Validation of an ICD-Code-Based Case Definition for Psychotic Illness Across Three Health Systems.

BACKGROUND AND HYPOTHESIS: Psychosis-associated diagnostic codes are increasingly being utilized as case definitions for electronic health record (EHR)-based algorithms to predict and detect psychosis. However, data on the validity of psychosis-related diagnostic codes is limited. We evaluated the positive predictive value (PPV) of International Classification of Diseases (ICD) codes for psychosis. STUDY DESIGN: Using EHRs at 3 health systems, ICD codes comprising primary psychotic disorders and mood disorders with psychosis were grouped into 5 higher-order groups. 1133 records were sampled for chart review using the full EHR. PPVs (the probability of chart-confirmed psychosis given ICD psychosis codes) were calculated across multiple treatment settings. STUDY RESULTS: PPVs across all diagnostic groups and hospital systems exceeded 70%: Mass General Brigham 0.72 [95% CI 0.68-0.77], Boston Children's Hospital 0.80 [0.75-0.84], and Boston Medical Center 0.83 [0.79-0.86]. Schizoaffective disorder PPVs were consistently the highest across sites (0.80-0.92) and major depressive disorder with psychosis were the most variable (0.57-0.79). To determine if the first documented code captured first-episode psychosis (FEP), we excluded cases with prior chart evidence of a diagnosis of or treatment for a psychotic illness, yielding substantially lower PPVs (0.08-0.62). CONCLUSIONS: We found that the first documented psychosis diagnostic code accurately captured true episodes of psychosis but was a poor index of FEP. These data have important implications for the case definitions used in the development of risk prediction models designed to predict or detect undiagnosed psychosis.

The Health Equity Explorer: An open-source resource for distributed health equity visualization and research across common data models.

Introduction: There is an urgent need to address pervasive inequities in health and healthcare in the USA. Many areas of health inequity are well known, but there remain important unexplored areas, and for many populations in the USA, accessing data to visualize and monitor health equity is difficult. Methods: We describe the development and evaluation of an open-source, R-Shiny application, the "Health Equity Explorer (H2E)," designed to enable users to explore health equity data in a way that can be easily shared within and across common data models (CDMs). Results: We have developed a novel, scalable informatics tool to explore a wide variety of drivers of health, including patient-reported Social Determinants of Health (SDoH), using data in an OMOP CDM research data repository in a way that can be easily shared. We describe our development process, data schema, potential use cases, and pilot data for 705,686 people who attended our health system at least once since 2016. For this group, 996,382 unique observations for questions related to food and housing security were available for 324,630 patients (at least one answer for all 46% of patients) with 65,152 (20.1% of patients with at least one visit and answer) reporting food or housing insecurity at least once. Conclusions: H2E can be used to support dynamic and interactive explorations that include rich social and environmental data. The tool can support multiple CDMs and has the potential to support distributed health equity research and intervention on a national scale.

The Longitudinal Reciprocal Relationships of Tooth Brushing Behaviors Between Underserved Children and Their Caregivers.

BACKGROUND: Tooth brushing is effective in preventing early childhood caries. However, it is unclear how children's and caregiver's tooth brushing are reciprocally related. PURPOSE: The current study investigated whether the longitudinal relationships between children and caregiver tooth brushing are moderated by a caregiver-targeted child oral health intervention and caregiver depression. METHODS: Secondary analysis of a randomized clinical trial that tested whether caregiver-targeted oral health text messages (OHT) outperformed child wellness text messages (CWT) on pediatric dental caries and oral health behaviors (n = 754, mean child age = 2.9 years, 56.2% Black, 68.3%

Tooth brushing is effective in preventing dental cavities in children, but we do not know if or how children and caregiver brushing frequencies are related. This is important because interventions targeting children's oral health may also have the potential to benefit their caregiver's behaviors. Our study examined whether caregiver brushing of their own teeth and caregiver brushing of their young child's teeth positively influenced each other over time. We also explored whether this relationship was less likely if caregivers experienced depressive symptoms and more likely if caregivers participated in a text message program focused on improving their child's oral health. Results showed that caregiver and child tooth brushing behaviors positively influenced each other over time, but this relationship was observed only in caregivers who received the child oral health program (as opposed to the control group) and who reported low depressive symptoms (in contrast to caregivers with high depression symptoms). Our findings suggest that while caregivers and children positively influence each other's tooth-brushing behaviors over time, additional support is essential for caregivers experiencing depression to fully realize these reciprocal benefits.

Validation of an ICD-code-based case definition for psychotic illness across three health systems.

Background and Hypothesis: Early detection of psychosis is critical for improving outcomes. Algorithms to predict or detect psychosis using electronic health record (EHR) data depend on the validity of the case definitions used, typically based on diagnostic codes. Data on the validity of psychosis-related diagnostic codes is limited. We evaluated the positive predictive value (PPV) of International Classification of Diseases (ICD) codes for psychosis. Study Design: Using EHRs at three health systems, ICD codes comprising primary psychotic disorders and mood disorders with psychosis were grouped into five higher-order groups. 1,133 records were sampled for chart review using the full EHR. PPVs (the probability of chart-confirmed psychosis given ICD psychosis codes) were calculated across multiple treatment settings. Study Results: PPVs across all diagnostic groups and hospital systems exceeded 70%: Massachusetts General Brigham 0.72 [95% CI 0.68-0.77], Boston Children's Hospital 0.80 [0.75-0.84], and Boston Medical Center 0.83 [0.79-0.86]. Schizoaffective disorder PPVs were consistently the highest across sites (0.80-0.92) and major depressive disorder with psychosis were the most variable (0.57-0.79). To determine if the first documented code captured first-episode psychosis (FEP), we excluded cases with prior chart evidence of a diagnosis of or treatment for a psychotic illness, yielding substantially lower PPVs (0.08-0.62). Conclusions: We found that the first documented psychosis diagnostic code accurately captured true episodes of psychosis but was a poor index of FEP. These data have important implications for the development of risk prediction models designed to predict or detect undiagnosed psychosis.

Racial Disparities in Diabetes Care and Outcomes for Patients with Visual Impairment: A Descriptive Analysis of the TriNetX Research Network.

Background: This research delves into the confluence of racial disparities and health inequities among individuals with disabilities, with a focus on those contending with both diabetes and visual impairment. Methods: Utilizing data from the TriNetX Research Network, which includes electronic medical records of roughly 115 million patients from 83 anonymous healthcare organizations, this study employs a directed acyclic graph (DAG) to pinpoint confounders and augment interpretation. We identified patients with visual impairments using ICD-10 codes, deliberately excluding diabetes-related ophthalmology complications. Our approach involved multiple race-stratified analyses, comparing co-morbidities like chronic pulmonary disease in visually impaired patients against their counterparts. We assessed healthcare access disparities by examining the frequency of annual visits, instances of two or more A1c measurements, and glomerular filtration rate (GFR) measurements. Additionally, we evaluated diabetes outcomes by comparing the risk ratio of uncontrolled diabetes (A1c > 9.0) and chronic kidney disease in patients with and without visual impairments. Results: The incidence of diabetes was substantially higher (nearly double) in individuals with visual impairments across White, Asian, and African American populations. Higher rates of chronic kidney disease were observed in visually impaired individuals, with a risk ratio of 1.79 for African American, 2.27 for White, and non-significant for the Asian group. A statistically significant difference in the risk ratio for uncontrolled diabetes was found only in the White cohort (0.843). White individuals without visual impairments were more likely to receive two A1c tests, a trend not significant in other racial groups. African Americans with visual impairments had a higher rate of glomerular filtration rate testing. However, White individuals with visual impairments were less likely to undergo GFR testing, indicating a disparity in kidney health monitoring. This pattern of disparity was not observed in the Asian cohort. Conclusions: This study uncovers pronounced disparities in diabetes incidence and management among individuals with visual impairments, particularly among White, Asian, and African American groups. Our DAG analysis illuminates the intricate interplay between SDoH, healthcare access, and frequency of crucial diabetes monitoring practices, highlighting visual impairment as both a medical and social issue.

Independent and joint association of cord plasma pantothenate and cysteine levels with autism spectrum disorders and other neurodevelopmental disabilities in children born term and preterm.

Background: Pantothenate (vitamin B5) is a precursor for coenzyme A (CoA) synthesis, which serves as a cofactor for hundreds of metabolic reactions. Cysteine is an amino acid in the CoA synthesis pathway. To date, research on the combined role of early life pantothenate and cysteine levels in childhood neurodevelopmental disabilities is scarce. Objective: To study the association between cord pantothenate and cysteine levels and risk of autism spectrum disorder (ASD), attention deficit hyperactivity disorder (ADHD) and other developmental disabilities (DD) in children born term and preterm. Methods: The study sample (n = 996, 177 born preterm) derived from the Boston Birth Cohort included 416 neurotypical children, 87 ASD, 269 ADHD, and 224 other DD children, who were mutually exclusive. Participants were enrolled at birth and were followed up prospectively (from October 1, 1998, to June 30, 2018) at the Boston Medical Center. Cord blood sample was collected at birth. Plasma pantothenate and cysteine levels were measured using liquid chromatography-tandem mass spectrometry. Results: Higher cord pantothenate (≥50th percentile vs. <50th percentile) was associated with a greater risk of ASD (adjusted odds ratio [aOR]: 1.94, 95% confidence interval [CI]: 1.06, 3.55) and ADHD (aOR: 1.66, 95% CI: 1.14, 2.40), after adjusting for potential confounders. However, cord cysteine alone was not associated with risk of ASD, ADHD, or other DD. When considering the joint association, greater ASD risk was noted when both cord pantothenate and cysteine levels were elevated (≥50th percentile) (aOR: 3.11, 95% CI: 1.24, 7.79), when compared to children with low cord pantothenate (<50th percentile) and high cysteine. Even though preterm and higher pantothenate independently increased the ASD risk, the greatest risk was found in preterm children who also had elevated pantothenate (≥50th percentile), which was true for all three outcomes: ASD (aOR: 5.36, 95% CI: 2.09, 13.75), ADHD (aOR: 3.31, 95% CI: 1.78, 6.16), and other DD (aOR: 3.39, 95% CI: 1.85, 6.24). Conclusions: In this prospective birth cohort, we showed that higher cord pantothenate individually and in combination with higher cysteine or preterm birth were associated with increased risk of ASD and ADHD. More study is needed to explore this biologically plausible pathway.

Maternal pre-pregnancy BMI, offspring epigenome-wide DNA methylation, and childhood obesity: findings from the Boston Birth Cohort.

BACKGROUND: Maternal pre-pregnancy obesity is an established risk factor for childhood obesity. Investigating epigenetic alterations induced by maternal obesity during fetal development could gain mechanistic insight into the developmental origins of childhood obesity. While obesity disproportionately affects underrepresented racial and ethnic mothers and children in the USA, few studies investigated the role of prenatal epigenetic programming in intergenerational obesity of these high-risk populations. METHODS: This study included 903 mother-child pairs from the Boston Birth Cohort, a predominantly urban, low-income minority birth cohort. Mother-infant dyads were enrolled at birth and the children were followed prospectively to age 18 years. Infinium Methylation EPIC BeadChip was used to measure epigenome-wide methylation level of cord blood. We performed an epigenome-wide association study of maternal pre-pregnancy body mass index (BMI) and cord blood DNA methylation (DNAm). To quantify the degree to which cord blood DNAm mediates the maternal BMI-childhood obesity, we further investigated whether maternal BMI-associated DNAm sites impact birthweight or childhood overweight or obesity (OWO) from age 1 to age 18 and performed corresponding mediation analyses. RESULTS: The study sample contained 52.8% maternal pre-pregnancy OWO and 63.2% offspring OWO at age 1-18 years. Maternal BMI was associated with cord blood DNAm at 8 CpG sites (genome-wide false discovery rate [FDR] < 0.05). After accounting for the possible interplay of maternal BMI and smoking, 481 CpG sites were discovered for association with maternal BMI. Among them 123 CpGs were associated with childhood OWO, ranging from 42% decrease to 87% increase in OWO risk for each SD increase in DNAm. A total of 14 identified CpG sites showed a significant mediation effect on the maternal BMI-child OWO association (FDR < 0.05), with mediating proportion ranging from 3.99% to 25.21%. Several of these 14 CpGs were mapped to genes in association with energy balance and metabolism (AKAP7) and adulthood metabolic syndrome (CAMK2B). CONCLUSIONS: This prospective birth cohort study in a high-risk yet understudied US population found that maternal pre-pregnancy OWO significantly altered DNAm in newborn cord blood and provided suggestive evidence of epigenetic involvement in the intergenerational risk of obesity.

The differential predictive utility of two caregiver-targeted self-efficacy measures to promote oral health of underserved children.

OBJECTIVE: Oral health self-efficacy is a modifiable determinant of early childhood caries, which is one of the most prevalent childhood diseases. Yet, two common measures of self-efficacy (i.e., context-specific and behavior-specific) lack validation and clarity in the prediction of children's oral health behaviors. This study examined the psychometric properties of two caregiver oral health self-efficacy measures and investigated the predictive ability and age-varying effects of caregiver oral health self-efficacy on child oral health behaviors. METHOD: In this secondary data analysis of caregiver-child dyads (n = 754, Mchild age = 2.4, 56.2% Black or African American, 68.3% below poverty level), caregivers reported their oral health self-efficacy and their child's tooth brushing frequency, diet, and sugar-sweetened beverage (SSB) consumption at baseline and 4, 12, and 24 months. Psychometrics were examined with confirmatory factor analyses (CFAs) and the predictive ability and age-varying effects of caregiver self-efficacy on child oral health behaviors were examined with time-varying effect models (TVEMs). RESULTS: The context- and behavior-specific oral health self-efficacy CFA models indicated mixed model fit. In the predictive TVEM models, greater behavior-specific, but not context-, oral health self-efficacy predicted greater child tooth brushing across all ages. Greater context-specific oral health self-efficacy predicted healthier child diet throughout childhood, but greater behavior-specific self-efficacy only predicted healthier child diet in older children. Greater behavior-specific self-efficacy predicted lower SSB consumption throughout childhood while greater context-specific self-efficacy only predicted lower SSB consumption in younger children. CONCLUSIONS: Both caregiver oral health self-efficacy measures were psychometrically comparable and differentially predicted oral health behaviors across varying childhood ages. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Umbilical cord DNA methylation is associated with body mass index trajectories from birth to adolescence.

BACKGROUND: DNA methylation (DNAm) in cord blood has been associated with various prenatal factors and birth outcomes. This study sought to fill an important knowledge gap: the link of cord DNAm with child postnatal growth trajectories from birth to age 18 years (y). METHODS: Using data from a US predominantly urban, low-income, multi-ethnic birth cohort (N = 831), we first applied non-parametric methods to identify body-mass-index percentile (BMIPCT) trajectories from birth to age 18 y (the outcome); then, conducted epigenome-wide association study (EWAS) of the outcome, interrogating over 700,000 CpG sites profiled by the Illumina Infinium MethylationEPIC BeadChip. Multivariate linear regression models and likelihood ratio tests (LRT) were applied to examine the DNAm-outcome association in the overall sample and sex strata. FINDINGS: We identified four distinct patterns of BMIPCT trajectories: normal weight (NW), Early overweight or obesity (OWO), Late OWO, and normal to very late OWO. DNAm at CpG18582997 annotated to TPGS1, CpG15241084 of TLR7, and cg24350936 of RAB31 were associated with BMIPCT at birth-to-3 y, 10 y, and 14 y, respectively (LRT FDR < 0.05 for all). INTERPRETATION: In this prospective birth cohort study, we identified 4 distinct and robust patterns of growth trajectories from birth to 18 y, which were associated with variations in cord blood DNAm at genes implicated in inflammation induction pathways. These findings, if further replicated, raise the possibility that these DNAm markers along with early assessment of BMIPCT trajectories may help identify young children at high-risk for obesity later in life. FUNDING: Detailed in the Acknowledgements section.

Plasma Insulin Concentration in Newborns and Children and Age at Menarche.

OBJECTIVE: To investigate the association of plasma insulin levels and their trajectories from birth to childhood with the timing of menarche. RESEARCH DESIGN AND METHODS: This prospective study included 458 girls recruited at birth between 1998 and 2011 and followed prospectively at the Boston Medical Center. Plasma nonfasting insulin concentrations were measured at two time points: at birth (cord blood) and in childhood (age 0.5-5 years). Age at menarche was obtained from a pubertal developmental questionnaire or abstracted from electronic medical records. RESULTS: Three hundred six (67%) of the girls had reached menarche. The median (range) age at menarche was 12.4 (9-15) years. Elevated plasma insulin concentrations at birth (n = 391) and in childhood (n = 335) were each associated with an earlier mean age at menarche: approximately 2 months earlier per doubling of insulin concentration (mean shift, -1.95 months, 95% CI, -0.33 to -3.53, and -2.07 months, 95% CI, -0.48 to -3.65, respectively). Girls with overweight or obesity in addition to elevated insulin attained menarche about 11-17 months earlier, on average, than those with normal weight and low insulin. Considering longitudinal trajectories (n = 268), having high insulin levels both at birth and in childhood was associated with a roughly 6 months earlier mean age at menarche (mean shift, -6.25 months, 95% CI, -0.38 to -11.88), compared with having consistently low insulin levels at both time points. CONCLUSIONS: Our data showed that elevated insulin concentrations in early life, especially in conjunction with overweight or obesity, contribute to the earlier onset of menarche, suggesting the need for early screening and intervention.

Cord Blood Plasma Metabolome-wide Associations With Height From Birth to Adolescence.

Although the maternal intrauterine metabolic environment has been recognized to have a profound impact on fetal growth and development with lifelong health implications, to our knowledge, there have been few large-scale birth cohort studies linking the cord metabolome (reflecting both the maternal and fetal metabolic state) with postnatal height measurements across the pediatric age range. Using data from the Boston Birth Cohort, an ongoing prospective birth cohort, this study investigated the association of cord plasma metabolites with children's height from birth to adolescence. Height was analyzed as attained height and longitudinal trajectories. Distinctive cord metabolite types were associated with attained height at different developmental windows: triacylglycerols [TAGs], diacylglycerols [DAGs], cholesterol ester [CEs], phospholipids, amino acids [AAs], acylcarnitines [ACs], and nucleotides in early (age 0-4 years) and middle (age 6-12 years) childhood; various metabolite types other than TAGs in later childhood (after age 14 years). Functional principal component analysis on children's repeated height measurements summarized two typical height trajectory components: loadings on first eigenfunction [FPC1] representing overall height by age, and loadings on second eigenfunction [FPC2] representing speed of pubertal height growth. Although only one cord metabolite was correlated with FPC1 after accounting for multiple testing, the study found 27 metabolites with significant overall effect on FPC2 among females and 18 among males. These metabolites were mostly phospholipids (including phosphatidylethanolamines [PEs], phosphatidylethanolamine plasmalogens [PE_Ps], phosphatidylcholines [PCs], lysophosphatidylethanolamines [LPEs], and lysophosphatidylcholines [LPCs]), AAs, and nucleotides. Their associations with height differed between overweight/obesity (OWO) and non-OWO children, especially among females. In this prospective study of US understudied urban, low-income, racially diverse children, we demonstrated that cord plasma metabolites were significantly associated with postnatal attained height at different age windows as well as height trajectories from birth to adolescence. We also revealed how these associations differed by children's sex and OWO status. Our findings help elucidate metabolic pathways underlying fetal origins of height growth across developmental stages. © 2023 The Authors. Journal of Bone and Mineral Research published by Wiley Periodicals LLC on behalf of American Society for Bone and Mineral Research (ASBMR).

Prospective Cohort Study of Emergency Department Visit Frequency and Diagnoses Before and During COVID-19 Pandemic in Urban, Low-Income, US- and Foreign-Born Mothers in Boston, MA.

Background: The coronavirus 2019 (COVID-19) pandemic fundamentally changed how populations interface with the healthcare system. Despite historical spikes in US mortality during the pandemic, emergency department (ED) visits were paradoxically low. This is a concerning phenomenon that raises a red flag regarding access to care, especially among vulnerable populations. In this study we sought to understand how ED utilization evolved during the COVID-19 pandemic among traditionally understudied, low-income, racially diverse US- and foreign-born mothers. Methods: This is a secondary analysis of a pre-existing dataset of 3,073 participants enrolled in the Boston Birth Cohort at birth and followed prospectively. We obtained ED visit diagnoses from 2019 and 2020 via electronic health records, categorized according to the International Classification of Diseases, 10th Revision, and compared them using graph plots, chi-square, and negative binomial regression. Results: The number of ED visits decreased by 29.1% (P < 0.001) from 2019 (1,376) to 2020 (976). However, visits for infectious and parasitic diseases, including COVID-19, increased by 90.6% (32:61) with COVID-19 accounting for 77% of those visits in 2020 (47/61). Mental health-related visits increased by 40.9% (44:62), with diagnoses of alcohol use disorder increasing by 183% (6:17). Regression analysis showed 50% less ED utilization among foreign- vs US-born participants; however, the increase in infectious diseases visits was greater among foreign-born compared to US-born mothers (185% vs 26%, P = 0.01), while the increase in mental health diagnoses was greater among US-born mothers (69% vs -33%, P = 0.10). Conclusion: Despite a decrease in total ED visits during the pandemic, there was an increase in COVID-19- (immigrant > US born) and mental health- (US-born only) related visits. Our findings demonstrate that EDs remain a critical access point for care for minority populations and have implications for preparedness, resources, and services of EDs in urban settings to better address the needs of communities. However, alternative avenues for healthcare services for these populations, particularly during health crises, warrant further investigation.

In-home environmental exposures predicted from geospatial characteristics of the built environment and electronic health records of children with asthma.

PURPOSE: Children may be exposed to numerous in-home environmental exposures (IHEE) that trigger asthma exacerbations. Spatially linking social and environmental exposures to electronic health records (EHR) can aid exposure assessment, epidemiology, and clinical treatment, but EHR data on exposures are missing for many children with asthma. To address the issue, we predicted presence of indoor asthma trigger allergens, and estimated effects of their key geospatial predictors. METHODS: Our study samples were comprised of children with asthma who provided self-reported IHEE data in EHR at a safety-net hospital in New England during 2004-2015. We used an ensemble machine learning algorithm and 86 multilevel features (e.g., individual, housing, neighborhood) to predict presence of cockroaches, rodents (mice or rats), mold, and bedroom carpeting/rugs in homes. We reduced dimensionality via elastic net regression and estimated effects by the G-computation causal inference method. RESULTS: Our models reasonably predicted presence of cockroaches (area under receiver operating curves [AUC] = 0.65), rodents (AUC = 0.64), and bedroom carpeting/rugs (AUC = 0.64), but not mold (AUC = 0.54). In models adjusted for confounders, higher average household sizes in census tracts were associated with more reports of pests (cockroaches and rodents). Tax-exempt parcels were associated with more reports of cockroaches in homes. Living in a White-segregated neighborhood was linked with lower reported rodent presence, and mixed residential/commercial housing and newer buildings were associated with more reports of bedroom carpeting/rugs in bedrooms. CONCLUSIONS: We innovatively applied a machine learning and causal inference mixture methodology to detail IHEE among children with asthma using EHR and geospatial data, which could have wide applicability and utility.

Advancing Interoperability of Patient-level Social Determinants of Health Data to Support COVID-19 Research.

Including social determinants of health (SDoH) data in health outcomes research is essential for studying the sources of healthcare disparities and developing strategies to mitigate stressors. In this report, we describe a pragmatic design and approach to explore the encoding needs for transmitting SDoH screening tool responses from a large safety-net hospital into the National Covid Cohort Collaborative (N3C) OMOP dataset. We provide a stepwise account of designing data mapping and ingestion for patient-level SDoH and summarize the results of screening. Our approach demonstrates that sharing of these important data - typically stored as non-standard, EHR vendor specific codes - is feasible. As SDoH screening gains broader use nationally, the approach described in this paper could be used for other screening instruments and improve the interoperability of these important data.

Development and validation of predictive models for COVID-19 outcomes in a safety-net hospital population.

OBJECTIVE: To develop predictive models of coronavirus disease 2019 (COVID-19) outcomes, elucidate the influence of socioeconomic factors, and assess algorithmic racial fairness using a racially diverse patient population with high social needs. MATERIALS AND METHODS: Data included 7,102 patients with positive (RT-PCR) severe acute respiratory syndrome coronavirus 2 test at a safety-net system in Massachusetts. Linear and nonlinear classification methods were applied. A score based on a recurrent neural network and a transformer architecture was developed to capture the dynamic evolution of vital signs. Combined with patient characteristics, clinical variables, and hospital occupancy measures, this dynamic vital score was used to train predictive models. RESULTS: Hospitalizations can be predicted with an area under the receiver-operating characteristic curve (AUC) of 92% using symptoms, hospital occupancy, and patient characteristics, including social determinants of health. Parsimonious models to predict intensive care, mechanical ventilation, and mortality that used the most recent labs and vitals exhibited AUCs of 92.7%, 91.2%, and 94%, respectively. Early predictive models, using labs and vital signs closer to admission had AUCs of 81.1%, 84.9%, and 92%, respectively. DISCUSSION: The most accurate models exhibit racial bias, being more likely to falsely predict that Black patients will be hospitalized. Models that are only based on the dynamic vital score exhibited accuracies close to the best parsimonious models, although the latter also used laboratories. CONCLUSIONS: This large study demonstrates that COVID-19 severity may accurately be predicted using a score that accounts for the dynamic evolution of vital signs. Further, race, social determinants of health, and hospital occupancy play an important role.

Clinical, social, and policy factors in COVID-19 cases and deaths: methodological considerations for feature selection and modeling in county-level analyses.

BACKGROUND: There is a need to evaluate how the choice of time interval contributes to the lack of consistency of SDoH variables that appear as important to COVID-19 disease burden within an analysis for both case counts and death counts. METHODS: This study identified SDoH variables associated with U.S county-level COVID-19 cumulative case and death incidence for six different periods: the first 30, 60, 90, 120, 150, and 180 days since each county had COVID-19 one case per 10,000 residents. The set of SDoH variables were in the following domains: resource deprivation, access to care/health resources, population characteristics, traveling behavior, vulnerable populations, and health status. A generalized variance inflation factor (GVIF) analysis was used to identify variables with high multicollinearity. For each dependent variable, a separate model was built for each of the time periods. We used a mixed-effect generalized linear modeling of counts normalized per 100,000 population using negative binomial regression. We performed a Kolmogorov-Smirnov goodness of fit test, an outlier test, and a dispersion test for each model. Sensitivity analysis included altering the county start date to the day each county reached 10 COVID-19 cases per 10,000. RESULTS: Ninety-seven percent (3059/3140) of the counties were represented in the final analysis. Six features proved important for both the main and sensitivity analysis: adults-with-college-degree, days-sheltering-in-place-at-start, prior-seven-day-median-time-home, percent-black, percent-foreign-born, over-65-years-of-age, black-white-segregation, and days-since-pandemic-start. These variables belonged to the following categories: COVID-19 related, vulnerable populations, and population characteristics. Our diagnostic results show that across our outcomes, the models of the shorter time periods (30 days, 60 days, and 900 days) have a better fit. CONCLUSION: Our findings demonstrate that the set of SDoH features that are significant for COVID-19 outcomes varies based on the time from the start date of the pandemic and when COVID-19 was present in a county. These results could assist researchers with variable selection and inform decision makers when creating public health policy.

Longitudinal trajectories and determinants of plasma per- and polyfluoroalkyl substance (PFAS) levels from birth to early childhood and metabolomic associations: A pilot study in the Boston Birth Cohort.

Background: Per- and polyfluoroalkyl substances (PFAS) are a major public health concern worldwide due to their ubiquitous exposures, environmental persistence, maternal-to-fetal transfer, and multi-organ toxicity. This pilot study aimed to generate preliminary data to inform future studies to address data gaps in the field, including early life PFAS exposure levels, longitudinal changes, determinants, and associated metabolomic alterations in understudied Black and Hispanic children in the United States (U.S.). Methods: This study leveraged existing biosamples and data in the Boston Birth Cohort and measured 12 legacy and emerging PFAS, including Me-PFOSA-AcOH, PFDA, PFDoA, PFHxS, PFNA, PFOA, PFOS, PFUnA, GenX, ADONA, 9Cl-PF3ONS, and PFHpS, in paired cord and early childhood plasma samples. Summary statistics and graphic plots were used to depict PFAS levels at the two time points and their longitudinal changes. Linear regression models were used to identify the early-life factors associated with cord and early childhood PFAS levels. Associations of cord PFAS with cord metabolites were explored using a metabolome-wide association approach and a targeted approach. Results: This study included 39 children, of whom 25 (64%) were Black, 14 (36%) were Hispanic, and 15 (38%) were female. PFOA, PFOS, PFNA, and PFHpS were detectable in all cord and early childhood plasma samples, while GenX and ADONA were not detectable in any sample. Cord PFAS levels were weakly-to-moderately correlated with early childhood PFAS levels (r = -0.03 to 0.40). Several maternal and child factors, including gestational age, year at blood collection, and race/ethnicity, were associated with cord and early childhood PFAS levels. The metabolome-wide association study and the targeted study identified several cord metabolites that may have been affected by in utero PFAS exposure. Conclusions: This pilot study found ubiquitous exposure to multiple PFAS in cord plasma (reflects in utero exposure) and in early childhood plasma (reflects both prenatal and postnatal exposure) among U.S. Black and Hispanic children. Metabolomic analysis suggests that in utero PFAS exposures may alter fetal metabolism. Future large-scale studies are needed to replicate the findings and further examine the associations of fetal PFAS exposure with long-term health outcomes and underlying metabolic pathways.

Boston Birth Cohort profile: rationale and study design.

In1998, the Boston Birth Cohort (BBC) was initiated at Boston Medical Center (BMC) in response to persistently high rates of preterm birth (PTB, defined as birth before 37 weeks of gestation) in the US population and the longstanding profound PTB disparity among Black, Indigenous, and people of color (BIPOC). The BBC encompasses two linked study protocols: The PTB Study serves as the baseline recruitment in the BBC. It aims to address fundamental questions about the causes and consequences of PTB. The study oversamples preterm babies using a case/control study design, in which cases are defined as mothers who deliver a preterm and/or low birthweight baby (<2500 grams regardless of gestational age). Controls are enrolled at a 2:1 control/case ratio and matched by maternal age (±5 years), self-reported race and ethnicity, and date of delivery (± 7 days for case delivery). From inception, it was designed as a comprehensive gene-environmental study of PTB. As a natural extension, the Children's Health Study, under a separate but linked Institutional Review Board protocol, is a longitudinal follow-up study of the participants who were recruited at birth in the PTB Study and who continue pediatric care at BMC. This linked model allows for investigation of early life origins of pediatric and chronic disease in a prospective cohort design. The BBC is one of the largest and longest National Institutes of Health-funded prospective birth cohort studies in the United States, consisting of 8733 mother-child dyads enrolled in the PTB Study at birth, and of those, 3592 children have been enrolled in the Children's Health Study, with a median follow-up of 14.5 years. The BBC mirrors the urban, underresourced, and underrepresented BIPOC population served by BMC. A high proportion of BBC children were born prematurely and had chronic health conditions (e.g., asthma, obesity, and elevated blood pressure) in childhood. The BBC's long-term goal has been to build a large, comprehensive database (epidemiological, clinical, and multiomics) and biospecimen repository to elucidate early life origins of pediatric and chronic diseases and identify modifiable upstream factors (e.g., psychosocial, environmental, and nutritional) to improve health across the life course for BIPOC mothers and children.