Faculty Bibliography

BACKGROUND:Chronic stress in African American individuals is multilayered amid the context of experiences of racism and discrimination. Cortisol dynamic range (CDR) may be an indicator of chronic stress, but CDR is understudied in African American populations compared with White populations, and is hypothesized to differ by sex. OBJECTIVE:Using a community-engaged approach within the Fostering African-American Improvement in Total Health! (FAITH!) Heart Health+ ancillary study, we assessed the feasibility of participant-centric CDR collection, and its association with measures for individual, interpersonal, and structural stress and exposure to racism in medically underserved African American women and men. METHODS:Participants residing in the Minneapolis-St Paul and Rochester, Minnesota areas provided survey data (everyday discrimination, perceived stress, mood, sleep quality, and high effort coping measures), and saliva samples (morning and afternoon) via at-home, self-collection kits for cortisol measurement. CDR was calculated as a difference in log cortisol levels (ie, log of the cortisol diurnal peak-to-nadir ratio). Geospatial Area Deprivation Index and the distance lived from George Floyd Square in Minneapolis were calculated. Linear regression examined the association between CDR and outcome variables. RESULTS:Of consented participants (n=53), 70% (37/53) provided cortisol samples. The final analytic sample included 32 participants with complete and physiological diurnal cortisol curves (mean age 57.5 years, 62.5% [20/32] women). Lower (less dynamic) CDR in women (n=20) was associated with greater perceived stress (β=-0.07, P=.01), greater anxiety (β=-0.06, P=.01), higher Superwoman Schema score (β=-0.02, P=.04), and greater distance from George Floyd Square (β=-0.02, P=.01). No associations were observed in men (P>.05). CONCLUSIONS:The current results suggest that CDR from participant-led saliva collection is feasible and may serve as a biomarker of chronic and physiological stress in African American women, particularly those residing in underresourced areas.

BACKGROUND:In many parts of the world including India, the prevalence of autoimmune inflammatory diseases such as neuromyelitis optica spectrum disorders (NMOSD), myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD), and multiple sclerosis (MS) is rising. A diagnosis is often delayed due to insufficient diagnostic tools. Machine learning (ML) models have accurately differentiated eyes of patients with MS from those of healthy controls (HCs) using optical coherence tomography (OCT)-based retinal images. Examining OCT characteristics may allow for early differentiation of these conditions. The objective of this study was to determine feasibility of ML analyses to distinguish between patients with different autoimmune inflammatory diseases, other ocular diseases, and HCs based on OCT measurements of the peripapillary retinal nerve fiber layer (pRNFL), ganglion cell-inner plexiform layer (GCIPL), and inner nuclear layers (INLs). METHODS:Eyes of people with MS (n = 99 patients), NMOSD (n = 40), MOGAD (n = 74), other ocular diseases (OTHER, n = 16), and HCs (n = 54) from the Mangalore Demyelinating Disease Registry were included. Support vector machine (SVM) classification models incorporating age, pRNFL, GCIPL, and INL were performed. Data were split into training (70%) and testing (30%) data and accounted for within-patient correlations. Cross-validation was used in training to choose the best parameters for the SVM model. Accuracy and area under receiver operating characteristic curves (AUROCs) were used to assess model performance. RESULTS:The SVM models distinguished between eyes of patients with each condition (i.e., MOGAD vs NMOSD, NMOSD vs HC, MS vs OTHER, etc) with strong discriminatory power demonstrated from the AUROCs for these comparisons ranging from 0.81 to 1.00. These models also performed with moderate to high accuracy, ranging from 0.66 to 0.81, with the exception of the MS vs NMOSD comparison, which had an accuracy of 0.53. CONCLUSIONS:ML models are useful for distinguishing between autoimmune inflammatory diseases and for distinguishing these from HCs and other ocular diseases based on OCT measures. This study lays the groundwork for future deep learning studies that use analyses of raw OCT images for identifying eyes of patients with such disorders and other etiologies of optic neuropathy.

OBJECTIVE:While timely interventions can improve medication adherence, it is challenging to identify which patients are at risk of nonadherence at point-of-care. We aim to develop and validate flexible machine learning (ML) models to predict a continuous measure of adherence to guideline-directed medication therapies (GDMTs) for heart failure (HF). MATERIALS AND METHODS/METHODS:We utilized a large electronic health record (EHR) cohort of 34,697 HF patients seen at NYU Langone Health with an active prescription for ≥1 GDMT between April 01, 2021 and October 31, 2022. The outcome was adherence to GDMT measured as proportion of days covered (PDC) at 6 months following a clinical encounter. Over 120 predictors included patient-, therapy-, healthcare-, and neighborhood-level factors guided by the World Health Organization's model of barriers to adherence. We compared performance of several ML models and their ensemble (superlearner) for predicting PDC with traditional regression model (OLS) using mean absolute error (MAE) averaged across 10-fold cross-validation, % increase in MAE relative to superlearner, and predictive-difference across deciles of predicted PDC. RESULTS:Superlearner, a flexible nonparametric prediction approach, demonstrated superior prediction performance. Superlearner and quantile random forest had the lowest MAE (mean [95% CI] = 18.9% [18.7%-19.1%] for both), followed by MAEs for quantile neural network (19.5% [19.3%-19.7%]) and kernel support vector regression (19.8% [19.6%-20.0%]). Gradient boosted trees and OLS were the 2 worst performing models with 17% and 14% higher MAEs, respectively, relative to superlearner. Superlearner demonstrated improved predictive difference. CONCLUSION/CONCLUSIONS:This development phase study suggests potential of linked EHR-pharmacy data and ML to identify HF patients who will benefit from medication adherence interventions. DISCUSSION/CONCLUSIONS:Fairness evaluation and external validation are needed prior to clinical integration.

OBJECTIVE:Early childhood advanced primary care models are promising ways of addressing child and family needs, but there is limited evidence to support short-term sustainability within current Medicaid payment structures. We evaluate claims-based outcomes associated with 3-2-1 IMPACT (IMPACT), an early childhood advanced primary care model, compared with the standard of care. METHODS:Using New York State Medicaid claims, we identified and matched children aged 1 to 35 months receiving care at 3 IMPACT sites and 3 comparison sites within a large public hospital system. Regression models were used to analyze use, expenditure, enrollment, and quality outcomes between groups. RESULTS:There were 6045 children at the treatment sites and 4832 matched children from the comparison sites. IMPACT was associated with a significant increase in 6 or more well-child visits and a decrease in emergency department visit rates. There was also a significant increase in 6 more well-child visits specifically for Black and Hispanic children seeking care at IMPACT sites compared with comparison sites. There were no significant differences in expenditures, other use types, or Medicaid enrollment across groups. CONCLUSION/CONCLUSIONS:An early childhood advanced primary care model that incorporates multiple evidence-based programs can show short-term, positive effects on preventative and acute care use and quality within Medicaid. These results highlight short-term strategies for sustainability while awaiting the long-term, cross-sector benefits expected from models like IMPACT. Future studies addressing additive model component effects and longer-term outcomes across mother-child dyadic and social-emotional outcomes are warranted.

BACKGROUND:Lifesaving medications for opioid use disorder (MOUD) exist; however, most people with opioid use disorder (OUD) do not receive treatment. Hospitalization is one important opportunity to engage people with OUD and offer treatment, including MOUD. Between 2018 and 2020, 6 public hospitals in New York City launched the "Consult for Addiction Treatment and Care in Hospitals" (CATCH) program to provide interprofessional addiction consult services to hospitalized patients. METHODS:This qualitative study aims to add perspectives from 30 racially and ethnically diverse people with opioid-related diagnoses who were hospitalized at a CATCH hospital between October 2019 and April 2021. We used purposive sampling to recruit demographically diverse individuals who accepted or declined aspects of CATCH services. Interviews were audio-recorded, transcribed, and coded for emergent themes using grounded theory techniques. The framework of structural vulnerability was utilized to highlight how social context impacts patients' experiences of healthcare, and in turn affects their addiction trajectories. RESULTS:Participants overwhelmingly accepted MOUD to manage withdrawal symptoms during hospitalization, and many planned to continue MOUD after discharge. Participants appreciated the interprofessional support of CATCH teams which included medical providers, social workers, addiction counselors, and peers. While participants felt that CATCH made holistic addiction treatment including MOUD more accessible, structural issues created barriers to continuing treatment long term. Some participants still felt stigmatized or "punished" for their drug use by non-CATCH providers. CONCLUSION/CONCLUSIONS:CATCH met an urgent need for nonjudgmental care and medical management of opioid withdrawal. Additional interventions that address broader needs, including housing and social supports, as well as trust-building healthcare encounters for patients who have been historically marginalized, are needed to meet the public health goal of preventing overdose and reducing drug-related morbidity for this population.

BACKGROUND AND OBJECTIVE/OBJECTIVE:Prostate cancer (PC) is the second most common cancer and a leading cause of death among males. In this systematic review we evaluated cohort studies and randomized controlled trials (RCTs) on the relationship between dietary patterns and PC risk, progression, mortality, and biomarkers. METHODS:A systematic search of MEDLINE, Embase, and Cochrane Central was conducted through June 2024 according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. A total of 63 studies (49 cohort studies, 14 RCTs reports) examining dietary patterns and PC outcomes were included. Study quality was assessed using Critical Appraisal Skills Programme checklists. KEY FINDINGS AND LIMITATIONS/UNASSIGNED:Among males without PC at baseline, plant-based and healthy dietary patterns (eg, higher Healthy Eating Index, lower dietary inflammatory and hyperinsulinemic scores) were generally associated with lower total PC risk. Among patients with PC, Mediterranean, plant-based, and low-inflammatory diets were more consistently linked to lower risk of progression and PC-specific mortality. RCTs testing various diet patterns showed mixed effects on prostate-specific antigen or tumor markers. Limitations include variations in diet definitions, outcomes, and follow-up duration, and residual confounding. CONCLUSIONS AND CLINICAL IMPLICATIONS/CONCLUSIONS:Healthy dietary patterns that support cardiometabolic health may also benefit PC prevention and management. While evidence appears stronger for diet in slowing PC progression after diagnosis, the impact of diet on reducing the risk of other PC outcomes should not be overlooked (eg, risk of developing PC or risk of PC death). Integrated strategies are needed to promote healthy eating, particularly for patients at risk of PC progression, as this population often has higher risk of cardiovascular disease and metabolic disorders such as diabetes.

Prostate-specific membrane antigen (PSMA) PET is a powerful tool for prostate cancer staging and restaging, providing higher sensitivity and specificity than conventional imaging. The recognition of interpretive pitfalls led to the development of various scoring systems and frameworks, which in turn created challenges for consistent interpretation. The Prostate Cancer Molecular Imaging Standardized Evaluation (PROMISE) version 2 classification integrates the five-point PRIMARY score for assessing local disease, the molecular imaging TNM stage for disease extent, and the PSMA expression score to assess eligibility for PSMA-targeted radioligand therapy. The PSMA Reporting and Data System (PSMA-RADS) classifies PSMA PET/CT findings on the basis of the likelihood of presence of prostate cancer. For assessing therapy response, PSMA PET Progression (PPP) criteria focus on new lesions and clinical or biochemical progression, whereas Response Evaluation Criteria in PSMA PET/CT (RECIP 1.0) assess new lesions and changes in total PSMA-positive total tumor volume. The European Association of Nuclear Medicine (EANM) E-PSMA guideline and EANM-Society of Nuclear Medicine and Molecular Imaging procedure guidelines provide standardized reporting recommendations, incorporating elements from existing systems such as PROMISE, PSMA-RADS, and PPP. Nevertheless, such systems can be essential for optimizing prostate cancer management and facilitating communication among imaging professionals, clinicians, and patients. This article outlines these systems and discusses potential strengths and weaknesses.

BACKGROUND/UNASSIGNED:Despite their efficacy, medications for opioid use disorder (MOUD) remain underutilized in patients with infections from intravenous opioid use (I-IOU). This study evaluates the impact of an Expanded MOUD Access Initiative (EMAI) on MOUD uptake and other clinical outcomes in patients hospitalized for I-IOU at an institution without addiction medicine consultation. METHODS/UNASSIGNED:We performed a retrospective pre-post study of hospital admissions for I-IOU before (January 2019-June 2021) and after (January 2022-December 2023) EMAI introduction. Data was collected via chart review. The EMAI eliminated restrictions on methadone use and established a new order set for buprenorphine inductions. The primary outcome was MOUD receipt; secondary outcomes included patient directed discharge (PDD) and 30-day re-hospitalization. RESULTS/UNASSIGNED:There were 129 hospitalizations prior to the intervention (control) and 98 after (EMAI). MOUD receipt was significantly higher in the EMAI group (75.5% vs 31.0%; OR, 6.86 [95% CI, 3.84-12.61]). In patients not receiving MOUD prior to admission (n = 176), new inductions occurred more frequently in the EMAI group (68.0% vs 11.9%; OR, 15.76 [95% CI, 7.50-35.78]). PDD was lower in the EMAI group (23.5% vs 48.8%; OR, 0.32 [95% CI, 0.10-0.57]), as was 30-day re-hospitalization (12.2% vs 22.5%; OR, 0.48 [95% CI, 0.22-0.98]). In a multivariable logistic regression model, the EMAI was the only variable to show a statistically significant association with MOUD receipt (aOR, 6.89 [95% CI, 3.75-13.11]). CONCLUSIONS/UNASSIGNED:The EMAI was associated with increased MOUD uptake, reduced PDD, and fewer 30-day re-hospitalizations despite the lack of addiction medicine consultation.

BACKGROUND/UNASSIGNED:; environmental injustice) by racial/ethnic segregation (social injustice) on dementia diagnosis in ESKD. METHODS/UNASSIGNED:concentrations (annualized and matched to older adults' residential ZIP code at dialysis initiation) and by segregation scores (Theil's H method). FINDINGS/UNASSIGNED:and segregation. INTERPRETATION/UNASSIGNED:experienced an increased risk of dementia; this risk was particularly pronounced among individuals in high segregation and predominantly minority neighborhoods. Environmental and social injustices likely drive racial and ethnic disparities in dementia for older adults with ESKD, underscoring the need for interventions and policies to mitigate these injustices. FUNDING/UNASSIGNED:National Institutes of Health.

Epigenetic aging in early life remains poorly characterized, and patterns of gene expression can provide biologically meaningful insights. Blood DNA methylation was measured using the Illumina EPICv1.0 array and RNA sequencing was performed in blood in 174 adolescent participants (age range: 14-15 years) from the CHAMACOS cohort. Thirteen widely used epigenetic clocks were calculated, and their associations with transcriptome-wide RNA expression were tested using the limma-voom pipeline. We found evidence for substantial shared associations with RNA expression between different epigenetic clocks, including differential expression of MYO6 and ZBTB38 across five clocks. The epiTOC2, principal component (PC) PhenoAge, Hannum, PedBE and PC Hannum clocks were associated with differential expression of the highest number of RNAs, exhibiting associations with 22, 8, 5, 3, and 2 transcripts respectively. Generally, biological clocks were associated with differential expression of more genes than chronological clocks, and PC clocks were associated with differential expression of more genes relative to their CpG-trained counterparts. A total of 17 associations in our study were replicated in an independent adult sample (age range: 40-54 years). Our findings support the biological relevance of epigenetic clocks in adolescents and provide direction for selection of epigenetic ageing biomarkers in adolescent research.