent variants for eGFR-decline unadjusted or adjusted for eGFR-baseline (11 novel, one known for this phenotype), including nine variants robustly associated across models were identified. All loci for eGFR-decline were known for cross-sectional eGFR and thus distinguished a subgroup of eGFR loci. Seven of the nine variants showed variant-by-age interaction on eGFR cross section (further about 350,000 individuals), which linked genetic associations for eGFR-decline with age-dependency of genetic cross-section associations. Clinically important were two to four-fold greater genetic effects on eGFR-decline in high-risk subgroups. Five variants associated also with chronic kidney disease progression mapped to genes with functional in-silico evidence (UMOD, SPATA7, GALNTL5, TPPP). An unfavorable versus favorable nine-variant genetic profile showed increased risk odds ratios of 1.35 for kidney failure (95% confidence intervals 1.03-1.77) and 1.27 for acute kidney injury (95% confidence intervals 1.08-1.50) in over 2000 cases each, with matched controls). Thus, we provide a large data resource, genetic loci, and prioritized genes for kidney function decline, which help inform drug development pipelines revealing important insights into the age-dependency of kidney function genetics.

Effects of Testing and Disclosing Ancestry-Specific Genetic Risk for Kidney Failure on Patients and Health Care Professionals: A Randomized Clinical Trial.

IMPORTANCE: Risk variants in the apolipoprotein L1 (APOL1 [OMIM 603743]) gene on chromosome 22 are common in individuals of West African ancestry and confer increased risk of kidney failure for people with African ancestry and hypertension. Whether disclosing APOL1 genetic testing results to patients of African ancestry and their clinicians affects blood pressure, kidney disease screening, or patient behaviors is unknown. OBJECTIVE: To determine the effects of testing and disclosing APOL1 genetic results to patients of African ancestry with hypertension and their clinicians. DESIGN, SETTING, AND PARTICIPANTS: This pragmatic randomized clinical trial randomly assigned 2050 adults of African ancestry with hypertension and without existing chronic kidney disease in 2 US health care systems from November 1, 2014, through November 28, 2016; the final date of follow-up was January 16, 2018. Patients were randomly assigned to undergo immediate (intervention) or delayed (waiting list control group) APOL1 testing in a 7:1 ratio. Statistical analysis was performed from May 1, 2018, to July 31, 2020. INTERVENTIONS: Patients randomly assigned to the intervention group received APOL1 genetic testing results from trained staff; their clinicians received results through clinical decision support in electronic health records. Waiting list control patients received the results after their 12-month follow-up visit. MAIN OUTCOMES AND MEASURES: Coprimary outcomes were the change in 3-month systolic blood pressure and 12-month urine kidney disease screening comparing intervention patients with high-risk APOL1 genotypes and those with low-risk APOL1 genotypes. Secondary outcomes compared these outcomes between intervention group patients with high-risk APOL1 genotypes and controls. Exploratory analyses included psychobehavioral factors. RESULTS: Among 2050 randomly assigned patients (1360 women [66%]; mean [SD] age, 53 [10] years), the baseline mean (SD) systolic blood pressure was significantly higher in patients with high-risk APOL1 genotypes vs those with low-risk APOL1 genotypes and controls (137 [21] vs 134 [19] vs 133 [19] mm Hg; P = .003 for high-risk vs low-risk APOL1 genotypes; P = .001 for high-risk APOL1 genotypes vs controls). At 3 months, the mean (SD) change in systolic blood pressure was significantly greater in patients with high-risk APOL1 genotypes vs those with low-risk APOL1 genotypes (6 [18] vs 3 [18] mm Hg; P = .004) and controls (6 [18] vs 3 [19] mm Hg; P = .01). At 12 months, there was a 12% increase in urine kidney disease testing among patients with high-risk APOL1 genotypes (from 39 of 234 [17%] to 68 of 234 [29%]) vs a 6% increase among those with low-risk APOL1 genotypes (from 278 of 1561 [18%] to 377 of 1561 [24%]; P = .10) and a 7% increase among controls (from 33 of 255 [13%] to 50 of 255 [20%]; P = .01). In response to testing, patients with high-risk APOL1 genotypes reported more changes in lifestyle (a subjective measure that included better dietary and exercise habits; 129 of 218 [59%] vs 547 of 1468 [37%]; P < .001) and increased blood pressure medication use (21 of 218 [10%] vs 68 of 1468 [5%]; P = .005) vs those with low-risk APOL1 genotypes; 1631 of 1686 (97%) declared they would get tested again. CONCLUSIONS AND RELEVANCE: In this randomized clinical trial, disclosing APOL1 genetic testing results to patients of African ancestry with hypertension and their clinicians was associated with a greater reduction in systolic blood pressure, increased kidney disease screening, and positive self-reported behavior changes in those with high-risk genotypes. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT02234063.

Type 2 Diabetes Partitioned Polygenic Scores Associate With Disease Outcomes in 454,193 Individuals Across 13 Cohorts.

OBJECTIVE: Type 2 diabetes (T2D) has heterogeneous patient clinical characteristics and outcomes. In previous work, we investigated the genetic basis of this heterogeneity by clustering 94 T2D genetic loci using their associations with 47 diabetes-related traits and identified five clusters, termed ß-cell, proinsulin, obesity, lipodystrophy, and liver/lipid. The relationship between these clusters and individual-level metabolic disease outcomes has not been assessed. RESEARCH DESIGN AND METHODS: Here we constructed individual-level partitioned polygenic scores (pPS) for these five clusters in 12 studies from the Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE) consortium and the UK Biobank (n = 454,193) and tested for cross-sectional association with T2D-related outcomes, including blood pressure, renal function, insulin use, age at T2D diagnosis, and coronary artery disease (CAD). RESULTS: Despite all clusters containing T2D risk-increasing alleles, they had differential associations with metabolic outcomes. Increased obesity and lipodystrophy cluster pPS, which had opposite directions of association with measures of adiposity, were both significantly associated with increased blood pressure and hypertension. The lipodystrophy and liver/lipid cluster pPS were each associated with CAD, with increasing and decreasing effects, respectively. An increased liver/lipid cluster pPS was also significantly associated with reduced renal function. The liver/lipid cluster includes known loci linked to liver lipid metabolism (e.g., GCKR, PNPLA3, and TM6SF2), and these findings suggest that cardiovascular disease risk and renal function may be impacted by these loci through their shared disease pathway. CONCLUSIONS: Our findings support that genetically driven pathways leading to T2D also predispose differentially to clinical outcomes.

Interaction-Based Feature Selection Algorithm Outperforms Polygenic Risk Score in Predicting Parkinson's Disease Status.

Polygenic risk scores (PRS) aggregating results from genome-wide association studies are the state of the art in the prediction of susceptibility to complex traits or diseases, yet their predictive performance is limited for various reasons, not least of which is their failure to incorporate the effects of gene-gene interactions. Novel machine learning algorithms that use large amounts of data promise to find gene-gene interactions in order to build models with better predictive performance than PRS. Here, we present a data preprocessing step by using data-mining of contextual information to reduce the number of features, enabling machine learning algorithms to identify gene-gene interactions. We applied our approach to the Parkinson's Progression Markers Initiative (PPMI) dataset, an observational clinical study of 471 genotyped subjects (368 cases and 152 controls). With an AUC of 0.85 (95% CI = [0.72; 0.96]), the interaction-based prediction model outperforms the PRS (AUC of 0.58 (95% CI = [0.42; 0.81])). Furthermore, feature importance analysis of the model provided insights into the mechanism of Parkinson's disease. For instance, the model revealed an interaction of previously described drug target candidate genes TMEM175 and GAPDHP25. These results demonstrate that interaction-based machine learning models can improve genetic prediction models and might provide an answer to the missing heritability problem.

Phe2vec: Automated disease phenotyping based on unsupervised embeddings from electronic health records.

Robust phenotyping of patients from electronic health records (EHRs) at scale is a challenge in clinical informatics. Here, we introduce Phe2vec, an automated framework for disease phenotyping from EHRs based on unsupervised learning and assess its effectiveness against standard rule-based algorithms from Phenotype KnowledgeBase (PheKB). Phe2vec is based on pre-computing embeddings of medical concepts and patients' clinical history. Disease phenotypes are then derived from a seed concept and its neighbors in the embedding space. Patients are linked to a disease if their embedded representation is close to the disease phenotype. Comparing Phe2vec and PheKB cohorts head-to-head using chart review, Phe2vec performed on par or better in nine out of ten diseases. Differently from other approaches, it can scale to any condition and was validated against widely adopted expert-based standards. Phe2vec aims to optimize clinical informatics research by augmenting current frameworks to characterize patients by condition and derive reliable disease cohorts.

Factors Associated With Longitudinal Psychological and Physiological Stress in Health Care Workers During the COVID-19 Pandemic: Observational Study Using Apple Watch Data.

BACKGROUND: The COVID-19 pandemic has resulted in a high degree of psychological distress among health care workers (HCWs). There is a need to characterize which HCWs are at an increased risk of developing psychological effects from the pandemic. Given the differences in the response of individuals to stress, an analysis of both the perceived and physiological consequences of stressors can provide a comprehensive evaluation of its impact. OBJECTIVE: This study aimed to determine characteristics associated with longitudinal perceived stress in HCWs and to assess whether changes in heart rate variability (HRV), a marker of autonomic nervous system function, are associated with features protective against longitudinal stress. METHODS: HCWs across 7 hospitals in New York City, NY, were prospectively followed in an ongoing observational digital study using the custom Warrior Watch Study app. Participants wore an Apple Watch for the duration of the study to measure HRV throughout the follow-up period. Surveys measuring perceived stress, resilience, emotional support, quality of life, and optimism were collected at baseline and longitudinally. RESULTS: A total of 361 participants (mean age 36.8, SD 10.1 years; female: n=246, 69.3%) were enrolled. Multivariate analysis found New York City's COVID-19 case count to be associated with increased longitudinal stress (P=.008). Baseline emotional support, quality of life, and resilience were associated with decreased longitudinal stress (P<.001). A significant reduction in stress during the 4-week period after COVID-19 diagnosis was observed in the highest tertial of emotional support (P=.03) and resilience (P=.006). Participants in the highest tertial of baseline emotional support and resilience had a significantly different circadian pattern of longitudinally collected HRV compared to subjects in the low or medium tertial. CONCLUSIONS: High resilience, emotional support, and quality of life place HCWs at reduced risk of longitudinal perceived stress and have a distinct physiological stress profile. Our findings support the use of these characteristics to identify HCWs at risk of the psychological and physiological stress effects of the pandemic.

Predictive Approaches for Acute Dialysis Requirement and Death in COVID-19.

BACKGROUND AND OBJECTIVES: AKI treated with dialysis initiation is a common complication of coronavirus disease 2019 (COVID-19) among hospitalized patients. However, dialysis supplies and personnel are often limited. DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS: Using data from adult patients hospitalized with COVID-19 from five hospitals from the Mount Sinai Health System who were admitted between March 10 and December 26, 2020, we developed and validated several models (logistic regression, Least Absolute Shrinkage and Selection Operator (LASSO), random forest, and eXtreme GradientBoosting [XGBoost; with and without imputation]) for predicting treatment with dialysis or death at various time horizons (1, 3, 5, and 7 days) after hospital admission. Patients admitted to the Mount Sinai Hospital were used for internal validation, whereas the other hospitals formed part of the external validation cohort. Features included demographics, comorbidities, and laboratory and vital signs within 12 hours of hospital admission. RESULTS: A total of 6093 patients (2442 in training and 3651 in external validation) were included in the final cohort. Of the different modeling approaches used, XGBoost without imputation had the highest area under the receiver operating characteristic (AUROC) curve on internal validation (range of 0.93-0.98) and area under the precision-recall curve (AUPRC; range of 0.78-0.82) for all time points. XGBoost without imputation also had the highest test parameters on external validation (AUROC range of 0.85-0.87, and AUPRC range of 0.27-0.54) across all time windows. XGBoost without imputation outperformed all models with higher precision and recall (mean difference in AUROC of 0.04; mean difference in AUPRC of 0.15). Features of creatinine, BUN, and red cell distribution width were major drivers of the model's prediction. CONCLUSIONS: An XGBoost model without imputation for prediction of a composite outcome of either death or dialysis in patients positive for COVID-19 had the best performance, as compared with standard and other machine learning models. PODCAST: This article contains a podcast at https://www.asn-online.org/media/podcast/CJASN/2021_07_09_CJN17311120.mp3.

Toward a fine-scale population health monitoring system.

Understanding population health disparities is an essential component of equitable precision health efforts. Epidemiology research often relies on definitions of race and ethnicity, but these population labels may not adequately capture disease burdens and environmental factors impacting specific sub-populations. Here, we propose a framework for repurposing data from electronic health records (EHRs) in concert with genomic data to explore the demographic ties that can impact disease burdens. Using data from a diverse biobank in New York City, we identified 17 communities sharing recent genetic ancestry. We observed 1,177 health outcomes that were statistically associated with a specific group and demonstrated significant differences in the segregation of genetic variants contributing to Mendelian diseases. We also demonstrated that fine-scale population structure can impact the prediction of complex disease risk within groups. This work reinforces the utility of linking genomic data to EHRs and provides a framework toward fine-scale monitoring of population health.

Discovery and fine-mapping of height loci via high-density imputation of GWASs in individuals of African ancestry.

Although many loci have been associated with height in European ancestry populations, very few have been identified in African ancestry individuals. Furthermore, many of the known loci have yet to be generalized to and fine-mapped within a large-scale African ancestry sample. We performed sex-combined and sex-stratified meta-analyses in up to 52,764 individuals with height and genome-wide genotyping data from the African Ancestry Anthropometry Genetics Consortium (AAAGC). We additionally combined our African ancestry meta-analysis results with published European genome-wide association study (GWAS) data. In the African ancestry analyses, we identified three novel loci (SLC4A3, NCOA2, ECD/FAM149B1) in sex-combined results and two loci (CRB1, KLF6) in women only. In the African plus European sex-combined GWAS, we identified an additional three novel loci (RCCD1, G6PC3, CEP95) which were equally driven by AAAGC and European results. Among 39 genome-wide significant signals at known loci, conditioning index SNPs from European studies identified 20 secondary signals. Two of the 20 new secondary signals and none of the 8 novel loci had minor allele frequencies (MAF) < 5%. Of 802 known European height signals, 643 displayed directionally consistent associations with height, of which 205 were nominally significant (p < 0.05) in the African ancestry sex-combined sample. Furthermore, 148 of 241 loci contained ≤20 variants in the credible sets that jointly account for 99% of the posterior probability of driving the associations. In summary, trans-ethnic meta-analyses revealed novel signals and further improved fine-mapping of putative causal variants in loci shared between African and European ancestry populations.

Association of SARS-CoV-2 viral load at admission with in-hospital acute kidney injury: A retrospective cohort study.

BACKGROUND: Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and the associated Coronavirus Disease 2019 (COVID-19) is a public health emergency. Acute kidney injury (AKI) is a common complication in hospitalized patients with COVID-19 although mechanisms underlying AKI are yet unclear. There may be a direct effect of SARS-CoV-2 virus on the kidney; however, there is currently no data linking SARS-CoV-2 viral load (VL) to AKI. We explored the association of SARS-CoV-2 VL at admission to AKI in a large diverse cohort of hospitalized patients with COVID-19. METHODS AND FINDINGS: We included patients hospitalized between March 13th and May 19th, 2020 with SARS-CoV-2 in a large academic healthcare system in New York City (N = 1,049) with available VL at admission quantified by real-time RT-PCR. We extracted clinical and outcome data from our institutional electronic health records (EHRs). AKI was defined by KDIGO guidelines. We fit a Fine-Gray competing risks model (with death as a competing risk) using demographics, comorbidities, admission severity scores, and log10 transformed VL as covariates and generated adjusted hazard ratios (aHR) and 95% Confidence Intervals (CIs). VL was associated with an increased risk of AKI (aHR = 1.04, 95% CI: 1.01-1.08, p = 0.02) with a 4% increased hazard for each log10 VL change. Patients with a viral load in the top 50th percentile had an increased adjusted hazard of 1.27 (95% CI: 1.02-1.58, p = 0.03) for AKI as compared to those in the bottom 50th percentile. CONCLUSIONS: VL is weakly but significantly associated with in-hospital AKI after adjusting for confounders. This may indicate the role of VL in COVID-19 associated AKI. This data may inform future studies to discover the mechanistic basis of COVID-19 associated AKI.