Faculty Bibliography

OBJECTIVE:To perform a detailed characterisation of diabetes burden and pre-diabetes risk in a rural county with previously documented poor health outcomes in order to understand the local within-county distribution of diabetes in rural areas of America. DESIGN, SETTING, AND PARTICIPANTS/METHODS:In 2021, we prospectively mailed health surveys to all households in Sullivan County, a rural county with the second-worst health outcomes of all counties in New York State. Our survey included questions on demographics, medical history and the American Diabetes Association's Pre-diabetes Risk Test. PRIMARY OUTCOME AND METHODS/UNASSIGNED:Our primary outcome was an assessment of diabetes burden within this rural county. To help mitigate non-response bias in our survey, raking adjustments were performed across strata of age, sex, race/ethnicity and health insurance. We analysed diabetes prevalence by demographic characteristics and used geospatial analysis to assess for clustering of diagnosed diabetes cases. RESULTS:After applying raking procedures for the 4725 survey responses, our adjusted diagnosed diabetes prevalence for Sullivan County was 12.9% compared with the 2019 Behavioural Risk Factor Surveillance System (BRFSS) estimate of 8.6%. In this rural area, diagnosed diabetes prevalence was notably higher among non-Hispanic Black (21%) and Hispanic (15%) residents compared with non-Hispanic White (12%) residents. 53% of respondents without a known history of pre-diabetes or diabetes scored as high risk for pre-diabetes. Nearest neighbour analyses revealed that hotspots of diagnosed diabetes were primarily located in the more densely populated areas of this rural county. CONCLUSIONS:Our mailed health survey to all residents in Sullivan County demonstrated higher diabetes prevalence compared with modelled BRFSS estimates that were based on small telephone samples. Our results suggest the need for better diabetes surveillance in rural communities, which may benefit from interventions specifically tailored for improving glycaemic control among rural residents.

OBJECTIVE/UNASSIGNED:To perform a geospatial analysis of food insecurity in a rural county known to have poor health outcomes and assess the effect of the COVID-19 pandemic. METHODS/UNASSIGNED:In 2020, we mailed a comprehensive cross-sectional survey to all households in Sullivan County, a rural county with the second-worst health outcomes among all counties in New York State. Surveys of households included validated food insecurity screening questions. Questions were asked in reference to 2019, prior to the pandemic, and for 2020, in the first year of the pandemic. Respondents also responded to demographic questions. Raking adjustments were performed using age, sex, race/ethnicity, and health insurance strata to mitigate non-response bias. To identify significant hotspots of food insecurity within the county, we also performed geospatial analysis. FINDINGS/UNASSIGNED:From the 28,284 households surveyed, 20% of households responded. Of 4725 survey respondents, 26% of households reported experiencing food insecurity in 2019, and in 2020, this proportion increased to 35%. In 2020, 58% of Black and Hispanic households reported experiencing food insecurity. Food insecurity in 2020 was also present in 58% of unmarried households with children and in 64% of households insured by Medicaid. The geospatial analyses revealed that hotspots of food insecurity were primarily located in or near more urban areas of the rural county. CONCLUSIONS/UNASSIGNED:Our countywide health survey in a high-risk rural county identified significant increases of food insecurity in the first year of the COVID-19 pandemic, despite national statistics reporting a stable rate. Responses to future crises should include targeted interventions to bolster food security among vulnerable rural populations.

Electronic health records (EHRs) contain rich clinical information for millions of patients and are increasingly used for public health research. However, non-random inclusion of subjects in EHRs can result in selection bias, with factors such as demographics, socioeconomic status, healthcare referral patterns, and underlying health status playing a role. While this issue has been well documented, little work has been done to develop or apply bias-correction methods, often due to the fact that most of these factors are unavailable in EHRs. To address this gap, we propose a series of Heckman type bias correction methods by incorporating social determinants of health selection covariates to model the EHR non-random sampling probability. Through simulations under various settings, we demonstrate the effectiveness of our proposed method in correcting biases in both the association coefficient and the outcome mean. Our method augments the utility of EHRs for public health inferences, as we show by estimating the prevalence of cardiovascular disease and its correlation with risk factors in the New York City network of EHRs.

INTRODUCTION:Traditional survey-based surveillance is costly, limited in its ability to distinguish diabetes types and time-consuming, resulting in reporting delays. The Diabetes in Children, Adolescents and Young Adults (DiCAYA) Network seeks to advance diabetes surveillance efforts in youth and young adults through the use of large-volume electronic health record (EHR) data. The network has two primary aims, namely: (1) to refine and validate EHR-based computable phenotype algorithms for accurate identification of type 1 and type 2 diabetes among youth and young adults and (2) to estimate the incidence and prevalence of type 1 and type 2 diabetes among youth and young adults and trends therein. The network aims to augment diabetes surveillance capacity in the USA and assess performance of EHR-based surveillance. This paper describes the DiCAYA Network and how these aims will be achieved. METHODS AND ANALYSIS:The DiCAYA Network is spread across eight geographically diverse US-based centres and a coordinating centre. Three centres conduct diabetes surveillance in youth aged 0-17 years only (component A), three centres conduct surveillance in young adults aged 18-44 years only (component B) and two centres conduct surveillance in components A and B. The network will assess the validity of computable phenotype definitions to determine diabetes status and type based on sensitivity, specificity, positive predictive value and negative predictive value of the phenotypes against the gold standard of manually abstracted medical charts. Prevalence and incidence rates will be presented as unadjusted estimates and as race/ethnicity, sex and age-adjusted estimates using Poisson regression. ETHICS AND DISSEMINATION:The DiCAYA Network is well positioned to advance diabetes surveillance methods. The network will disseminate EHR-based surveillance methodology that can be broadly adopted and will report diabetes prevalence and incidence for key demographic subgroups of youth and young adults in a large set of regions across the USA.

BACKGROUND:Global longitudinal strain (GLS) is a sensitive marker for identifying subclinical myocardial dysfunction in obstructive coronary artery disease (CAD). Little is known about the relationship between GLS and ischemia in patients with myocardial ischemia and no obstructive CAD (INOCA). OBJECTIVES/OBJECTIVE:To investigate the relationship between resting GLS and ischemia on stress echocardiography (SE) in patients with INOCA. METHODS:Left ventricular GLS was calculated offline on resting SE images at enrollment (n = 144) and 1-year follow-up (n = 120) in the CIAO-ISCHEMIA (Changes in Ischemia and Angina over One year in International Study of Comparative Health Effectiveness with Medical and Invasive Approaches trial screen failures with no obstructive CAD on computed tomography [CT] angiography) study, which enrolled participants with moderate or severe ischemia by local SE interpretation (≥3 segments with new or worsening wall motion abnormality and no obstructive (<50% stenosis) on coronary computed tomography angiography. RESULTS:Global longitudinal strain values were normal in 83.3% at enrollment and 94.2% at follow-up. Global longitudinal strain values were not associated with a positive SE at enrollment (GLS = -21.5% positive SE vs GLS = -19.9% negative SE, P = .443) or follow-up (GLS = -23.2% positive SE vs GLS = -23.1% negative SE, P = .859). Significant change in GLS was not associated with positive SE in follow-up (P = .401). Regional strain was not associated with colocalizing ischemia at enrollment or follow-up. Changes in GLS and number of ischemic segments from enrollment to follow-up showed a modest but not clinically meaningful correlation (β = 0.41; 95% CI, 0.16, 0.67; P = .002). CONCLUSIONS:In this cohort of INOCA patients, resting GLS values were largely normal and did not associate with the presence, severity, or location of stress-induced ischemia. These findings may suggest the absence of subclinical myocardial dysfunction detectable by echocardiographic strain analysis at rest in INOCA.

INTRODUCTION/UNASSIGNED:There is growing interest in using electronic health records (EHRs) for chronic disease surveillance. However, these data are convenience samples of in-care individuals, which are not representative of target populations for public health surveillance, generally defined, for the relevant period, as resident populations within city, state, or other jurisdictions. We focus on using EHR data for estimation of diabetes prevalence among young adults in New York City, as rising diabetes burden in younger ages call for better surveillance capacity. METHODS/UNASSIGNED:This article applies common nonprobability sampling methods, including raking, post-stratification, and multilevel regression with post-stratification, to real and simulated data for the cross-sectional estimation of diabetes prevalence among those aged 18-44 years. Within real data analyses, we externally validate city- and neighborhood-level EHR-based estimates to gold-standard estimates from a local health survey. Within data simulations, we probe the extent to which residual biases remain when selection into the EHR sample is non-ignorable. RESULTS/UNASSIGNED:Within the real data analyses, these methods reduced the impact of selection biases in the citywide prevalence estimate compared to gold standard. Residual biases remained at the neighborhood-level, where prevalence tended to be overestimated, especially in neighborhoods where a higher proportion of residents were captured in the sample. Simulation results demonstrated these methods may be sufficient, except when selection into the EHR is non-ignorable, depending on unmeasured factors or on diabetes status. CONCLUSIONS/UNASSIGNED:While EHRs offer potential to innovate on chronic disease surveillance, care is needed when estimating prevalence for small geographies or when selection is non-ignorable.

IMPORTANCE:Biomarkers may improve prediction of cardiovascular events for patients with stable coronary artery disease (CAD), but their importance in addition to clinical tests of inducible ischemia and CAD severity is unknown. OBJECTIVES:To evaluate the prognostic value of multiple biomarkers in stable outpatients with obstructive CAD and moderate or severe inducible ischemia. DESIGN AND SETTING:The ISCHEMIA and ISCHEMIA CKD trials randomized 5,956 participants with CAD to invasive or conservative management from July 2012 to January 2018; 1,064 participated in the biorepository. MAIN OUTCOME MEASURES:Primary outcome was cardiovascular death, myocardial infarction (MI), or hospitalization for unstable angina, heart failure, or resuscitated cardiac arrest. Secondary outcome was cardiovascular death or MI. Improvements in prediction were assessed by cause-specific hazard ratios (HR) and area under the receiver operating characteristics curve (AUC) for an interquartile increase in each biomarker, controlling for other biomarkers, in a base clinical model of risk factors, left ventricular ejection fraction (LVEF) and ischemia severity. Secondary analyses were performed among patients in whom core-lab confirmed severity of CAD was ascertained by computed cardiac tomographic angiography (CCTA). EXPOSURES:Baseline levels of interleukin-6 (IL-6), high sensitivity troponin T (hsTnT), growth differentiation factor 15 (GDF-15), N-terminal pro-B-type natriuretic peptide (NT-proBNP), lipoprotein a (Lp[a]), high sensitivity C-reactive protein (hsCRP), Cystatin C, soluble CD 40 ligand (sCD40L), myeloperoxidase (MPO), and matrix metalloproteinase 3 (MMP3). RESULTS:Among 757 biorepository participants, median (IQR) follow-up was 3 (2-5) years, age was 67 (61-72) years, and 144 (19%) were female; 508 had severity of CAD by CCTA available. In an adjusted multimarker model with hsTnT, GDF-15, NT-proBNP and sCD40L, the adjusted HR for the primary outcome per interquartile increase in each biomarker was 1.58 (95% CI 1.22, 2.205), 1.60 (95% CI 1.16, 2.20), 1.61 (95% 1.22, 2.14), and 1.46 (95% 1.12, 1.90), respectively. The adjusted multimarker model also improved prediction compared with the clinical model, increasing the AUC from 0.710 to 0.792 (P < .01) and 0.714 to 0.783 (P < .01) for the primary and secondary outcomes, respectively. Similar findings were observed after adjusting for core-lab confirmed atherosclerosis severity. CONCLUSIONS AND RELEVANCE:Among ISCHEMIA biorepository participants, biomarkers of myocyte injury/distension, inflammation, and platelet activity improved cardiovascular event prediction in addition to risk factors, LVEF, and assessments of ischemia and atherosclerosis severity. These biomarkers may improve risk stratification for patients with stable CAD.

Iron deficits have been reported as a risk factor for psychotic spectrum disorders (PSD). However, examinations of brain iron in PSD remain limited. The current study employed quantitative MRI to examine iron content in several iron-rich subcortical structures in 49 young adult individuals with PSD (15 schizophrenia, 17 schizoaffective disorder, and 17 bipolar disorder with psychotic features) compared with 35 age-matched healthy controls (HC). A parametric approach based on a two-pool magnetization transfer model was applied to estimate longitudinal relaxation rate (R₁), which reflects both iron and myelin, and macromolecular proton fraction (MPF), which is specific to myelin. To describe iron content, a synthetic effective transverse relaxation rate (R₂*) was modeled using a linear fitting of R₁ and MPF. PSD patients compared to HC showed significantly reduced R₁ and synthetic R₂* across examined regions including the pallidum, ventral diencephalon, thalamus, and putamen areas. This finding was primarily driven by decreases in the subgroup with schizophrenia, followed by schizoaffective disorder. No significant group differences were noted for MPF between PSD and HC while for regional volume, significant reductions in patients were only observed in bilateral caudate, suggesting that R₁ and synthetic R₂* reductions in schizophrenia and schizoaffective patients likely reflect iron deficits that either occur independently or precede structural and myelin changes. Subcortical R₁ and synthetic R₂* were also found to be inversely related to positive symptoms within the PSD group and to schizotypal traits across the whole sample. These findings that decreased iron in subcortical regions are associated with PSD risk and symptomatology suggest that brain iron deficiencies may play a role in PSD pathology and warrant further study.

BACKGROUND:The ISCHEMIA and the ISCHEMIA-CKD trials found no statistical difference in the primary clinical endpoint between initial invasive management and initial conservative management of patients with chronic coronary disease and moderate to severe ischemia on stress testing without or with advanced chronic kidney disease (CKD). In ISCHEMIA, there was numerically lower cardiovascular mortality but higher non-cardiovascular mortality with no significant difference in all-cause death with an initial invasive strategy when compared with a conservative strategy. However, an invasive strategy increased peri-procedural myocardial infarction (MI) but decreased spontaneous MI with continued separation of curves over time, which potentially may lead to reduced risk of cardiovascular and all-cause mortality. Thus, the long-term effect of invasive management strategy on mortality remains unclear. In ISCHEMIA-CKD, the treatment and cause-specific mortality rates were similar during follow-up. METHODS:Funded by the National Heart, Lung, and Blood Institute, the ISCHEMIA-EXTEND observational study is the long-term follow-up of surviving participants (projected median of 10 years) with chronic coronary disease from the ISCHEMIA trial. In the ISCHEMIA trial, 5,179 participants with moderate or severe stress-induced ischemia were randomized to initial invasive management with angiography, revascularization when feasible, and guideline-directed medical therapy (GDMT), or initial conservative management with GDMT alone and angiography reserved for failure of medical therapy. ISCHEMIA-CKD EXTEND is the long-term follow-up of surviving participants (projected median of 9 years) from the ISCHEMIA-CKD trial, a companion trial that included 777 patients with advanced CKD. Ascertainment of death will be conducted via direct participant contact, medical record review, and/or vital status registry search. The overarching objective of long-term follow-up is to assess whether there are between-group differences in long-term all-cause, cardiovascular, and non-cardiovascular mortality, and increase precision around the treatment effect estimates for risk of all-cause, cardiovascular, and non-cardiovascular mortality. We will conduct Bayesian survival modeling to take advantage of rich inferences using the posterior distribution of the treatment effect. CONCLUSIONS:The long-term effect of an initial invasive versus conservative strategy on all-cause, cardiovascular, and non-cardiovascular mortality will be assessed. The findings of ISCHEMIA-EXTEND and ISCHEMIA-CKD EXTEND will inform patients, practitioners, practice guidelines, and health policy.