Abstract P368: Neighborhoods and Subclinical Cardiovascular Disease

Applying Framingham Risk Score 2008 to Predict the 10-Year Risk of Cardiovascular Disease in a Group of Office Workers in Hanoi, Vietnam

Mobile location data has emerged as a valuable data source for studying human mobility patterns in various contexts, including virus spreading, urban planning, and hazard evacuation. However, these data are often anonymized overviews derived from a panel of traced mobile devices, and the representativeness of these panels is not well documented. Without a clear understanding of the data representativeness, the interpretations of research based on mobile location data may be questionable. This article presents a comprehensive examination of the potential biases associated with mobile location data using SafeGraph Patterns data in the United States as a case study. The research rigorously scrutinizes and documents the bias from multiple dimensions, including spatial, temporal, urbanization, demographic, and socioeconomic, over a five-year period from 2018 to 2022 across diverse geographic levels, including state, county, census tract, and census block group. Our analysis of the SafeGraph Patterns dataset revealed an average sampling rate of 7.5% with notable temporal dynamics, geographic disparities, and urban-rural differences. The number of sampled devices was strongly correlated with the census population at the county level over the five years for both urban (r > 0.97) and rural counties (r > 0.91), but less so at the census tract and block group levels. We observed minor sampling biases among groups such as gender, age, and moderate-income, with biases typically ranging from -0.05 to +0.05. However, minority groups such as Hispanic populations, low-income households, and individuals with low levels of education generally exhibited higher levels of underrepresentation bias that varied over space, time, urbanization, and across geographic levels. These findings provide important insights for future studies that utilize SafeGraph data or other mobile location datasets, highlighting the need to thoroughly evaluate the spatiotemporal dynamics of the bias across spatial scales when employing such data sources.

BackgroundAssessments of environmental exposure and health risks that utilize Geographic Information Systems (GIS) often make simplifying assumptions when using: (a) one or more discrete buffer distances to define the spatial extent of impacted regions, and (b) aggregated demographic data at the level of census enumeration units to derive the characteristics of the potentially exposed population. A case-study of school children in Orange County, Florida, is used to demonstrate how these limitations can be overcome by the application of cumulative distribution functions (CDFs) and individual geocoded locations. Exposure potential for 159,923 school children was determined at the childrens' home residences and at school locations by determining the distance to the nearest gasoline station, stationary air pollution source, and industrial facility listed in the Toxic Release Inventory (TRI). Errors and biases introduced by the use of discrete buffer distances and data aggregation were examined.ResultsThe use of discrete buffers distances in proximity-based exposure analysis introduced substantial bias in terms of determining the potentially exposed population, and the results are strongly dependent on the choice of buffer distance(s). Comparisons of exposure potential between home and school locations indicated that different buffer distances yield different results and contradictory conclusions. The use of a CDF provided a much more meaningful representation and is not based on the a-priori assumption that any particular distance is more relevant than another. The use of individual geocoded locations also provided a more accurate characterization of the exposed population and allowed for more reliable comparisons among sub-groups. In the comparison of children's home residences and school locations, the use of data aggregated at the census block group and tract level introduced variability as well as bias, leading to incorrect conclusions as to whether exposure potential was higher at school or at home.ConclusionThe use of CDFs in distance-based environmental exposure assessment provides more robust results than the use of discrete buffer distances. Unless specific circumstances warrant the use of discrete buffer distances, their applcation should be discouraged in favor of CDFs. The use of aggregated data at the census tract or block group level introduces substantial bias in environmental exposure assessment, which can be reduced through individual geocoding. The use of aggregation should be minimized when individual-level data are available. Existing GIS analysis techniques are well suited to determine CDFs as well as reliably geocode large datasets, and computational issues do not present a barrier for their more widespread use in environmental exposure and risk assessment.

Linking neighborhood- and patient-level data provides valuable information about the influence of upstream social determinants of health (SDOH). However, sharing of these data across health systems presents challenges. We set out to develop a pipeline to acquire, deidentify, and share neighborhood-level SDOH data across multiple health systems. We created a pipeline centered around Decentralized Geomarker Assessment for Multi-Site Studies (DeGAUSS) that utilizes containerization to geocode patient addresses and obtain neighborhood-level SDOH variables. We compared DeGAUSS to a third-party vendor geocoding tool available at Duke Health using a cohort of adult patients referred for abdominal transplant from January 1, 2016, to December 31, 2022. We calculated Cohen's Kappa and percent disagreement at census block group and tract levels, and by Area Deprivation Index, urbanicity, and year. The pipeline successfully generated SDOH data for 97.8% of addresses. There was high concordance between DeGAUSS and the vendor tool at the census block group (0.93) and tract levels (0.95). At the block group level, disagreement proportion differed by year and urbanicity, with larger disagreement in the rural category than in micropolitan and metropolitan categories (13%, 7%, 6.2%, respectively). We describe a novel pipeline that can facilitate the secure acquisition and sharing of neighborhood-level SDOH without sharing PHI. The pipeline can be scaled to include additional social, climate, and environmental variables, and can be extended to an unlimited number of health systems.

Socioeconomic status (SES) predicts coronary heart disease independently of the Framingham risk-scoring factors included in cholesterol treatment guidelines, possibly resulting in undertreatment of lower SES persons. We examined whether hybrid SES measures (based on area measures of income and individual education) address this bias and derived an approach to incorporating SES information into treatment guidelines. The Atherosclerosis Risk in Communities study data (initiated in 1987 with a 10-year follow-up of 15,495 adults aged 45 to 64 years in 4 southern and midwestern communities) were used to assess the calibration bias of 4 Cox models predicting 10-year coronary heart disease risk: Framingham risk score alone, and Framingham risk score plus SES using an individual-based measure (income less than 150% federal poverty level or less then 12 years of schooling), and 2 hybrid SES measures substituting area-based income measures (block group or zip code median incomes of less than 25th national percentiles) for the individual income component. Revised cholesterol treatment thresholds based on SES risk were also derived. Use of either the block group hybrid or individual-based SES measures eliminated the significant SES bias observed using Framingham risk score alone. Cholesterol treatment guideline thresholds of 10% and 20% coronary heart disease risk (based on the Framingham risk score) were reduced to 6% and 13% for those with low SES. Using patient income based on block group and individual education minimizes the SES bias in Framingham risk scoring and suggests more aggressive cholesterol treatment thresholds for low-SES persons.

To determine the association between altitude and the Framingham risk score in the Peruvian population, we performed a cross-sectional analytical study of data collected by the 2017–2018 Food and Nutrition Surveillance by Life Stages survey. The outcome of this study was the Framingham 10-year cardiovascular disease event risk prediction, which is composed of six modifiable and non-modifiable coronary risk factors. A generalized linear model (GLM) of the gamma family and log link function was used to report the crude and adjusted β coefficients. Several sensitivity analyses were performed to assess the association of interest. Data from a total of 833 surveyed participants were included. After adjusting for educational level, poverty level, alcohol consumption, physical activity level, the presence of any limitation, obesity, and area of residence, it was observed that altitude ≥ 2500 m above sea level (β = −0.42 [95% CI: −0.69 to −0.16]) was negatively and significantly associated with a decrease in the Framingham 10-year risk score. High altitude was significantly and negatively associated with Framingham 10-year risk scores. Our results will allow prevention strategies considering modifiable risk factors to avoid the development of cardiovascular diseases, especially in people living at low altitudes.

Limited data are available on hepatic steatosis (HS) in HIV patients who are not infected with hepatitis C virus (HCV). The aims of this study were to assess the prevalence of HS and its risk factors in HIV patients with and without HCV infection, and to evaluate whether HS correlates with advanced liver fibrosis and/or cardiovascular disease risk. Fifty-seven HIV mono-infected and 61 HIV/HCV co-infected patients were enrolled consecutively. All patients underwent liver ultrasound and transient elastography. The main parameters of liver function, HIV and HCV viral loads, CD4+ cell counts, and data on highly active antiretroviral therapy (HAART) were recorded. Cardiovascular disease risk was evaluated using the 10-year Framingham risk score. HS prevalence in the whole HIV population was 53% (54% in mono-infected patients and 51% in co-infected patients). HS was associated with lipodystrophy and triglyceride values (p<0.0001), metabolic syndrome (p<0.0004), and total cholesterol levels (p<0.001) in both HIV groups. In HIV mono-infected patients, HS was linked with HAART exposure of >1 year (p<0.01). By multivariate analysis, only triglyceride levels (p<0.02) and Framingham risk score (p<0.05) were independently associated with HS in both HIV groups. No correlation was observed between HS and advanced liver fibrosis, measured by transient elastography. HS was common in HIV patients, occurring in about half of the population. HS was found to be linked with the Framingham risk score, but was not correlated with advanced liver fibrosis. We suggest that in our HIV population with HS, the burden of cardiovascular disease risk is greater than that of liver disease progression.

Background/Introduction The 2019 European Society of Cardiology (ESC) guidelines currently recommend the use of coronary computed tomography angiography (CCTA) as the initial test for diagnosing coronary artery disease (CAD) in symptomatic patients in whom obstructive CAD cannot be excluded by clinical assessment alone. Purpose The purpose of this study is to identify the prevalence of obstructive CAD in patients with stable chest pain, and the correlation between several clinical cardiovascular disease (CVD) risk factors and CCTA findings presented by Coronary Artery Disease Reporting and Data System (CAD-RADS). Methods The present study is a single-center retrospective cross-sectional study. A total of 1,892 patients with stable chest pain who underwent CCTA were enrolled in this study. Diamond-Forrester classification, Framingham risk score (FRS), atherosclerotic CVD (ASCVD) 10-year risk score, coronary artery calcium score (CACS) and CAD-RADS category were obtained from every patient. Results Among 1,892 patients (mean age, 60.5±8.6 years; men, 59.3%), 356 (18.8%) had obstructive CAD according to CCTA. Patients with high and intermediate ASCVD 10-year risk score had 2.59 times (aOR 2.59, 95% CI; 1.58 to 4.23) and 1.66 times (aOR 1.66, 95% CI; 1.04 to 2.65) higher odds of having obstructive CAD than patients with low ASCVD 10-year risk score, respectively (adjusted for Diamond-Forrester classification and CACS group). Higher ASCVD risk scores were significantly associated with higher CAD-RADS category (p&amp;lt;0.001), and patients with CAD-RADS category 3 had ASCVD 10-year risk score of 20.1±12.7. CACS showed the highest discrimination in presence of obstructive CAD, followed by ASCVD 10-year risk score, FRS, and Diamond-Forrester classification (AUC: 0.821 [95% CI; 0.797–0.845]; 0.711 [95% CI; 0.683–0.740]; 0.675 [95% CI; 0.646–0.704]; 0.600 [95% CI; 0.569–0.632], respectively). Conclusion This is the first study of CCTA findings in stable chest pain patients in Korea. The prevalence of obstructive CAD in patients with stable chest pain was 18.8%. Higher ASCVD score is significantly associated with presence of obstructive CAD and higher CAD-RADS category. As coronary stenosis of 50%-69% had a mean ASCVD score of 20.1, we should consider CCTA for identifying obstructive CAD in patients with ASCVD score over 20 with stable chest pain. Funding Acknowledgement Type of funding sources: None.

This report describes a comparison of two geocoding methods used by the Cohorts for Environmental Exposures and Cancer Risks (CEECR) consortium: ArcGIS Geocoding by Esri and the SAS GEOCODE Procedure. The goal of this report is to determine the comparability of data sets that employ different approaches for linking survey data with spatial surrogates of exposure to environmental and socioeconomic factors. ArcGIS and SAS GEOCODE were selected as two platforms for this comparison because both programs are being used by one or more CEECR cohort study teams and they can be used locally offline. The latter minimizes confidentiality issues related to online data linkages. Residential addresses from 3,238 Wisconsin residents in the Cancer & COVID Study and the Wisconsin in situ Cohort (WISC) were geocoded and linked to eight different publicly available datasets of environmental and socioeconomic factors at various geographic scales using both geocoding platforms. Since the two analytic platforms vary in geocoding approaches, the validity and accuracy of both platforms were compared to examine differences when assigning surrogate measurements of exposure based on spatial locations. ArcGIS offered a higher specificity for matched addresses with slightly more latitude/longitude point and street matches (97.7%) than SAS (95.9%), with the remainder matching at the zip code level. The two geocoding platforms showed high concordance in assignment at the county (99.6%), census tract (96.5%), and census block group (94.7%). As a result, the correlations based on census tracts and block groups were very strong for linked exposure measures of socioeconomic status, environmental justice, urban/rural residence, air pollution, proximity to industrial sites, and cancer risk (all intraclass correlation coefficients ≥98%). Slightly lower concordance was observed for point source linkages (intraclass correlation coefficients 96-97%). Approximately ~4% of addresses were mis-matched largely in rural areas where census areas are larger and accurate geocoding base-layers are less widely available than in urban areas. For researchers that are already utilizing SAS, the GEOCODE procedure can be a logical choice as it is included in base SAS software and does not require an additional cost. However, SAS and ArcGIS provide similar options for the vast majority of study address locations.

The excess stroke mortality in the southeastern states of the United States (stroke-belt states) is well known; however, the factors associated with this pattern have not been fully elucidated. We measured the contribution of several demographic factors by analyzing stroke mortality data (2005-2009) at the census block group (BG) level in the state of Arkansas. Census BGs were used as proxies for neighborhoods. BGs were stratified using 5 census measures: poverty (percent of population below federal poverty level), population density (population per square mile), education (percent of population aged >25 years who did not graduate from high school), population mobility (percent of population who resided at the same address 1 year ago), and the percent of non-Hispanic blacks (percent of population that is black). Generalized additive models were used to estimate the variation in stroke mortality among BGs and to assess the impact of different demographic variables. From 2005 to 2009, there were 8930 stroke deaths in Arkansas. There was considerable variation in the relative risk even between adjacent BGs within a single county. The geographically weighted regression analyses indicated that 4.5% to 9% of deviance in stroke mortality among BGs could be explained by poverty, education, population density, and population mobility. Race/ethnicity (non-Hispanic blacks) explains <2% of the deviance in stroke mortality among BGs. Our study shows that primordial risk factors such as poverty and education drive disparities in stroke mortality among neighborhoods in Arkansas.

The purpose of this study was to examine the contribution of neighborhood socioeconomic conditions to risk of police-reported domestic violence in relation to victim's race. Data on race came from police forms legally mandated for the reporting of domestic violence and sexual assault. Using 1990 U.S. census block group data and data for the years 1996-1998 from Rhode Island's domestic violence surveillance system, the authors generated annual and relative risk of police-reported domestic violence and estimates of trends stratified by age, race (black, Hispanic, or white), and neighborhood measures of socioeconomic conditions. Race-specific linear regression models were constructed with average annual risk of police-reported domestic violence as the dependent variable. Across all levels of neighborhood poverty (< 5% to 100% of residents living below the federal poverty level), the risk of police-reported domestic violence was higher for Hispanic and black women than for white women. Results from the linear regression models varied by race. For black women, living in a census block group in which fewer than 10% of adults ages > or = 25 years were college-educated contributed independently to risk of police-reported domestic violence. Block group measures of relative poverty (> or = 20% of residents living below 200% of the poverty line) and unemployment (> or = 10% of adults ages > or = 16 years in the labor force but unemployed) did not add to this excess. For Hispanic women, three neighborhood-level measures were significant: percentage of residents living in relative poverty, percentage of residents without college degrees, and percentage of households monolingual in Spanish. A higher degree of linguistic isolation, as defined by the percentage of monolingual Spanish households, decreased risk among the most isolated block groups for Hispanic women. For white women, neighborhood-level measures of poverty, unemployment, and education were significant determinants of police-reported domestic violence. When data on neighborhood conditions at the block group level and their interaction with individual racial position are linked to population-based surveillance systems, domestic violence intervention and prevention efforts can be improved.

Objectives: The objectives of this study was to evaluate the degree of correlation between conventional risk models such as Framingham risk score (FRS) risk categories, and the burden of coronary artery disease (CAD) as assessed by coronary computed tomographic angiography (CCTA) using calcium score, Coronary Artery Disease-Reporting and Data System (CAD-RADS) score, segment involvement score (SIS), segment stenosis score (SSS), and modified Duke’s prognostic index (MDPI). Methodology: One hundred and forty-four patients with suspected CAD referred for CCTA were included in the study. FRS was determined, and patients were categorized into low, moderate, and high risk. On CCTA, calcium score, CAD-RADS, SIS, SSS, and MDPI were calculated. Correlation between FRS and CT scores was assessed using Pearson’s correlation. Results: Of 144 patients, 21.5% had CAD on CCTA. 10-year FRS of 10%–19% showed 6.8 times higher risk of CAD on CCTA; those with 10-year FRS of ≥20% showed 11.71 times higher risk of CAD as compared to those with FRS &lt;10%. FRS and CAD-RADS, SIS, SSS, and MDPI showed weak-to-moderate correlation ( r = 0.452, 0.401, 0.356, and 0.436, respectively, P &lt; 0.001). No one with FRS &lt;10% showed CADRDAS &gt;3, suggesting a good correlation between FRS and various CT scores. Conclusion: A strong correlation between FRS and CAD burden in the low-risk (&lt;10% risk) group was observed, with none of them showing CAD on CCTA. Obstructive CAD was seen only in the high-risk group (&gt;20%), indicating the importance of imaging in this group to decide preventive measures or therapy.

BackgroundCardiovascular disease (CVD) remains the major cause of global mortality. Applying a comprehensive interventional program may reduce the incidence of cardiovascular disease and its complications.ObjectiveThis study compared the effects of a three-month intervention involving lifestyle modification and physical activity with standard care in women ≥30 years having a moderate to high risk of CVD, with respect to improving physical activity and cardiovascular disease risk factors at the National Guard Residential City in Jeddah, Saudi Arabia, in 2015.MethodsThe effects of this community-based lifestyle program were assessed through a randomized controlled trial from January 1st to September 6th, 2015. Women in the intervention group (n = 31) received health education, exercise training, and diet counselling as individuals and in groups according to the participant’s risk. Women in the control group (n = 28) received one health education session at the screening site. The primary outcome was the proportion of women with moderate Framingham risk scores (FRS) reducing their risk by 10% and the proportion of women with high FRS reducing their risk by 25%. The secondary outcome was the proportion of women reducing their risk by ≥1 risk category.ResultsThe mean participant age was 42 ± 8 years. At three-month’s follow-up, reductions were greater in the intervention group and the difference between groups was statistically significant (p < 0.05). Lifestyle intervention program significantly reduced systolic blood pressure (-9.2 mmHg), blood glucose (-45 mg/dL) and Framingham risk score (-13.6). Linear regression analysis revealed a significant improvement in the Framingham risk score (p < 0.01).ConclusionIn a population of women with moderate-to-high risk of CVD, a personalized lifestyle modification program showed positive association in improving the 10-year cardiovascular Framingham risk score after three months.

To examine the change in Framingham risk score (FRS) arising from short-term treatment with ziprasidone or olanzapine. Hospitalized adults with a primary DSM-IV diagnosis of schizophrenia or schizo-affective disorder were randomly assigned to 6 weeks of double-blind treatment with ziprasidone or olanzapine from November 21, 1998 to September 28, 2000. Data on fasting lipid levels were collected at screening and endpoint, and blood pressure was measured at screening and baseline and weekly until week 6 of treatment (or last visit). FRS for patients aged ≥30 years was calculated using an algorithm derived from the Framingham Heart Study. Baseline-to-endpoint least-squares mean changes in age-adjusted FRS by gender were compared using analysis of covariance (baseline adjusted). Men who received olanzapine demonstrated a mean increase in their total cholesterol levels (+18.5 mg/dL; N = 53) and low-density lipoprotein cholesterol levels (+13.0 mg/dL; N = 45), whereas men who received ziprasidone demonstrated a mean decrease in their total cholesterol levels (-8.5 mg/dL; N = 44) and low-density lipoprotein cholesterol levels (-7.2 mg/dL; N = 40) (p = .0006 and p = .004, respectively). Additionally, men who received olanzapine showed an increase in baseline FRS (+7.69%; N = 53), whereas men who received ziprasidone showed a decrease in baseline FRS (-11.06%; N = 42) (p = .09). In women, treatment differences in FSR numerically favored ziprasidone but were not statistically significant. Neither treatment had a significant effect on blood pressure. In short-term treatment, olanza-pine was associated with a significant worsening of lipid profile compared with ziprasidone, with a consequent increase in FRS versus ziprasidone. These findings, coupled with the significant weight gain in patients treated with olanzapine versus ziprasidone, warrant investigation in longer-term trials.

Introduction: Left ventricular hypertrophy (LVH) is more prevalent in blacks than whites, and is a major independent predictor of coronary heart disease (CHD)/CVD survival in blacks. We evaluated the ability of LVH to predict CHD outcomes beyond traditional cardiovascular risk factors in blacks, compared with whites from a large community-based cohort. Methods: Data were analyzed on 14,489 participants (mean age 54 +/- 5.7 years, 43.5% men, and 26% black) within the ARIC cohort, with baseline (1987-1989) electrocardiograms (ECG), followed through 2009. Risk estimates for incident CHD were assessed using the 10-year Framingham Risk Score (FRS). Model 1 was the Framingham base model, while model 2 included the base model plus LVH by any of 11 traditional ECG-LVH criteria (Table). Net reclassification improvement (NRI) was calculated, and the distribution of risk was compared using model 2 vs. model 1. Results: There were 690 (4.8%) 10-year and 1515 (10.5%) 20-year CHD events. LVH defined by any criteria was associated with CHD events in the entire cohort [HR (95% CI): 1.42 (1.20-1.7)]. LVH defined by the Framingham ECG score and LV strain criteria were the criteria most associated with CHD overall. LVH by Framingham ECG score was most associated with CHD in blacks [HR (95%CI): 2.53 (1.65-3.89)], while LV strain showed the strongest association with CHD in whites [1.73 (1.18-2.56)]. No statistically significant improvement in NRI or C-statistic was observed in model 2 [C-statistic (95% CI): 0.779 (0.763-0.794), NRI = 0.006 (p = 0.41)], compared with the base model [0.777 (0.762-0.792)]; and no racial interactions were observed. Findings were unchanged when the base model was replaced with the 10 and 20-year ARIC risk model (includes diabetes) for CHD. Conclusions: In this cohort of black and white men and women, LVH (defined by ECG) was significantly associated with CHD after adjustment for FRS; but did not significantly improve CHD risk prediction beyond the FRS. No significant black-white differences in risk prediction were observed. Table. Reclassification of Coronary Heart Disease by the addition of each Criterion for Left Ventricular Hypertrophy, based on a 10-year Framingham Risk Model * Base model factors in age, gender, current smoking, diabetes, systolic blood pressure, diastolic blood pressure, HDL cholesterol and total cholesterol as separate variables. * NRI categorized as &lt;10%, 10-20% and &gt;20%. Abbreviations: NRI = Net Reclassification Index, IDI = Integrated Discrimination Index, HR = Hazard Ratio, CI = Confidence Interval