The NYCKidSeq randomized controlled trial: Impact of GUÍA digitally enhanced genetic results disclosure in diverse families.

Digital solutions are needed to support rapid increases in the application of genetic/genomic tests (GTs) in diverse clinical settings and patient populations. We developed GUÍA, a bilingual digital application that facilitates disclosure of GT results. The NYCKidSeq randomized controlled trial enrolled diverse children with neurologic, cardiac, and immunologic conditions who underwent GTs. The trial evaluated GUÍA's impact on understanding the GT results by randomizing families to results disclosure genetic counseling with GUÍA (intervention) or standard of care (SOC). Parents/legal guardians (participants) completed surveys at baseline, post-results disclosure, and 6 months later. Survey measures assessed the primary study outcomes of participants' perceived understanding of and confidence in explaining their child's GT results and the secondary outcome of objective understanding. The analysis included 551 diverse participants, 270 in the GUÍA arm and 281 in SOC. Participants in the GUÍA arm had significantly higher perceived understanding post-results (OR = 2.8, CI[1.004, 7.617], p = 0.049) and maintained higher objective understanding over time (OR = 1.1, CI[1.004, 1.127], p = 0.038) compared to SOC. There was no impact on perceived confidence. Hispanic/Latino(a) individuals in the GUÍA arm maintained higher perceived understanding (OR = 3.9, CI[1.603, 9.254], p = 0.003), confidence (OR = 2.7, CI[1.021, 7.277], p = 0.046), and objective understanding (OR = 1.1, CI[1.009, 1.212], p = 0.032) compared to SOC. This trial demonstrates that GUÍA positively impacts understanding of GT results in diverse parents of children with suspected genetic conditions and builds a case for utilizing GUÍA to deliver complex results. Continued development and evaluation of digital applications in diverse populations are critical for equitably scaling GT offerings in specialty clinics.

Physician services and costs after disclosure of diagnostic sequencing results in the NYCKidSeq program.

PURPOSE: To better understand the effects of returning diagnostic sequencing results on clinical actions and economic outcomes for pediatric patients with suspected genetic disorders. METHODS: Longitudinal physician claims data after diagnostic sequencing were obtained for patients aged 0 to 21 years with neurologic, cardiac, and immunologic disorders with suspected genetic etiology. We assessed specialist consultation rates prompted by primary diagnostic results, as well as marginal effects on overall 18-month physician services and costs. RESULTS: We included data on 857 patients (median age: 9.6 years) with a median follow-up of 17.3 months after disclosure of diagnostic sequencing results. The likelihood of having ≥1 recommendation for specialist consultation in 155 patients with positive findings was high (72%) vs 23% in 443 patients with uncertain findings and 21% in 259 patients with negative findings (P < .001). Follow-through consultation occurred in 30%. Increases in 18-month physician services and costs following a positive finding diminished after multivariable adjustment. Also, no significant differences between those with uncertain and negative findings were demonstrated. CONCLUSION: Our study did not provide evidence for significant increases in downstream physician services and costs after returning positive or uncertain diagnostic sequencing findings. More large-scale longitudinal studies are needed to confirm these findings.

Physician and informal care use explained by the Pediatric Quality of Life Inventory (PedsQL) in children with suspected genetic disorders.

PURPOSE: To examine associations between Pediatric Quality of Life Inventory (PedsQL) 4.0 Generic Core Scales and PedsQL Infant Scales with formal health care resource utilization (HCRU) and informal caregiver burden. METHODS: We studied a pediatric cohort of 837 patients (median age: 8.4 years) with suspected genetic disorders enrolled January 2019 through July 2021 in the NYCKidSeq program for diagnostic sequencing. Using linked ~ nine-month longitudinal survey and physician claims data collected through May 2022, we modeled the association between baseline PedsQL scores and post-baseline HCRU (median follow-up: 21.1 months) and informal care. We also assessed the longitudinal change in PedsQL scores with physician services using linear mixed-effects models. RESULTS: Lower PedsQL total and physical health scores were independently associated with increases in 18-month physician services, encounters, and weekly informal care. Comparing low vs. median total scores, increases were 10.6 services (95% CI: 1.0-24.6), 3.3 encounters (95% CI: 0.5-6.8), and $668 (95% CI: $350-965), respectively. For the psychosocial domain, higher scores were associated with decreased informal care. Based on adjusted linear mixed-effects modeling, every additional ten physician services was associated with diminished improvement in longitudinal PedsQL total score trajectories by 1.1 point (95% confidence interval: 0.6-1.6) on average. Similar trends were observed in the physical and psychosocial domains. CONCLUSION: PedsQL scores were independently associated with higher utilization of physician services and informal care. Moreover, longitudinal trajectories of PedsQL scores became less favorable with increased physician services. Adding PedsQL survey instruments to conventional measures for improved risk stratification should be evaluated in further research.

The Pediatric Quality of Life Inventory (PedsQL) is widely used to measure health-related quality of life in pediatric patients; however, few studies have examined whether the PedsQL is indicative of longitudinal outcomes of morbidity and health care needs. This study captures associations between PedsQL scores with utilization of physician and informal care in children with suspected genetic disorders. We demonstrate that lower PedsQL total and physical health scores are independently associated with greater utilization of physician services and informal care. Moreover, longitudinal trajectories of PedsQL scores become less favorable with increased physician services. Results can inform future applications of PedsQL instruments.

Trends and disparities in prevalence of cardiometabolic diseases by food security status in the United States.

BACKGROUND: Previous studies have demonstrated the association between food security and cardiometabolic diseases (CMDs), yet none have investigated trends in prevalence of CMDs by food security status in the United States (US). METHODS: Serial cross-sectional analysis of the US nationally representative data from National Health and Nutrition Examination Survey (1999-2018) was conducted among adults aged 20 years or older. Food security status was defined by the US Household Food Security Survey Module (full, marginal, low, and very low food security). We estimated the age-adjusted prevalence of CMDs including obesity, hypertension, diabetes, and coronary heart disease by food security status. Racial and ethnic disparities in age-adjusted prevalence of CMDs by food security status were also assessed. RESULTS: A total of 49,738 participants were included in this analysis (weighted mean age 47.3 years; 51.3% women). From 1999 to 2018, the age-adjusted prevalence of CMDs was lower in full food secure group as compared with other groups. For example, trends in hypertension decreased from 49.7% (47.5-51.8%) to 45.9% (43.8-48.0%) (P-trend = 0.002) among the full and from 54.2% (49.9-58.5%) to 49.7% (46.8-52.6%) (P-trend = 0.02) among the marginal but remained stable among the low at 49.7% (47.9-51.6%) and among the very low at 51.1% (48.9-53.3%) (P-interaction = 0.02). Prevalence of diabetes increased from 8.85% (8.15-9.60%) to 12.2% (11.1-13.5%) among the full (P-trend < 0.001), from 16.5% (13.2-20.4%) to 20.9% (18.6-23.5%) (P-trend = 0.045) among the marginal and from 14.6% (11.1-19.0%) to 20.9% (18.8-23.3%) (P-trend = 0.001) among the low but remained stable at 18.8% (17.0-20.9) among the very low (P-trend = 0.35) (P-interaction = 0.03). Racial and ethnic differences in prevalence of CMD by food security status were observed. For example, among individuals with full food secure status, the prevalence of diabetes was 9.08% (95% CI, 8.60-9.59%) for non-Hispanic whites, 17.3% (95% CI, 16.4-18.2%) for non-Hispanic blacks, 16.1% (95% CI, 15.0-17.4%) for Hispanics and 14.9% (95% CI, 13.3-16.7%) for others. CONCLUSIONS AND RELEVANCE: Prevalence of CMDs was greatest among those experiencing food insecurity, and food insecurity disproportionately affected racial/ethnic minorities. Disparities in CMD prevalence by food security status persisted or worsened, especially among racial/ethnic minorities.

Molecular diagnostic yield of genome sequencing versus targeted gene panel testing in racially and ethnically diverse pediatric patients.

PURPOSE: Adoption of genome sequencing (GS) as a first-line test requires evaluation of its diagnostic yield. We evaluated the GS and targeted gene panel (TGP) testing in diverse pediatric patients (probands) with suspected genetic conditions. METHODS: Probands with neurologic, cardiac, or immunologic conditions were offered GS and TGP testing. Diagnostic yield was compared using a fully paired study design. RESULTS: A total of 645 probands (median age 9 years) underwent genetic testing, and 113 (17.5%) received a molecular diagnosis. Among 642 probands with both GS and TGP testing, GS yielded 106 (16.5%) and TGPs yielded 52 (8.1%) diagnoses (P < .001). Yield was greater for GS vs TGPs in Hispanic/Latino(a) (17.2% vs 9.5%, P < .001) and White/European American (19.8% vs 7.9%, P < .001) but not in Black/African American (11.5% vs 7.7%, P = .22) population groups by self-report. A higher rate of inconclusive results was seen in the Black/African American (63.8%) vs White/European American (47.6%; P = .01) population group. Most causal copy number variants (17 of 19) and mosaic variants (6 of 8) were detected only by GS. CONCLUSION: GS may yield up to twice as many diagnoses in pediatric patients compared with TGP testing but not yet across all population groups.

Identification of copy number variants with genome sequencing: Clinical experiences from the NYCKidSeq program.

Copy number variations (CNVs) play a significant role in human disease. While chromosomal microarray has traditionally been the first-tier test for CNV detection, use of genome sequencing (GS) is increasing. We report the frequency of CNVs detected with GS in a diverse pediatric cohort from the NYCKidSeq program and highlight specific examples of its clinical impact. A total of 1052 children (0-21 years) with neurodevelopmental, cardiac, and/or immunodeficiency phenotypes received GS. Phenotype-driven analysis was used, resulting in 183 (17.4%) participants with a diagnostic result. CNVs accounted for 20.2% of participants with a diagnostic result (37/183) and ranged from 0.5 kb to 16 Mb. Of participants with a diagnostic result (n = 183) and phenotypes in more than one category, 5/17 (29.4%) were solved by a CNV finding, suggesting a high prevalence of diagnostic CNVs in participants with complex phenotypes. Thirteen participants with a diagnostic CNV (35.1%) had previously uninformative genetic testing, of which nine included a chromosomal microarray. This study demonstrates the benefits of GS for reliable detection of CNVs in a pediatric cohort with variable phenotypes.

Conducting clinical genomics research during the COVID-19 pandemic: Lessons learned from the CSER consortium experience.

Clinical research studies have navigated many changes throughout the COVID-19 pandemic. We sought to describe the pandemic's impact on research operations in the context of a clinical genomics research consortium that aimed to enroll a majority of participants from underrepresented populations. We interviewed (July to November 2020) and surveyed (May to August 2021) representatives of six projects in the Clinical Sequencing Evidence-Generating Research (CSER) consortium, which studies the implementation of genome sequencing in the clinical care of patients from populations that are underrepresented in genomics research or are medically underserved. Questions focused on COVID's impact on participant recruitment, enrollment, and engagement, and the transition to teleresearch. Responses were combined and thematically analyzed. Projects described factors at the project, institutional, and community levels that affected their experiences. Project factors included the project's progress at the pandemic's onset, the urgency of in-person clinical care for the disease being studied, and the degree to which teleresearch procedures were already incorporated. Institutional and community factors included institutional guidance for research and clinical care and the burden of COVID on the local community. Overall, being responsive to community experiences and values was essential to how CSER navigated evolving challenges during the COVID-19 pandemic.

Detection of mosaic variants using genome sequencing in a large pediatric cohort.

The increased use of next-generation sequencing has expanded our understanding of the involvement and prevalence of mosaicism in genetic disorders. We describe a total of eleven cases: nine in which mosaic variants detected by genome sequencing (GS) and/or targeted gene panels (TGPs) were considered to be causative for the proband's phenotype, and two of apparent parental mosaicism. Variants were identified in the following genes: PHACTR1, SCN8A, KCNT1, CDKL5, NEXMIF, CUX1, TSC2, GABRB2, and SMARCB1. In addition, we identified one large duplication including three genes, UBE3A, GABRB3, and MAGEL2, and one large deletion including deletion of ARFGAP1, EEF1A2, CHRNA4, and KCNQ2. All patients were enrolled in the NYCKidSeq study, a research program studying the communication of genomic information in clinical care, as well as the clinical utility and diagnostic yield of GS for children with suspected genetic disorders in diverse populations in New York City. We observed variability in the correlation between reported variant allele fraction and the severity of the patient's phenotype, although we were not able to determine the mosaicism percentage in clinically relevant tissue(s). Although our study was not sufficiently powered to assess differences in mosaicism detection between the two testing modalities, we saw a trend toward better detection by GS as compared with TGP testing. This case series supports the importance of mosaicism in childhood-onset genetic conditions and informs guidelines for laboratory and clinical interpretation of mosaic variants detected by GS.

The Genomic Medicine Integrative Research Framework: A Conceptual Framework for Conducting Genomic Medicine Research.

Conceptual frameworks are useful in research because they can highlight priority research domains, inform decisions about interventions, identify outcomes and factors to measure, and display how factors might relate to each other to generate and test hypotheses. Discovery, translational, and implementation research are all critical to the overall mission of genomic medicine and prevention, but they have yet to be organized into a unified conceptual framework. To fill this gap, our diverse team collaborated to develop the Genomic Medicine Integrative Research (GMIR) Framework, a simple but comprehensive tool to aid the genomics community in developing research questions, strategies, and measures and in integrating genomic medicine and prevention into clinical practice. Here we present the GMIR Framework and its development, along with examples of its use for research development, demonstrating how we applied it to select and harmonize measures for use across diverse genomic medicine implementation projects. Researchers can utilize the GMIR Framework for their own research, collaborative investigations, and clinical implementation efforts; clinicians can use it to establish and evaluate programs; and all stakeholders can use it to help allocate resources and make sure that the full complexity of etiology is included in research and program design, development, and evaluation.

Integration of stakeholder engagement from development to dissemination in genomic medicine research: Approaches and outcomes from the CSER Consortium.

PURPOSE: There is a critical need for genomic medicine research that reflects and benefits socioeconomically and ancestrally diverse populations. However, disparities in research populations persist, highlighting that traditional study designs and materials may be insufficient or inaccessible to all groups. New approaches can be gained through collaborations with patient/community stakeholders. Although some benefits of stakeholder engagement are recognized, routine incorporation into the design and implementation of genomics research has yet to be realized. METHODS: The National Institutes of Health-funded Clinical Sequencing Evidence-Generating Research (CSER) consortium required stakeholder engagement as a dedicated project component. Each CSER project planned and carried out stakeholder engagement activities with differing goals and expected outcomes. Examples were curated from each project to highlight engagement strategies and outcomes throughout the research lifecycle from development through dissemination. RESULTS: Projects tailored strategies to individual study needs, logistical constraints, and other challenges. Lessons learned include starting early with engagement efforts across project stakeholder groups and planned flexibility to enable adaptations throughout the project lifecycle. CONCLUSION: Each CSER project used more than 1 approach to engage with relevant stakeholders, resulting in numerous adaptations and tremendous value added throughout the full research lifecycle. Incorporation of community stakeholder insight improves the outcomes and relevance of genomic medicine research.

The Clinical Sequencing Evidence-Generating Research Consortium: Integrating Genomic Sequencing in Diverse and Medically Underserved Populations.

The Clinical Sequencing Evidence-Generating Research (CSER) consortium, now in its second funding cycle, is investigating the effectiveness of integrating genomic (exome or genome) sequencing into the clinical care of diverse and medically underserved individuals in a variety of healthcare settings and disease states. The consortium comprises a coordinating center, six funded extramural clinical projects, and an ongoing National Human Genome Research Institute (NHGRI) intramural project. Collectively, these projects aim to enroll and sequence over 6,100 participants in four years. At least 60% of participants will be of non-European ancestry or from underserved settings, with the goal of diversifying the populations that are providing an evidence base for genomic medicine. Five of the six clinical projects are enrolling pediatric patients with various phenotypes. One of these five projects is also enrolling couples whose fetus has a structural anomaly, and the sixth project is enrolling adults at risk for hereditary cancer. The ongoing NHGRI intramural project has enrolled primarily healthy adults. Goals of the consortium include assessing the clinical utility of genomic sequencing, exploring medical follow up and cascade testing of relatives, and evaluating patient-provider-laboratory level interactions that influence the use of this technology. The findings from the CSER consortium will offer patients, healthcare systems, and policymakers a clearer understanding of the opportunities and challenges of providing genomic medicine in diverse populations and settings, and contribute evidence toward developing best practices for the delivery of clinically useful and cost-effective genomic sequencing in diverse healthcare settings.

GenomeDiver: a platform for phenotype-guided medical genomic diagnosis.

PURPOSE: Making a diagnosis from clinical genomic sequencing requires well-structured phenotypic data to guide genotype interpretation. A patient's phenotypic features can be documented using the Human Phenotype Ontology (HPO), generating terms used to prioritize genes potentially causing the patient's disease. We have developed GenomeDiver to provide a user interface for clinicians that allows more effective collaboration with the clinical diagnostic laboratory, with the goal of improving the success of the diagnostic process. METHODS: GenomeDiver uses genomic data to prompt reverse phenotyping of patients undergoing genetic testing, enriching the amount and quality of structured phenotype data for the diagnostic laboratory, and helping clinicians to explore and flag diseases potentially causing their patient's presentation. RESULTS: We show how GenomeDiver communicates the clinician's informed insights to the diagnostic lab in the form of HPO terms for interpretation of genomic sequencing data. We describe our user-driven design process, the engineering of the software for efficiency, security and portability, and examples of the performance of GenomeDiver using genomic testing data. CONCLUSION: GenomeDiver is a first step in a new approach to genomic diagnostics that enhances laboratory-clinician interactions, with the goal of directly engaging clinicians to improve the outcome of genomic diagnostic testing.

GUÍA: a digital platform to facilitate result disclosure in genetic counseling.

PURPOSE: Use of genomic sequencing is increasing at a pace that requires technological solutions to effectively meet the needs of a growing patient population. We developed GUÍA, a web-based application, to enhance the delivery of genomic results and related clinical information to patients and families. METHODS: GUÍA development occurred in five overlapping phases: formative research, content development, stakeholder/community member input, user interface design, and web application development. Development was informed by formative qualitative research involving parents (N = 22) whose children underwent genomic testing. Participants enrolled in the NYCKidSeq pilot study (N = 18) completed structured feedback interviews post-result disclosure using GUÍA. Genetic specialists, researchers, patients, and community stakeholders provided their perspectives on GUÍA's design to ensure technical, cultural, and literacy appropriateness. RESULTS: NYCKidSeq participants responded positively to the use of GUÍA to deliver their children's results. All participants (N = 10) with previous experience with genetic testing felt GUÍA improved result disclosure, and 17 (94%) participants said the content was clear. CONCLUSION: GUÍA communicates complex genomic information in an understandable and personalized manner. Initial piloting demonstrated GUÍA's utility for families enrolled in the NYCKidSeq pilot study. Findings from the NYCKidSeq clinical trial will provide insight into GUÍA's effectiveness in communicating results among diverse, multilingual populations.

Establishing the value of genomics in medicine: the IGNITE Pragmatic Trials Network.

PURPOSE: A critical gap in the adoption of genomic medicine into medical practice is the need for the rigorous evaluation of the utility of genomic medicine interventions. METHODS: The Implementing Genomics in Practice Pragmatic Trials Network (IGNITE PTN) was formed in 2018 to measure the clinical utility and cost-effectiveness of genomic medicine interventions, to assess approaches for real-world application of genomic medicine in diverse clinical settings, and to produce generalizable knowledge on clinical trials using genomic interventions. Five clinical sites and a coordinating center evaluated trial proposals and developed working groups to enable their implementation. RESULTS: Two pragmatic clinical trials (PCTs) have been initiated, one evaluating genetic risk APOL1 variants in African Americans in the management of their hypertension, and the other to evaluate the use of pharmacogenetic testing for medications to manage acute and chronic pain as well as depression. CONCLUSION: IGNITE PTN is a network that carries out PCTs in genomic medicine; it is focused on diversity and inclusion of underrepresented minority trial participants; it uses electronic health records and clinical decision support to deliver the interventions. IGNITE PTN will develop the evidence to support (or oppose) the adoption of genomic medicine interventions by patients, providers, and payers.

Public Attitudes toward Consent and Data Sharing in Biobank Research: A Large Multi-site Experimental Survey in the US.

Individuals participating in biobanks and other large research projects are increasingly asked to provide broad consent for open-ended research use and widespread sharing of their biosamples and data. We assessed willingness to participate in a biobank using different consent and data sharing models, hypothesizing that willingness would be higher under more restrictive scenarios. Perceived benefits, concerns, and information needs were also assessed. In this experimental survey, individuals from 11 US healthcare systems in the Electronic Medical Records and Genomics (eMERGE) Network were randomly allocated to one of three hypothetical scenarios: tiered consent and controlled data sharing; broad consent and controlled data sharing; or broad consent and open data sharing. Of 82,328 eligible individuals, exactly 13,000 (15.8%) completed the survey. Overall, 66% (95% CI: 63%-69%) of population-weighted respondents stated they would be willing to participate in a biobank; willingness and attitudes did not differ between respondents in the three scenarios. Willingness to participate was associated with self-identified white race, higher educational attainment, lower religiosity, perceiving more research benefits, fewer concerns, and fewer information needs. Most (86%, CI: 84%-87%) participants would want to know what would happen if a researcher misused their health information; fewer (51%, CI: 47%-55%) would worry about their privacy. The concern that the use of broad consent and open data sharing could adversely affect participant recruitment is not supported by these findings. Addressing potential participants' concerns and information needs and building trust and relationships with communities may increase acceptance of broad consent and wide data sharing in biobank research.

Correction: Developing a common framework for evaluating the implementation of genomic medicine interventions in clinical care: the IGNITE Network's Common Measures Working Group.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

"Not Alone Anymore": The Experiences of Adults With Diabetes in New York's Medicaid Health Home Program.

BACKGROUND: New York State Medicaid's Health Home program is an example of a natural experiment that could affect individuals with diabetes. While evaluations of interventions such as the Health Home program are generally based solely on clinical and administrative data and rarely examine patients' experience, patients may add to the understanding of the intervention's implementation and mechanisms of impact. OBJECTIVE: The objective of this study was to qualitatively examine the health and nonmedical challenges faced by Medicaid-insured patients with diabetes and their experiences with the services provided by New York's Health Homes to address these challenges. RESEARCH DESIGN: We performed 10 focus groups and 23 individual interviews using a guide developed in collaboration with a stakeholder board. We performed a thematic analysis to identify cross-cutting themes. SUBJECTS: A total of 63 Medicaid-insured individuals with diabetes, 31 of whom were enrolled in New York's Health Home program. RESULTS: While participants were not generally familiar with the term "Health Home," they described and appreciated services consistent with Health Home enrollment delivered by care managers. Services addressed challenges in access to care, especially by facilitating and reminding participants about appointments, and nonmedical needs, such as transportation, housing, and help at home. Participants valued their personal relationships with care managers and the psychosocial support they provided. CONCLUSIONS: From the perspective of its enrollees, the Health Home program primarily addressed access to care, but also addressed material and psychosocial needs. These findings have implications for Health Home entities and for research assessing their impact.

Network Engagement in Action: Stakeholder Engagement Activities to Enhance Patient-centeredness of Research.

BACKGROUND: Stakeholders (ie, patients, policymakers, clinicians, advocacy groups, health system leaders, payers, and others) offer critical input at various stages in the research continuum, and their contributions are increasingly recognized as an important component of effective translational research. Natural experiments, in particular, may benefit from stakeholder feedback in addressing real-world issues and providing insight into future policy decisions, though best practices for the engagement of stakeholders in observational studies are limited in the literature. METHODS: The Natural Experiments for Translation in Diabetes 2.0 (NEXT-D2) network utilizes rigorous methods to evaluate natural experiments in health policy and program delivery with a focus on diabetes-related outcomes. Each of the 8 partnering institutions incorporates stakeholder engagement throughout multiple study phases to enhance the patient-centeredness of results. NEXT-D2 dedicates a committee to Engagement for resource sharing, enhancing engagement approaches, and advancing network-wide engagement activities. Key stakeholder engagement activities include Study Meetings, Proposal Development, Trainings & Educational Opportunities, Data Analysis, and Results Dissemination. Network-wide patient-centered resources and multimedia have also been developed through the broad expertise of each site's stakeholder group. CONCLUSIONS: This collaboration has created a continuous feedback loop wherein site-level engagement approaches are informed via the network and network-level engagement efforts are shaped by individual sites. Emerging best practices include: incorporating stakeholders in multiple ways throughout the research, building on previous relationships with stakeholders, enhancing capacity through stakeholder and investigator training, involving stakeholders in refining outcome choices and understanding the meaning of variables, and recognizing the power of stakeholders in maximizing dissemination.

Machine Learning to Predict Mortality and Critical Events in a Cohort of Patients With COVID-19 in New York City: Model Development and Validation.

BACKGROUND: COVID-19 has infected millions of people worldwide and is responsible for several hundred thousand fatalities. The COVID-19 pandemic has necessitated thoughtful resource allocation and early identification of high-risk patients. However, effective methods to meet these needs are lacking. OBJECTIVE: The aims of this study were to analyze the electronic health records (EHRs) of patients who tested positive for COVID-19 and were admitted to hospitals in the Mount Sinai Health System in New York City; to develop machine learning models for making predictions about the hospital course of the patients over clinically meaningful time horizons based on patient characteristics at admission; and to assess the performance of these models at multiple hospitals and time points. METHODS: We used Extreme Gradient Boosting (XGBoost) and baseline comparator models to predict in-hospital mortality and critical events at time windows of 3, 5, 7, and 10 days from admission. Our study population included harmonized EHR data from five hospitals in New York City for 4098 COVID-19-positive patients admitted from March 15 to May 22, 2020. The models were first trained on patients from a single hospital (n=1514) before or on May 1, externally validated on patients from four other hospitals (n=2201) before or on May 1, and prospectively validated on all patients after May 1 (n=383). Finally, we established model interpretability to identify and rank variables that drive model predictions. RESULTS: Upon cross-validation, the XGBoost classifier outperformed baseline models, with an area under the receiver operating characteristic curve (AUC-ROC) for mortality of 0.89 at 3 days, 0.85 at 5 and 7 days, and 0.84 at 10 days. XGBoost also performed well for critical event prediction, with an AUC-ROC of 0.80 at 3 days, 0.79 at 5 days, 0.80 at 7 days, and 0.81 at 10 days. In external validation, XGBoost achieved an AUC-ROC of 0.88 at 3 days, 0.86 at 5 days, 0.86 at 7 days, and 0.84 at 10 days for mortality prediction. Similarly, the unimputed XGBoost model achieved an AUC-ROC of 0.78 at 3 days, 0.79 at 5 days, 0.80 at 7 days, and 0.81 at 10 days. Trends in performance on prospective validation sets were similar. At 7 days, acute kidney injury on admission, elevated LDH, tachypnea, and hyperglycemia were the strongest drivers of critical event prediction, while higher age, anion gap, and C-reactive protein were the strongest drivers of mortality prediction. CONCLUSIONS: We externally and prospectively trained and validated machine learning models for mortality and critical events for patients with COVID-19 at different time horizons. These models identified at-risk patients and uncovered underlying relationships that predicted outcomes.

Trends and Racial Disparities of Palliative Care Use among Hospitalized Patients with ESKD on Dialysis.

BACKGROUND: Study findings show that although palliative care decreases symptom burden, it is still underused in patients with ESKD. Little is known about disparity in use of palliative care services in such patients in the inpatient setting. METHODS: To investigate the use of palliative care consultation in patients with ESKD in the inpatient setting, we conducted a retrospective cohort study using the National Inpatient Sample from 2006 to 2014 to identify admitted patients with ESKD requiring maintenance dialysis. We compared palliative care use among minority groups (black, Hispanic, and Asian) and white patients, adjusting for patient and hospital variables. RESULTS: We identified 5,230,865 hospitalizations of such patients from 2006 through 2014, of which 76,659 (1.5%) involved palliative care. The palliative care referral rate increased significantly, from 0.24% in 2006 to 2.70% in 2014 (P<0.01). Black and Hispanic patients were significantly less likely than white patients to receive palliative care services (adjusted odds ratio [aOR], 0.72; 95% confidence interval [95% CI], 0.61 to 0.84, P<0.01 for blacks and aOR, 0.46; 95% CI, 0.30 to 0.68, P<0.01 for Hispanics). These disparities spanned across all hospital subtypes, including those with higher proportions of minorities. Minority patients with lower socioeconomic status (lower level of income and nonprivate health insurance) were also less likely to receive palliative care. CONCLUSIONS: Despite a clear increase during the study period in provision of palliative care for inpatients with ESKD, significant racial disparities occurred and persisted across all hospital subtypes. Further investigation into causes of racial and ethnic disparities is necessary to improve access to palliative care services for the vulnerable ESKD population.