Faculty Bibliography

PURPOSE/UNASSIGNED:Race disparities in the healthcare system and the resulting inequality in clinical data among different races hinder the ability to generate equitable prediction results. This study aims to reduce healthcare disparities arising from data imbalance by leveraging advanced transfer learning (TL) methods. METHOD/UNASSIGNED:We examined the ophthalmic healthcare disparities at a population level using electronic medical records data from a study cohort (N = 785) receiving care at an academic institute. Regression-based TL models were usesd, transferring valuable information from the dominant racial group (White) to improve visual field mean deviation (MD) rate of change prediction particularly for data-disadvantaged African American (AA) and Asian racial groups. Prediction results of TL models were compared with two conventional approaches. RESULTS/UNASSIGNED:Disparities in socioeconomic status and baseline disease severity were observed among the AA and Asian racial groups. The TL approach achieved marked to comparable improvement in prediction accuracy compared to the two conventional approaches as evident by smaller mean absolute errors or mean square errors. TL identified distinct key features of visual field MD rate of change for each racial group. CONCLUSIONS/UNASSIGNED:The study introduces a novel application of TL that improved reliability of the analysis in comparison with conventional methods, especially in small sample size groups. This can improve assessment of healthcare disparity and subsequent remedy approach. TRANSLATIONAL RELEVANCE/UNASSIGNED:TL offers an equitable and efficient approach to mitigate healthcare disparities analysis by enhancing prediction performance for data-disadvantaged group.

PURPOSE/UNASSIGNED:stimation, especially designed for longitudinal ophthalmic studies. LIMBARE accommodates repeated measurements from both eyes and overtime, and effectively address the presence of outliers. METHODS/UNASSIGNED:The model setup of LIMBARE and computing algorithm for point and confidence interval estimates of the breakpoint was introduced. The performance of LIMBARE and other competing methods was assessed via comprehensive simulation studies and application to a longitudinal ophthalmic study with 216 eyes (145 subjects) followed for an average of 3.7±1.3 years to examine the longitudinal association between structural and functional measurements. RESULTS/UNASSIGNED:In simulation studies, LIMBARE showed the smallest bias and mean squared error (MSE) for estimating the breakpoint, with empirical coverage probability of corresponding CI estimate closest to the nominal level for scenarios with and without outlier data points. In the application to the longitudinal ophthalmic study, LIMBARE detected two breakpoints between visual field mean deviation (MD) and retinal nerve fiber layer thickness (RNFL) and one breakpoint between MD and cup to disc ratio (CDR), while the cross-sectional analysis approach only detected one and none, respectively. CONCLUSIONS/UNASSIGNED:LIMBARE enhances breakpoint estimation accuracy in longitudinal ophthalmic studies, while cross-sectional analysis approach is not recommended for future studies. TRANSLATIONAL RELEVANCE/UNASSIGNED:Our proposed method and companion software R package provides a valuable computational tool for advancing longitudinal ophthalmology research and exploring the association relationships between ophthalmic variables.

BACKGROUND:It is critical to assess implementation fidelity of evidence-based interventions and factors moderating fidelity, to understand the reasons for their success or failure. However, fidelity and fidelity moderators are seldom systematically reported. The study objective was to conduct a concurrent implementation fidelity evaluation and examine fidelity moderators of CHORD (Community Health Outreach to Reduce Diabetes), a pragmatic, cluster-randomized, controlled trial to test the impact of a Community Health Workers (CHW)-led health coaching intervention to prevent incident type 2 Diabetes Mellitus in New York (NY). METHODS:We applied the Conceptual Framework for Implementation Fidelity to assess implementation fidelity and factors moderating it across the four core intervention components: patient goal setting, education topic coaching, primary care (PC) visits, and referrals to address social determinants of health (SDH), using descriptive statistics and regression models. PC patients with prediabetes receiving care from safety-net patient-centered medical homes (PCMHs) at either, VA NY Harbor or at Bellevue Hospital (BH) were eligible to be randomized into the CHW-led CHORD intervention or usual care. Among 559 patients randomized and enrolled in the intervention group, 79.4% completed the intake survey and were included in the analytic sample for fidelity assessment. Fidelity was measured as coverage, content adherence and frequency of each core component, and the moderators assessed were implementation site and patient activation measure. RESULTS:Content adherence was high for three components with nearly 80.0% of patients setting ≥ 1 goal, having ≥ 1 PC visit and receiving ≥ 1 education session. Only 45.0% patients received ≥ 1 SDH referral. After adjusting for patient gender, language, race, ethnicity, and age, the implementation site moderated adherence to goal setting (77.4% BH vs. 87.7% VA), educational coaching (78.9% BH vs. 88.3% VA), number of successful CHW-patient encounters (6 BH vs 4 VA) and percent of patients receiving all four components (41.1% BH vs. 25.7% VA). CONCLUSIONS:The fidelity to the four CHORD intervention components differed between the two implementation sites, demonstrating the challenges in implementing complex evidence-based interventions in different settings. Our findings underscore the importance of measuring implementation fidelity in contextualizing the outcomes of randomized trials of complex multi-site behavioral interventions. TRIAL REGISTRATION:The trial was registered with ClinicalTrials.gov on 30/12/2016 and the registration number is NCT03006666 .

OBJECTIVES/OBJECTIVE:Autoantibody seroconversion has been extensively studied in the context of COVID-19 infection but data regarding post-vaccination autoantibody production is lacking. Here we aimed to determine the incidence of common autoantibody formation following mRNA COVID-19 vaccines in patients with inflammatory arthritis (IA) and in healthy controls. METHODS:Autoantibody seroconversion was measured by serum ELISA in a longitudinal cohort of IA participants and healthy controls before and after COVID-19 mRNA-based immunization. RESULTS:Overall, there was a significantly lower incidence of ANA seroconversion in participants who did not contract COVID-19 prior to vaccination compared with those who been previously infected (7.4% vs 24.1%, p= 0.014). Incidence of de novo anti-cyclic citrullinated protein (CCP) seroconversion in all participants was low at 4.9%. Autoantibody levels were typically of low titer, transient, and not associated with increase in IA flares. CONCLUSIONS:In both health and inflammatory arthritis, the risk of autoantibody seroconversion is lower following mRNA-based immunization than following natural SARS-CoV-2 infection. Importantly, seroconversion does not correlate with self-reported IA disease flare risk, further supporting the encouragement of mRNA-based COVID-19 immunization in the IA population.

INTRODUCTION:Psoriatic arthritis (PsA) is a complex, immune-mediated disease associated with skin psoriasis that, if left untreated, can lead to joint destruction. Up to 30% of patients with psoriasis progress to PsA. In most cases, psoriasis precedes synovio-entheseal inflammation by an average of 5-7 years, providing a unique opportunity for early and potentially preventive intervention in a susceptible and identifiable population. Guselkumab is an effective IL-23p19 inhibitor Food and Drug Administration (FDA-approved for treatment of moderate-to-severe psoriasis and PsA. The Preventing Arthritis in a Multicentre Psoriasis At-Risk cohort (PAMPA) study aims to evaluate the efficacy of guselkumab in preventing PsA and decreasing musculoskeletal power Doppler ultrasound (PDUS) abnormalities in a population of patients with psoriasis who are at-increased risk for PsA progression. METHODS AND ANALYSIS:The PAMPA study is a multicentre, randomised, double-blind, placebo-controlled, interventional, preventive trial comparing PDUS involvement and conversion to PsA in patients with psoriasis at-increased risk for progression treated with guselkumab compared with non-biological standard of care. The study includes a screening period, a double-blind treatment period (24 weeks) and an open-label follow-up period (72 weeks). At baseline, 200 subjects will be randomised (1:1) to receive either guselkumab 100 mg (arm 1) or placebo switching to guselkumab 100 mg starting at week 24 (arm 2). Arm 3 will follow 150 at-risk psoriasis patients who decline biological therapy and randomisation. Changes from baseline in the PDUS score at week 24 and the difference in proportion of patients transitioning to PsA at 96 weeks will be examined as the coprimary endpoints. ETHICS AND DISSEMINATION:Ethics approval for this study was granted by the coordinating centre's (NYU School of Medicine) Institutional Review Board (IRB). Each participating site received approval through their own IRBs. The findings will be shared in peer-reviewed articles and scientific conference presentations. TRIAL REGISTRATION NUMBER:NCT05004727.

Recent studies have suggested that the temporal dynamics of the human microbiome may have associations with human health and disease. An increasing number of longitudinal microbiome studies, which record time to disease onset, aim to identify candidate microbes as biomarkers for prognosis. Owing to the ultra-skewness and sparsity of microbiome proportion (relative abundance) data, directly applying traditional statistical methods may result in substantial power loss or spurious inferences. We propose a novel joint modeling framework [JointMM], which is comprised of two sub-models: a longitudinal sub-model called zero-inflated scaled-Beta generalized linear mixed-effects regression to depict the temporal structure of microbial proportions among subjects; and a survival sub-model to characterize the occurrence of an event and its relationship with the longitudinal microbiome proportions. JointMM is specifically designed to handle the zero-inflated and highly skewed longitudinal microbial proportion data and examine whether the temporal pattern of microbial presence and/or the non-zero microbial proportions are associated with differences in the time to an event. The longitudinal sub-model of JointMM also provides the capacity to investigate how the (time-varying) covariates are related to the temporal microbial presence/absence patterns and/or the changing trend in non-zero proportions. Comprehensive simulations and real data analyses are used to assess the statistical efficiency and interpretability of JointMM. This article is protected by copyright. All rights reserved.

BACKGROUND:With the rapid accumulation of microbiome-wide association studies, a great amount of microbiome data are available to study the microbiome's role in human disease and advance the microbiome's potential use for disease prediction. However, the unique features of microbiome data hinder its utility for disease prediction. METHODS:Motivated from the polygenic risk score framework, we propose a microbial risk score (MRS) framework to aggregate the complicated microbial profile into a summarized risk score that can be used to measure and predict disease susceptibility. Specifically, the MRS algorithm involves two steps: (1) identifying a sub-community consisting of the signature microbial taxa associated with disease and (2) integrating the identified microbial taxa into a continuous score. The first step is carried out using the existing sophisticated microbial association tests and pruning and thresholding method in the discovery samples. The second step constructs a community-based MRS by calculating alpha diversity on the identified sub-community in the validation samples. Moreover, we propose a multi-omics data integration method by jointly modeling the proposed MRS and other risk scores constructed from other omics data in disease prediction. RESULTS:Through three comprehensive real-data analyses using the NYU Langone Health COVID-19 cohort, the gut microbiome health index (GMHI) multi-study cohort, and a large type 1 diabetes cohort separately, we exhibit and evaluate the utility of the proposed MRS framework for disease prediction and multi-omics data integration. In addition, the disease-specific MRSs for colorectal adenoma, colorectal cancer, Crohn's disease, and rheumatoid arthritis based on the relative abundances of 5, 6, 12, and 6 microbial taxa, respectively, are created and validated using the GMHI multi-study cohort. Especially, Crohn's disease MRS achieves AUCs of 0.88 (0.85-0.91) and 0.86 (0.78-0.95) in the discovery and validation cohorts, respectively. CONCLUSIONS:The proposed MRS framework sheds light on the utility of the microbiome data for disease prediction and multi-omics integration and provides a great potential in understanding the microbiome's role in disease diagnosis and prognosis. Video Abstract.

BACKGROUND/UNASSIGNED:With the rapid accumulation of microbiome-wide association studies, a great amount of microbiome data are available to study the microbiome's role in human disease and advance the microbiome's potential use for disease prediction. However, the unique features of microbiome data hinder its utility for disease prediction. METHODS/UNASSIGNED:Motivated from the polygenic risk score framework, we propose a microbial risk score (MRS) framework to aggregate the complicated microbial profile into a summarized risk score that can be used to measure and predict disease susceptibility. Specifically, the MRS algorithm involves two steps: 1) identifying a sub-community consisting of the signature microbial taxa associated with disease, and 2) integrating the identified microbial taxa into a continuous score. The first step is carried out using the existing sophisticated microbial association tests and pruning and thresholding method in the discovery samples. The second step constructs a community-based MRS by calculating alpha diversity on the identified sub-community in the validation samples. Moreover, we propose a multi-omics data integration method by jointly modeling the proposed MRS and other risk scores constructed from other omics data in disease prediction. RESULTS/UNASSIGNED:Through three comprehensive real data analyses using the NYU Langone Health COVID-19 cohort, the gut microbiome health index (GMHI) multi-study cohort, and a large type 1 diabetes cohort separately, we exhibit and evaluate the utility of the proposed MRS framework for disease prediction and multi-omics data integration. In addition, the disease-specific MRSs for colorectal adenoma, colorectal cancer, Crohn's disease, and rheumatoid arthritis based on the relative abundances of 5, 6, 12, and 6 microbial taxa respectively are created and validated using the GMHI multi-study cohort. Especially, Crohn's disease MRS achieves AUCs of 0.88 ([0.85-0.91]) and 0.86 ([0.78-0.95]) in the discovery and validation cohorts, respectively. CONCLUSIONS/UNASSIGNED:The proposed MRS framework sheds light on the utility of the microbiome data for disease prediction and multi-omics integration, and provides great potential in understanding the microbiome's role in disease diagnosis and prognosis.