Causal Mediation Research Articles

Causal Mediation Analysis for Childhood Cancer Survival Disparity in Texas, 2005 to 2014

This study aimed to examine the disparities of childhood cancer survival among different racial and ethnic groups in Texas. The analysis was mediated by socioeconomic status (SES) and spatial accessibility to Children Oncology Group (COG) hospitals. The relationship between race-ethnicity and overall survival was measured using the Cox proportional hazards model with a robust variance estimator. The counterfactual model measures the total effect of race-ethnicity on survival through all mediating pathways while adjusting for baseline confounders (age, sex, and stage at diagnosis), which are then decomposed into natural direct and indirect effects. Considering all cancer site groups, African Americans showed a statistically significant higher hazard ratio in death (HR = 3.63; 95% CI = 1.87 - 6.62) compared with non-Hispanic White children. At the same time, the mortality hazard ratio among Hispanic children is not significant (HR = 1.23; 95% CI = 0.80 - 1.93) when compared with non-Hispanic White children. Analysis results also suggested that both mediators significantly contribute to racial-ethnic survival disparities for specific cancer site groups such as Leukemia for African American children. This study builds knowledge and understanding about underlying factors (mediators) responsible for the disparities in the outcomes among childhood cancer patients.

The cortisol/DHEA ratio mediates the association between early life stress and externalizing problems in adolescent boys

BackgroundDespite evidence that early life stress (ELS) can influence the functioning of the hypothalamic-pituitary-adrenal (HPA) axis and increase maladaptive behaviors in adolescence, less attention has been paid to the role of the coordinated effects of the two primary adrenal hormones, cortisol and dehydroepiandrosterone (DHEA), in these associations. Methods138 typically developing adolescents (76 females) reported the stressful events experienced during childhood and early adolescence across 30 domains. Two years later we assessed levels of externalizing problems and obtained salivary levels of cortisol and DHEA. Using causal moderated mediation analyses, we examined whether the ratio of cortisol to DHEA (CD ratio) mediates the association between ELS and subsequent externalizing problems. ResultsWe found that ELS is associated with both a lower CD ratio and more externalizing problems. Importantly, a lower CD ratio mediated the association between ELS and externalizing problems in boys. ConclusionsAn imbalance in adrenal hormones may be a mechanism through which ELS leads to an increase in externalizing problems in adolescent boys. These findings underscore the utility of using the CD ratio to index HPA-axis functioning.

Do sleep and circadian characteristics predict alcohol use in adult drinkers?

While sleep and circadian rhythms are recognized contributors to the risk for alcohol use and related problems, few studies have examined whether objective sleep and circadian measures can predict future alcohol use in humans, and no such studies have been conducted in adults. This study examined whether any baseline sleep and/or circadian characteristics of otherwise healthy adults predicted their alcohol use over the subsequent 12 months. Participants (21-42 years) included 28 light and 50 heavy drinkers. At baseline, a comprehensive range of self-reported and objective sleep/circadian measures was assessed via questionnaires, wrist actigraphy, and measurement of dim light melatonin onset and circadian photoreceptor responsivity. Following this, the number of alcoholic drinks per week and binge drinking episodes per month were assessed quarterly over the subsequent 12 months. Anticipated effects of alcohol (stimulation, sedation, and rewarding aspects) were also assessed quarterly over the 12 months. Analyses included generalized linear mixed-effects models and causal mediation analysis. Across the range of measures, only self-reported insomnia symptoms and a longer total sleep time at baseline predicted more drinks per week and binges per month (ps <0.02). There was a trend for the anticipated alcohol effect of wanting more alcohol at the 6-month timepoint to mediate the relationship between insomnia symptoms at baseline and drinks per week at 12 months (p = 0.069). These results suggest that in otherwise healthy adults, insomnia symptoms, even if subclinical, are a significant predictor of future drinking, and appear to outweigh the influence of circadian factors on future drinking, at least in otherwise healthy adults. Insomnia symptoms may be a modifiable target for reducing the risk of alcohol misuse.

Using causal mediation to examine self-efficacy as a mechanism through which continuing care interventions reduce alcohol use.

Understanding the causal mechanisms through which telephone and mobile health continuing care approaches reduce alcohol use can help develop more efficient interventions that effectively target these mechanisms. Self-efficacy for successfully coping with high-risk alcohol relapse situations is a theoretically and empirically supported mediator of alcohol treatment. This secondary analysis aims to examine self-efficacy as a mechanism through which remote-delivered continuing care interventions reduce alcohol use. The study included 262 adults (Mage = 46.9, SD = 7.4) who had completed 3 weeks of an intensive outpatient alcohol treatment program. The sample was predominantly male (71%), African American (82%), and completed a high school education (71%). The four-arm randomized clinical trial compared three active continuing care interventions (telephone monitoring and counseling [TMC], addiction comprehensive health enhancement support system [ACHESS], and combined delivery of TMC and ACHESS) to usual care and assessed longitudinal measures of alcohol use and self-efficacy. Analyses employed the potential outcomes framework and sensitivity analyses to address threats to causal inference resulting from an observed mediator variable. Relative to usual care, the two intervention conditions that included TMC reduced alcohol use through improvements to self-efficacy. There was no evidence that self-efficacy mediated the effect of ACHESS on alcohol use. Based on our findings, self-efficacy is an important mechanism through which telephone continuing care interventions affect alcohol use. Future research to identify which components of TMC influence self-efficacy and factors that mediate ACHESS effects could enhance the effectiveness of remote delivery of continuing care. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Cumulative environmental stress and emerging cardiometabolic risk during childhood.

To prospectively evaluate the relationship between cumulative environmental stress and cardiometabolic risk in middle childhood, and to examine whether hair cortisol, a measure of hypothalamic pituitary adrenal-axis activity, mediates this relationship. In a cohort of children from low-income households (n = 320; 59% Hispanic, 23% Black, body mass index (BMI) percentile >50th at enrollment), environmental stressors including family and neighbourhood factors representing disadvantage/deprivation, and cortisol concentrations from hair samples, were measured over five timepoints beginning when children were 2-4 years old. Cardiometabolic risk factors (i.e., BMI, blood pressure, lipids, blood sugar, C-reactive protein) were measured at the final timepoint when children were 7-11 years of age. In adjusted logistic regression models, greater cumulative environmental stress was associated with a higher likelihood of elevated cardiometabolic risk in middle childhood (p = 0.01). Children from minoritized racial/ethnic groups had a higher prevalence of both stressors and cardiometabolic risk factors. Cumulative environmental stress was associated with higher hair cortisol concentrations (p < 0.01). However, hair cortisol was not directly associated with cardiometabolic risk factors and did not explain the association between environmental stress and cardiometabolic risk in causal mediation analysis. The influence of cumulative stress on cardiometabolic health can be observed in middle childhood and may contribute to cardiometabolic health disparities, highlighting the importance of public health interventions to mitigate disadvantage.

Smoking, Lung Cancer Stage, and Prognostic Factors-Findings from the National Lung Screening Trial.

Low-dose computed tomography (LDCT) increases the early detection of lung cancer. Identifying modifiable behaviors that may affect tumor progression in LDCT-detected patients increases the likelihood of long-term survival and a good quality of life. We examined cigarette smoking behaviors on lung cancer stage, progression, and survival in 299 ever-smoking patients with low-dose CT-detected tumors from the National Lung Screening Trial. Univariate and multivariate Cox models were used to estimate the hazard ratio (HR) for smoking variables on survival time. Current vs. former smokers and early morning smokers (≤5 min after waking, i.e., time to first cigarette (TTFC) ≤ 5 min) had more advanced-stage lung cancer. The adjusted HR for current vs. former smokers was 1.3 (95% confidence interval [CI] 0.911-1.98, p = 0.136) for overall survival (OS) and 1.3 (0.893-1.87, p = 0.1736) for progression-free survival (PFS). The univariate hazard ratios for TTFC ≤ 5 min vs. >5 min were 1.56 (1.1-2.2, p = 0.013) for OS and 1.53 (1.1-2.12, p = 0.01) for PFS. Among current smokers, the corresponding HRs for early TTFC were 1.78 (1.16-2.74, p = 0.0088) and 1.95 (1.29-2.95, p = 0.0016) for OS and PFS, respectively. In causal mediation analysis, the TTFC effect on survival time was mediated entirely through lung cancer stage. The current findings indicate smoking behaviors at diagnosis may affect lung cancer stage and prognosis.

Abstract 6128: Association between neighborhood disinvestment and all-cause survival moderated by epigenetic age among women with breast cancer

Abstract Background: DNA methylation is hypothesized to mediate relationships between chronic social disadvantage and cancer outcomes. We investigated whether epigenetic age acceleration mediates the association between modifiable neighborhood disinvestment and survival within a cohort of women with breast cancer. Methods: Individual-level data were from the Women’s Circle of Health Follow-up Study, a breast cancer cohort comprising women self-identifying as African American or Black, diagnosed 2013-2019 in New Jersey residents. Demographic, socioeconomic, health behavior, and dietary characteristics were self-reported. Peripheral blood samples from 312 participants collected 18-24 months after diagnosis were profiled for DNA methylation using Illumina’s MethylationEPIC array. After pre-processing, pace of biological aging was estimated using DunedinPACE. Neighborhood disinvestment was assessed using a previously validated virtual audit of 6 disinvestment indicators - garbage, graffiti, dumpsters, poor building conditions, poor yard conditions, abandoned buildings - in 14,671 Google Streetview streetscapes. Accelerated failure time models of all-cause mortality as functions of neighborhood disinvestment and pace of biologic aging were fit to estimate survival time ratios (TR), adjusted for stage at diagnosis and neighborhood socioeconomic composition using causal mediation analyses. End of follow-up was August 13, 2023. Results: There were 46 all-cause deaths through follow-up (median=6.9 years), and 5-year overall survival probability was 0.90 (95% CI: 0.86, 0.93). Minimum, median, and maximum pace of biologic aging was 0.87, 1.26, and 1.72 indicating accelerated aging (&amp;gt; 1.0) among most participants. There was no evidence for mediation of the neighborhood disinvestment - survival effect by pace of biologic aging (indirect effect p-value=0.33). There was evidence of an interaction between pace of biologic aging and neighborhood disinvestment in the adjusted model: survival time decreased 37% (95% CI: 4% - 59%, p=0.03) per 1 standard deviation (SD) increase in neighborhood disinvestment at pace of biologic aging values 1 SD below the average (&amp;lt; 1.13). There was no evidence of such association at pace of biologic aging values 1 SD above the average (p=0.86). Conclusion: Greater neighborhood disinvestment may be associated with shorter survival following a breast cancer diagnosis among those with slower than average epigenetic age acceleration, due to the cumulative nature of neighborhood disinvestment as a chronic indicator of social disadvantage. Citation Format: Jesse J. Plascak, Kellie Archer, Adana A. Llanos, Stephen J. Mooney, Cathleen Y. Xing, Andrew G. Rundle, Helmut Zarbl, B Qin, Nur Zeinomar, Mario Schootman, Song Yao, Christine Ambrosone, Karen Pawlish, Kitaw Demissie, Elisa V. Bandera, Chi-Chen Hong. Association between neighborhood disinvestment and all-cause survival moderated by epigenetic age among women with breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6128.

Abstract 6544: Targeting resistance mechanisms to AR-targeted therapy in prostate cancer through inhibition of CREBBP/EP300

Abstract Standard initial treatment for men with metastatic prostate cancer involves targeting the androgen receptor (AR) with androgen deprivation therapy or androgen-receptor signaling inhibitors. Resistance eventually develops, such as through AR-activating alterations, leaving few targeted options for patients. Here, we utilize data from 4,509 prostate cancer patients with longitudinal clinical and DNA/RNA sequencing data to elucidate mechanisms of resistance and identify potential therapeutic vulnerabilities in resistant disease. By constructing comprehensive regulatory networks, termed hybrid networks, built across different disease stages, combining features from WGCNA, causal mediation analysis, probabilistic causal network inference, and public knowledge graphs, we uncover subnetworks strongly associated with disease progression and resistance. These subnetworks reflect biological processes for cell cycle, E2F, DNA repair, MYC, and AR, comprehensively identifying broad transcriptional changes that drive disease progression and resistance. Selected subnetwork scores, derived from RNA expression, were significantly associated with outcomes in both the hormone-sensitive and castration-resistant settings, demonstrating that the subnetworks capture early and late resistance. The hybrid networks identified the transcriptional co-activators CREBBP and EP300 as key modulators of these subnetworks, suggesting that targeting CREBBP/EP300 may be effective in treating both early and late-stage disease. Prostate cancer cell lines were treated with FT-6876, a small molecule inhibitor of CREBBP/EP300, and growth inhibition was measured in both enzalutamide sensitive and resistant cell lines, covering a wide range of resistance mechanisms. The selected subnetworks were enriched for genes down-regulated following treatment with FT-6876, suggesting that targeting CREBBP/P300 addresses multiple resistance-related pathways. Differential changes in protein abundance and protein acetylation between enzalutamide sensitive and resistant cell lines identified changes in fatty acid metabolism as a common, downstream event in response to treatment with FT-6876, suggesting an additional mechanism for response to FT-6876. Together, our integrated theoretical and experimental analyses support the hypothesis that targeting CREBBP/EP300 may provide therapeutic benefits to prostate cancer patients. Further, the subnetworks identified may serve as mechanistically-informed biomarkers for identifying patients who may benefit from inhibition of CREBBP/EP300. Citation Format: Bonnie V. Dougherty, Beatrice C. Thomas, Alice M. Walsh, Veronica Calvo, Kathryne E. Ware, Andrew J. Armstrong, Jens Renstrup, Jason A. Somarelli, Eric E. Schadt. Targeting resistance mechanisms to AR-targeted therapy in prostate cancer through inhibition of CREBBP/EP300 [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6544.

An empirical evaluation of approximate and exact regression-based causal mediation approaches for a binary outcome and a continuous or a binary mediator for case-control study designs

BackgroundIn the causal mediation analysis framework, several parametric regression-based approaches have been introduced in past years for decomposing the total effect of an exposure on a binary outcome into a direct effect and an indirect effect through a target mediator. In this context, a well-known strategy involves specifying a logistic model for the outcome and invoking the rare outcome assumption (ROA) to simplify estimation. Recently, exact estimators for natural direct and indirect effects have been introduced to circumvent the challenges prompted by the ROA. As for the approximate approaches relying on the ROA, these exact approaches cannot be used as is on case-control data where the sampling mechanism depends on the outcome.MethodsConsidering a continuous or a binary mediator, we empirically compare the approximate and exact approaches using simulated data under various case-control scenarios. An illustration of these approaches on case-control data is provided, where the natural mediation effects of long-term use of oral contraceptives on ovarian cancer, with lifetime number of ovulatory cycles as the mediator, are estimated.ResultsIn the simulations, we found few differences between the performances of the approximate and exact approaches when the outcome was rare, both marginally and conditionally on variables. However, the performance of the approximate approaches degraded as the prevalence of the outcome increased in at least one stratum of variables. Differences in behavior were also observed among the approximate approaches. In the data analysis, all studied approaches were in agreement with respect to the natural direct and indirect effects estimates.ConclusionsIn the case where a violation of the ROA applies or is expected, approximate mediation approaches should be avoided or used with caution, and exact estimators favored.

Improved Child Feces Management Mediates Reductions in Childhood Diarrhea from an On-Site Sanitation Intervention: Causal Mediation Analysis of a Cluster-Randomized Trial in Rural Bangladesh

BackgroundThe WASH benefits Bangladesh trial multi-component sanitation intervention reduced diarrheal disease among children < 5 years. Intervention components included latrine upgrades, child feces management tools, and behavioral promotion. It remains unclear which components most impacted diarrhea.MethodsWe conducted mediation analysis within a subset of households (n = 720) from the sanitation and control arms. Potential mediators were categorized into indicators of latrine quality, latrine use practices, and feces management practices. We estimated average causal mediation effects (ACME) as prevalence differences (PD), defined as the intervention’s effect on diarrhea through its effect on the mediator.ResultsThe intervention improved all indicators compared to controls. We found significant mediation through multiple latrine use and feces management practice indicators. The strongest mediators during monsoon seasons were reduced open defecation among children aged < 3 and 3–8 years, and increased disposal of child feces into latrines. The strongest mediators during dry seasons were access to a flush/pour-flush latrine, reduced open defecation among children aged 3–8 years, and increased disposal of child feces into latrines. Individual mediation effects were small (PD = 0.5–2 percentage points) compared to the overall intervention effect but collectively describe significant mediation pathways.DiscussionThe effect of the WASH Benefits Bangladesh sanitation intervention on diarrheal disease was mediated through improved child feces management and reduced child open defecation. Although the intervention significantly improved latrine quality, relatively high latrine quality at baseline may have limited benefits from additional improvements. Targeting safe child feces management may increase the health benefits of rural sanitation interventions.

Testing Whether Protective Parenting is a Causal Mediator of Intervention Effects on Decreased Delinquency Using a Randomized Prevention Trial

Protective parenting practices, including parental monitoring and establishing nurturing and supportive rules, are thought to affect the risk of children’s involvement in delinquency. However, there has yet to be any direct test of this hypothesis using experimental data designed to test causality better. Data from 346 Black couples with an 11-year-old child were assigned randomly to Protecting Strong African American Families (ProSAAF) intervention or control condition. Results from traditional mediation models, causal mediation analysis, and the complier analysis indicated that the intervention had a significant indirect effect on youth delinquent behaviors and that effects on protective parenting mediated this effect. The results demonstrated that ProSAAF, which is designed to promote family communication, is effective in deterring delinquent behaviors among Black Americans who reside in resource-scarce communities by enhancing parenting practices.

Association between dietary magnesium intake and gallstones: the mediating role of atherogenic index of plasma.

Dyslipidemia and abnormalities in cholesterol metabolism are commonly observed in individuals with gallstone disease. Previous research has demonstrated that dietary magnesium can influence lipid metabolism. The atherogenic index of plasma (AIP) has emerged as a novel lipid marker. This study aimed to examine the possible correlation between dietary magnesium intake and gallstones and the potential mediating role of AIP in US adults. A total of 4,841 adults were included in this study from the National Health and Nutrition Examination Survey (NHANES) conducted from 2017 to 2020. A variety of statistical techniques such as logistic regression, subgroup analysis, smoothed curve fitting, and causal mediation analysis were utilized to analyze the information collected from the participants. In the fully adjusted model, a statistically noteworthy inverse relationship was observed between dietary magnesium intake and the presence of gallstones, as indicated by an odds ratio (OR) of 0.58 and a 95% confidence interval (CI) of (0.42, 0.81). Causal intermediary analysis revealed that the association between magnesium intake and gallstones was partially mediated by AIP, with a mediation ratio of 3.2%. According to this study, dietary magnesium intake had a significant linear negative association with the prevalence of gallstones, in which AIP played a mediating role. This discovery offers novel perspectives on the prevention and management of gallstones.

Abstract 27: Chronotype, Life’s Essential 8, and Risk of Cardiovascular Disease: A Prospective Cohort Study in UK Biobank

Introduction: Individuals with evening chronotype often experience circadian misalignment, which may disrupt health behaviors and circadian regulation of cardiometabolic functions such as blood pressure and heart rate. However, the association of chronotype with cardiovascular disease (CVD) and the interplay with other CVD risk factors are not fully understood. Hypothesis: We assessed the hypothesis that individuals with an evening chronotype, compared to those with a morning chronotype, have a poorer cardiovascular health as measured by the Life’s Essential 8 (LE8), and consequently an increased CVD risk. We further explored if the associations differed by sex. Methods: We conducted a prospective study in the UK Biobank among 477,878 adults aged 39-74 years and free of CVD at baseline in 2006-2010. Chronotype was self-reported using a single representative question from the Morningness-Eveningness Questionnaire. The LE8 score was calculated based on 8 modifiable CVD risk factors, and ranged from 0 to 100 with higher scores indicating better cardiovascular health. Incident CVD was defined as first myocardial infarction (MI) or stroke leading to hospitalization or death, identified via validated algorithms that mapped national medical records and death certificates to ICD-10 codes until May 2022. We used Cox proportional hazards models to estimate the association between chronotype and CVD risk. Models were adjusted for demographics, shift work, and family history of CVD. We applied causal mediation framework to decompose the total chronotype-CVD effect into estimated natural direct effect (i.e., independent of LE8) and natural indirect effect (i.e., mediated by LE8; NIE). Results: Participants with a “definite evening” chronotype were 75% more likely to have a low LE8 score (&lt;50) compared to “intermediate” type (prevalence ratio 95% CI: 1.67, 1.84). This association was notably stronger among women ( P -interaction: 0.0001). Over 12.7 years of follow-up, there were 27,346 documented incident CVD events (17,180 MI; 11,329 stroke). The hazard ratio (HR) for total CVD was 1.03 (95% CI: 1.00, 1.06) for the “definite morning” and 1.15 (95% CI: 1.10, 1.20) for “definite evening” compared with “intermediate” chronotype ( P -trend: 0.06). This association was stronger in women ( P -interaction: 0.05), and similar for MI and stroke. The NIE of the chronotype-CVD association comparing “definite evening” to “intermediate” was 1.11 (95% CI: 1.09, 1.13), equivalent to 95% of the association being mediated by LE8. Conclusions: Evening chronotype was associated with a modest increase in CVD risk, compared to intermediate chronotype, which appeared to be mainly explained by overall poor cardiovascular health. Our results suggest potential benefits of targeted interventions in evening chronotype for CVD prevention. Ongoing work is exploring potential effect modification by night shift work.

Associations of Frailty with Neuropsychiatric Symptoms of Alzheimer's Disease: A Longitudinal Study.

Frailty is a vulnerability state increasing the risk of many adverse health outcomes, but little is known about the effects of frailty on neuropsychiatric health. To explore the associations between frailty and the risk of neuropsychiatric symptoms (NPSs) in Alzheimer's disease (AD), especially in its different clinical stages. We included 2,155 individuals assessed using modified frailty index-11 (mFI-11), Neuropsychiatric Inventory (NPI) and Neuropsychiatric Inventory Questionnaire (NPI-Q) in the Alzheimer's Disease Neuroimaging Initiative (ADNI). The relationships between frailty and NPSs were explored with logistic regression models and Cox proportional hazard regression models. Causal mediation analyses were conducted to explore the mediation factors between frailty and NPSs. Among mild cognitive impairment (MCI) participants, frailty was cross-sectionally associated with an increased risk of apathy, and longitudinally associated with increased risk of depression and apathy. Among AD participants, frailty was cross-sectionally associated with increased risk of depression and anxiety, and longitudinally associated with an increased risk of apathy. Among participants with cognitive progression, frailty was associated with increased risk of depression and apathy. In MCI participants, the influence of frailty on NPSs was partially mediated by hippocampus volume, whole brain volume, and monocytes, with mediating proportions ranging from 8.40% to 9.29%. Frailty was associated with NPSs such as depression, anxiety, and apathy among MCI, AD, and cognitive progression participants. Atrophy of the hippocampus and whole brain, as well as peripheral immunity may be involved in the potential mechanisms underlying the above associations.

Abstract P374: Understanding the Mediating Role of Allostatic Load on the Relationship Between Long-Term Financial Strain and Subclinical CVD Among Black Women: The Study of Women's Health Across the Nation (SWAN)

Introduction: Black adults are disproportionately burdened by poor cardiovascular health, and Black women bear the brunt of this disparity. Cardiovascular disease (CVD) disparities among Black women have been attributed to chronic exposure to psychosocial stressors and the cumulative “wear and tear” of these stressors on the body (i.e., allostatic load). Financial strain has been linked to both higher allostatic load and CVD risk. We aimed to determine whether allostatic load (AL) mediates the relationship between financial strain and cumulative CVD risk among Black women. Methods: This study utilized longitudinal data from the Black participants in the Study of Women's Health Across the Nation (SWAN). We conducted a causal mediation analysis with fourfold decomposition to determine if the relationship between long-term financial strain and subclinical CVD was mediated by AL. The study outcome was defined as the presence of any indicator of subclinical CVD from carotid ultrasounds conducted in visit 12/13: carotid artery intima-media thickness (cIMT) ≥ 75 th percentile of the study sample, carotid adventitial diameter (cAD) ≥ 75 th percentile of the study sample, or plaque index (a measure of number and size of plaques) &gt;2. Long-term financial strain was defined as reporting “money problems” or “difficulty paying for basics” at &gt;50% of the 6 most recent visits prior to the carotid ultrasound. AL was calculated using a validated summary score method based on 10 component measures from visit 12: systolic and diastolic blood pressure, serum glucose, BMI, triglycerides, C-reactive protein, HDL-C, total cholesterol, waist-to-hip ratio, and dehydroepiandrosterone sulfate. AL was considered “high” if the score was ≥75 th percentile of the study sample. Results: Of the 390 Black women in our sample [Mean (SD) age at subclinical CVD visit = 60.0 (2.7)], 42.3% had subclinical CVD (24.9% with cIMT or cAD, 9.7% with plaque), 56.9% reported long-term financial strain, and 38.0% had high AL. Both long-term financial strain (β=0.132, p &lt;0.01) and high AL (β=0.134, p &lt;0.01) were associated with a significantly higher likelihood of subclinical CVD. However, causal mediation analysis indicated that the effect of long-term financial strain on subclinical CVD was not mediated by AL (β=0.004, p =0.66). Conclusions: Though higher AL is associated with higher CVD risk among Black women, it did not mediate the causal relationship between long-term financial strain and CVD risk in these midlife women. Future research should examine the pathways by which financial strain and AL influence premature cardiovascular weathering among midlife Black women and explore potential large-scale interventions.

A novel spatiotemporal graph convolutional network framework for functional connectivity biomarkers identification of Alzheimer’s disease

BackgroundFunctional connectivity (FC) biomarkers play a crucial role in the early diagnosis and mechanistic study of Alzheimer’s disease (AD). However, the identification of effective FC biomarkers remains challenging. In this study, we introduce a novel approach, the spatiotemporal graph convolutional network (ST-GCN) combined with the gradient-based class activation mapping (Grad-CAM) model (STGC-GCAM), to effectively identify FC biomarkers for AD.MethodsThis multi-center cross-racial retrospective study involved 2,272 participants, including 1,105 cognitively normal (CN) subjects, 790 mild cognitive impairment (MCI) individuals, and 377 AD patients. All participants underwent functional magnetic resonance imaging (fMRI) and T1-weighted MRI scans. In this study, firstly, we optimized the STGC-GCAM model to enhance classification accuracy. Secondly, we identified novel AD-associated biomarkers using the optimized model. Thirdly, we validated the imaging biomarkers using Kaplan–Meier analysis. Lastly, we performed correlation analysis and causal mediation analysis to confirm the physiological significance of the identified biomarkers.ResultsThe STGC-GCAM model demonstrated great classification performance (The average area under the curve (AUC) values for different categories were: CN vs MCI = 0.98, CN vs AD = 0.95, MCI vs AD = 0.96, stable MCI vs progressive MCI = 0.79). Notably, the model identified specific brain regions, including the sensorimotor network (SMN), visual network (VN), and default mode network (DMN), as key differentiators between patients and CN individuals. These brain regions exhibited significant associations with the severity of cognitive impairment (p < 0.05). Moreover, the topological features of important brain regions demonstrated excellent predictive capability for the conversion from MCI to AD (Hazard ratio = 3.885, p < 0.001). Additionally, our findings revealed that the topological features of these brain regions mediated the impact of amyloid beta (Aβ) deposition (bootstrapped average causal mediation effect: β = -0.01 [-0.025, 0.00], p < 0.001) and brain glucose metabolism (bootstrapped average causal mediation effect: β = -0.02 [-0.04, -0.001], p < 0.001) on cognitive status.ConclusionsThis study presents the STGC-GCAM framework, which identifies FC biomarkers using a large multi-site fMRI dataset.

HOW DO DIFFERENT FORMS OF DIGITALIZATION AFFECT INCOME INEQUALITY?

This article examines how different forms of digitalization affect inequality in Europe. Using a cross-national dataset of economic development and digitalization across a range of regression specifications including country and time fixed effects, this article explores the heterogeneous relationships of disparate forms of digitalization – human capital, broadband connectivity, integration of digital technology into small and medium enterprises, and digital public services – with income inequality. Fixed country and time effects models show that only the digitalization of human capital and integration of digital technology by SMEs are associated with decreases in income inequality. Causal mediation analysis reveals that tertiary education, despite its oft-cited connection to digital technology uptake, has no causal effect on the pathways through which digitalization of labour and SME operations lower inequality, which are direct. The findings tentatively suggest that there exist informal sources of digital skills training apart from formal tertiary education and point to SMEs as a potentially impactful area for investing in digitalization as pathways for income redistribution.

Which foreign vaccine should the government purchase in a pandemic? Evidence from a survey experiment in the United States

BackgroundFor many countries confronting a future pandemic, the initial vaccines available will come from abroad. Public hesitancy to receive these foreign vaccines is important, as it may create an incentive for governments to forego procuring them for public use. We investigate the influence of the vaccine's country of origin on public support for government procurement during the early stages of a pandemic and examine whether endorsements from the WHO can mitigate such biases. MethodsIn the summer of 2023, we carried out a survey experiment of 1,110 U.S. residents where we asked respondents to rate their support for vaccine purchasing policies for 20 hypothetical vaccines (13,320 evaluations). We varied the vaccine's country of origin and its endorsement status from the WHO, while also randomizing other vaccine attributes. ResultsCompared to foreign vaccines from countries Americans see favorable (e.g., Germany, the United Kingdom), those originating from less favorable countries (e.g., China, Russia), garnered lower support for government procurement. Our causal mediation analysis indicates that this country-of-origin effect is primarily driven by participants' sentiments toward the vaccine. Surprisingly, WHO endorsement does little to mitigate the effect of the vaccine's country of origin. These findings are consistent across various sample subsets and considerations of vaccine quality. ConclusionOur study advances previous work on vaccine country-of-origin effects by assessing its impact on policy preferences for procuring initial vaccines from overseas (as opposed to uptake intentions), identifying a mechanism by which vaccine favoritism occurs, and documenting that neither personal disease susceptibility nor vaccine quality fully mitigates country of origin effects. We conclude by discussing why the study of “vaccine diplomacy” ought to not only include interstate dynamics governing vaccine purchasing and availability but also consider vaccine-producing countries' more general reputations.

Does innovation help to explain the effect of export on productivity? Evidence from the French dairy industry

Abstract While there is strong evidence that becoming an exporter increases a firm’s productivity, underlying mechanisms that explain such a relationship remain largely unexplored. This paper analyses the contribution of the complementarity between exporting and investment in technology as a potential driver of export-related productivity gains. We employ firm-level data on production and trade combined with information on new products in the French dairy industry to conduct a causal mediation analysis between exporting, innovation investment, innovation output and productivity. Our estimation result show that starting exports increase the productivity by eight percentage points. Our mediation analysis reveals that innovation (investment and output) explains 31 per cent of this productivity growth.

Association between dietary inflammatory potential and frailty is mediated by inflammation among patients with colorectal cancer: A cross-sectional study

Patients with colorectal cancer (CRC) are at high risk of frailty, leading to reduced quality of life and survival. Diet is associated with frailty in the elderly through regulating inflammation. Thus, we hypothesized that dietary inflammatory potential (as assessed by dietary inflammatory index [DII]) might be associated with frailty in patients with CRC through regulating inflammatory biomarkers. A total of 231 patients with CRC were included in this cross-sectional study. Dietary intake was evaluated by 3-day, 24-hour dietary recalls, and frailty status was assessed in accordance with the Fried frailty criteria. Plasma inflammatory cytokines were determined in 126 blood samples. A total of 67 patients (29.0%) were frail, with significantly higher DII scores than nonfrail patients, accompanied with significantly increased interleukin-6 (IL-6) and decreased interleukin-10 (IL-10) concentrations. Each 1-point increase of DII was related to a 25.0% increased risk of frailty. IL-6 was positively correlated with frailty and DII, whereas IL-10 was negatively correlated. After adjusting for age, sex, body mass index, education level, smoking status, and energy, mediation analysis revealed that the association between DII and frailty was significantly mediated by IL-6 (average causal mediation effect [ACME], 0.052; 95% confidence interval, 0.020–0.087; P = .002) and IL-10 (ACME, 0.025; 95% confidence interval, 0.004–0.063; P = .016). The ρ values for the sensitivity measure at which estimated ACMEs were zero were 0.3 and –0.2 for IL-6 and IL-10, respectively. Therefore, a pro-inflammatory diet was associated with frailty in patients with CRC possibly in part by affecting circulating IL-6 and IL-10 concentrations.