Search

BackgroundCannabis exposures reported to the California Poison Control System increased following the initiation of recreational cannabis sales on 1 January 2018 (i.e., "commercialization"). We evaluated whether local cannabis control policies adopted by 2021 were associated with shifts in harmful cannabis exposures.MethodsUsing cannabis control policies collected for all 539 California cities and counties in 2020-2021, we applied a differences-in-differences design with negative binomial regression to test the association of policies with harmful cannabis exposures reported to California Poison Control System (2011-2020), before and after commercialization. We considered three policy categories: bans on storefront recreational retail cannabis businesses, overall restrictiveness, and specific recommended provisions (restricting product types or potency, packaging and labeling restrictions, and server training requirements).ResultsLocalities that ultimately banned storefront recreational retail cannabis businesses had fewer harmful cannabis exposures for children aged 13 years (rate ratio = 0.97; 95% confidence interval = 0.85, 1.11). Of 167 localities ultimately permitting recreational cannabis sales, overall restrictiveness was not associated with harmful cannabis exposures among children aged 13 years, a 1-standard deviation increase in ultimate restrictiveness was associated with fewer harmful cannabis exposures (rate ratio = 0.93; 95% confidence interval = 0.86, 1.01). For recommended provisions, estimates were generally too imprecise to detect associations with harmful cannabis exposures.ConclusionBans on storefront retail and other restrictive approaches to regulating recreational cannabis may be associated with fewer harmful cannabis exposures for some age groups following statewide commercialization.

BackgroundInfants experiencing bronchiolitis are at increased risk for asthma, but few studies have identified modifiable risk factors. We assessed whether early life air pollution influenced child asthma and wheeze at age 4-6 years among children with a history of bronchiolitis in the first postnatal year.MethodsChildren with caregiver-reported physician-diagnosed bronchiolitis were drawn from ECHO-PATHWAYS, a pooled longitudinal cohort from six US cities. We estimated their air pollution exposure from age 1 to 3 years from validated spatiotemporal models of fine particulate matter (PM 2.5 ), nitrogen dioxide (NO 2 ), and ozone (O 3 ). Caregivers reported children's current wheeze and asthma at age 4-6 years. We used modified Poisson regression to estimate relative risks (RR) and 95% confidence intervals (CI), adjusting for child, maternal, and home environmental factors. We assessed effect modification by child sex and maternal history of asthma with interaction models.ResultsA total of 224 children had caregiver-reported bronchiolitis. Median (interquartile range) 2-year pollutant concentrations were 9.3 (7.8-9.9) µg/m 3 PM 2.5 , 8.5 (6.4-9.9) ppb NO 2 , and 26.6 (25.6-27.7) ppb O 3 . RRs (CI) for current wheeze per 2-ppb higher O 3 were 1.3 (1.0-1.7) and 1.4 (1.1-1.8) for asthma. NO 2 was inversely associated with wheeze and asthma whereas associations with PM 2.5 were null. We observed interactions between NO 2 and PM 2.5 and maternal history of asthma, with lower risks observed among children with a maternal history of asthma.ConclusionOur results are consistent with the hypothesis that exposure to modest postnatal O 3 concentrations increases the risk of asthma and wheeze among the vulnerable subpopulation of infants experiencing bronchiolitis.

BackgroundEpidemiologic evidence linking prenatal exposure to per- and polyfluoroalkyl substances (PFAS) with altered neurodevelopment is inconclusive, and few large studies have focused on autism-related outcomes. We investigated whether blood concentrations of PFAS in pregnancy are associated with child autism-related outcomes.MethodsWe included 10 cohorts from the National Institutes of Health (NIH)-funded Environmental influences on Child Health Outcomes (ECHO) program (n = 1,429). We measured 14 PFAS analytes in maternal blood collected during pregnancy; eight analytes met detection criteria for analysis. We assessed quantitative autism-related traits in children via parent report on the Social Responsiveness Scale (SRS). In multivariable linear models, we examined relationships of each PFAS (natural log-transformed) with SRS scores. We further modeled PFAS as a complex mixture using Bayesian methods and examined modification of these relationships by child sex.ResultsMost PFAS in maternal blood were not associated with child SRS T-scores. Perfluorononanoic acid (PFNA) showed the strongest and most consistent association: each 1-unit increase in ln-transformed PFNA was associated with greater autism-related traits (adjusted β [95% confidence interval (CI)] = 1.5 [-0.1, 3.0]). The summed mixture, which included six PFAS detected in >70% of participants, was not associated with SRS T-scores (adjusted β [95% highest posterior density interval] = 0.7 [-1.4, 3.0]). We did not observe consistent evidence of sex differences.ConclusionsPrenatal blood concentrations of PFNA may be associated with modest increases in child autism-related traits. Future work should continue to examine the relationship between exposures to both legacy and emerging PFAS and additional dimensional, quantitative measures of childhood autism-related outcomes.

Background: Although the indoor environment has been proposed to be associated with childhood sleep health, to our knowledge no study has investigated the association between home renovation and childhood sleep problems. Methods: The study included 186,470 children aged 6-18 years from the National Chinese Children Health Study (2012-2018). We measured childhood sleeping problems via the Chinese version of the Sleep Disturbance Scale for Children (C-SDSC). Information on home renovation exposure within the recent 2 years was collected via parent report. We estimated associations between home renovation and various sleeping problems, defined using both continuous and categorized (binary) C-SDSC t-scores, using generalized mixed models. We fitted models with city as a random effect variable, and other covariates as fixed effects. Results: Out of the overall participants, 89,732 (48%) were exposed to recent home renovations. Compared to the unexposed group, children exposed to home renovations had higher odds of total sleep disorder (odd ratios [OR] = 1.3; 95% confidence interval [CI] = 1.2, 1.4). Associations varied when we considered different types of home renovation materials. Children exposed to multiple types of home renovation had higher odds of sleeping problems. We observed similar findings when considering continuous C-SDSC t-scores. Additionally, sex and age of children modified the associations of home renovation exposure with some of the sleeping problem subtypes. Conclusions: We found that home renovation was associated with higher odds of having sleeping problems and that they varied when considering the type of renovation, cumulative exposure, sex, and age differences. [ABSTRACT FROM AUTHOR]

BackgroundCannabis outlets may affect health and health disparities. Local governments can regulate outlets, but little is known about the effectiveness of local policies in limiting outlet densities and discouraging disproportionate placement of outlets in vulnerable neighborhoods.MethodsFor 241 localities in California, we measured seven policies pertaining to density or location of recreational cannabis outlets. We geocoded outlets using web-scraped data from the online finder Weedmaps between 2018 and 2020. We applied Bayesian spatiotemporal models to evaluate associations of local cannabis policies with Census block group-level outlet counts, accounting for confounders and spatial autocorrelation. We assessed whether associations differed by block group median income or racial-ethnic composition.ResultsSeventy-six percent of localities banned recreational cannabis outlets. Bans were associated with fewer outlets, particularly in block groups with higher median income, fewer Hispanic residents, and more White and Asian residents. Outlets were disproportionately located in block groups with lower median income [posterior RR (95% credible interval): 0.76 (0.70, 0.82) per $10,000], more Hispanic residents [1.05 (1.02, 1.09) per 5%], and fewer Black residents [0.91 (0.83, 0.98) per 5%]. For the six policies in jurisdictions permitting outlets, two policies were associated with fewer outlets and two with more; two policy associations were uninformative. For these policies, we observed no consistent heterogeneity in associations by median income or racial-ethnic composition.ConclusionsSome local cannabis policies in California are associated with lower cannabis outlet densities, but are unlikely to deter disproportionate placement of outlets in racial-ethnic minority and low-income neighborhoods.

BackgroundBicycling is an important form of physical activity in populations. Research assessing the effect of infrastructure on bicycling with high-resolution smartphone data is emerging in several places, but it remains limited in low-bicycling US settings, including the Southeastern US. The Atlanta area has been expanding its bicycle infrastructure, including off-street paved trails such as the Atlanta BeltLine and some protected bike lanes.MethodsUsing the generalized synthetic-control method, we estimated effects of five groups of off-street paved trails and protected bike lanes on bicycle ridership in their corresponding areas. To measure bicycling, we used 2 years (October 1, 2016 to September 30, 2018) of monthly Strava data in Atlanta's urban core along with data from 15 on-the-ground counters to adjust for spatiotemporal variation in app use.ResultsConsidering all infrastructure as one joint intervention, an estimated 1.10 (95% confidence interval [CI]: 0.99, 1.18) times more bicycle-distance was ridden than would have been expected in the same areas had the infrastructure not been built, when defining treatment areas by the narrower of two definitions (defined in text). The Atlanta BeltLine Westside Trail and Proctor Creek Greenway had especially strong effect estimates, e.g., ratios of 1.45 (95% CI: 1.12, 1.86) and 1.55 (1.10, 2.14) under each treatment-area definition, respectively. We estimated that other infrastructure had weaker positive or no effects on bicycle-distance ridden.ConclusionsThis study advances research on the topic because of its setting in the US Southeast, simultaneous assessment of several infrastructure groups, and data-driven approach to estimating effects. See video abstract at, http://links.lww.com/EDE/B936.

BackgroundDichlorodiphenyltrichloroethane (DDT) or pyrethroid insecticides are sprayed inside dwellings for malaria vector control, resulting in high exposure to millions of people, including pregnant women. These chemicals disrupt endocrine function and may affect child growth. To our knowledge, few studies have investigated the potential impact of prenatal exposure to DDT or pyrethroids on growth trajectories.MethodsWe investigated associations between gestational insecticide exposure and child growth trajectories in the Venda Health Examination of Mothers, Babies and their Environment, a birth cohort of 751 children born between 2012 and 2013 in South Africa. Based on child weight measured at follow-up and abstracted from medical records, we modeled weight trajectories from birth to 5 years using SuperImposition, Translation and Rotation, which estimated two child-specific parameters: size (average weight) and tempo (age at peak weight velocity). We estimated associations between peripartum maternal concentrations of serum DDT, dichlorodiphenyldichloroethylene, or urinary pyrethroid metabolites and SuperImposition, Translation and Rotation parameters using marginal structural models.ResultsWe observed that a 10-fold increase in maternal concentrations of the pyrethroid metabolite trans-3-(2,2,-dicholorvinyl)-2,2-dimethyl-cyclopropane carboxylic acid was associated with a 21g (95% confidence interval = -40, -1.6) smaller size among boys but found no association among girls (Pinteraction = 0.07). Estimates suggested that pyrethroids may be associated with earlier tempo but were imprecise. We observed no association with serum DDT or dichlorodiphenyldichloroethylene.ConclusionsInverse associations between pyrethroids and weight trajectory parameters among boys are consistent with hypothesized disruption of androgen pathways and with our previous research in this population, and support the endocrine-disrupting potential of pyrethroids in humans.

BackgroundChronic pain is the leading cause of disability worldwide and is strongly associated with the epidemic of opioid overdosing events. However, the causal links between chronic pain, opioid prescriptions, and mortality remain unclear.MethodsThis study included 13,884 US adults aged ≥20 years who provided data on chronic pain in the National Health and Nutrition Examination Survey 1999-2004 with linkage to mortality databases through 2015. We employed the generalized form of the front-door formula within the structural causal model framework to investigate the causal effect of chronic pain on all-cause mortality mediated by opioid prescriptions.ResultsWe identified a total of 718 participants at 3 years of follow-up and 1260 participants at 5 years as having died from all causes. Opioid prescriptions increased the risk of all-cause mortality with an estimated odds ratio (OR) (95% confidence interval) = 1.5 (1.1, 1.9) at 3 years and 1.3 (1.1, 1.6) at 5 years. The front-door formula revealed that chronic pain increased the risk of all-cause mortality through opioid prescriptions; OR = 1.06 (1.01, 1.11) at 3 years and 1.03 (1.01, 1.06) at 5 years. Our bias analysis showed that our findings based on the front-door formula were likely robust to plausible sources of bias from uncontrolled exposure-mediator or mediator-outcome confounding.ConclusionsChronic pain increased the risk of all-cause mortality through opioid prescriptions. Our findings highlight the importance of careful guideline-based chronic pain management to prevent death from possibly inappropriate opioid prescriptions driven by chronic pain.

Accurate measurement of daily infection incidence is crucial to epidemic response. However, delays in symptom onset, testing, and reporting obscure the dynamics of transmission, necessitating methods to remove the effects of stochastic delays from observed data. Existing estimators can be sensitive to model misspecification and censored observations; many analysts have instead used methods that exhibit strong bias. We develop an estimator with a regularization scheme to cope with stochastic delays, which we term the robust incidence deconvolution estimator. We compare the method to existing estimators in a simulation study, measuring accuracy in a variety of experimental conditions. We then use the method to study COVID-19 records in the United States, highlighting its stability in the face of misspecification and right censoring. To implement the robust incidence deconvolution estimator, we release incidental, a ready-to-use R implementation of our estimator that can aid ongoing efforts to monitor the COVID-19 pandemic.

Postauthorization observational studies play a key role in understanding COVID-19 vaccine effectiveness following the demonstration of efficacy in clinical trials. Although bias due to confounding, selection bias, and misclassification can be mitigated through careful study design, unmeasured confounding is likely to remain in these observational studies. Phase III trials of COVID-19 vaccines have shown that protection from vaccination does not occur immediately, meaning that COVID-19 risk should be similar in recently vaccinated and unvaccinated individuals, in the absence of confounding or other bias. Several studies have used the estimated effectiveness among recently vaccinated individuals as a negative control exposure to detect bias in vaccine effectiveness estimates. In this paper, we introduce a theoretical framework to describe the interpretation of such a bias indicator in test-negative studies, and outline strong assumptions that would allow vaccine effectiveness among recently vaccinated individuals to serve as a negative control exposure.

Background: Despite its potential value in register-based pharmacoepidemiologic research, recorded information on "indication for use" (INDO) in the Danish National Prescription Registry has rarely been used, likely because of questions about the variable's validity, which to our knowledge no study has systematically assessed. Methods: We extracted data on 80,814 prescriptions from the software systems (PharmaNet and C2) of five Danish community pharmacies filled between 4 and 16 February 2019 and 2020. Using the indication information recorded in the pharmacy software systems as the gold standard, we evaluated the extent and quality of the corresponding information from the Prescription Registry. Results: Of all prescriptions identified, we captured >99% in the Prescription Registry. The proportion of prescriptions with recorded indication codes in the Prescription Registry was 82% (n = 66,164) but was lower for C2 than PharmaNet. Correcting for the overrepresentation of C2 data in our sample, the estimated proportion of registration was ≈88%. Almost 100% (66,158 of 66,164) of the prescriptions with recorded indication codes in the Prescription Registry had correctly recorded indication codes. Nonspecific indication codes were present in 5.6%–36% of selected drugs and drug classes. Conclusions: Prescriptions filled at Danish community pharmacies are accurately captured by the Danish National Prescription Registry, and the recorded information on indication is generally valid and usable in research. However, minor concerns remain about missingness, nonspecific recorded indication codes, and lower validity, and a higher proportion of missingness of recorded indication codes is expected before 2017. A supplementary digital video by the article's co-author, Hanin Harbi, is available at http://links.lww.com/EDE/C89 [ABSTRACT FROM AUTHOR]

RationaleAsthma and obesity often co-occur. It has been hypothesized that asthma may contribute to childhood obesity onset.ObjectivesTo determine if childhood asthma is associated with incident obesity and examine the role of asthma medication in this association.MethodsWe studied 8,716 children between ages 6 and 18.5 years who were nonobese at study entry participating in 18 US cohorts of the Environmental influences on Child Health Outcomes program (among 7,299 children with complete covariate data mean [SD] study entry age = 7.2 [1.6] years and follow up = 5.3 [3.1] years).Measurements and main resultsWe defined asthma based on caregiver report of provider diagnosis. Incident obesity was defined as the first documented body mass index ≥95th percentile for age and sex following asthma status ascertainment. Over the study period, 26% of children had an asthma diagnosis and 11% developed obesity. Cox proportional hazards models with sex-specific baseline hazards were fitted to assess the association of asthma diagnosis with obesity incidence. Children with asthma had a 23% (95% confidence intervals [CI] = 4, 44) higher risk for subsequently developing obesity compared with those without asthma. A novel mediation analysis was also conducted to decompose the total asthma effect on obesity into pathways mediated and not mediated by asthma medication use. Use of asthma medication attenuated the total estimated effect of asthma on obesity by 64% (excess hazard ratios = 0.64; 95% CI = -1.05, -0.23).ConclusionsThis nationwide study supports the hypothesis that childhood asthma is associated with later risk of obesity. Asthma medication may reduce this association and merits further investigation as a potential strategy for obesity prevention among children with asthma.

BackgroundGeoffrey Rose's paper "Sick Individuals, Sick Populations" highlights the counterintuitive finding that the largest share of morbidity arises from populations engaging in low- to moderate-risk behavior. Scholars refer to this finding as the prevention paradox. We examine whether this logic applies to SARS-CoV-2 infected persons considered low to moderate risk.MethodsWe conducted a population-representative survey and sero-surveillance study for SARS-CoV-2 among adults in Orange County, California. Participants answered questions about health behaviors and provided a finger-pin-prick sample from 10 July to 16 August 2020.ResultsOf the 2979 adults, those reporting low- and moderate-risk behavior accounted for between 78% and 92% of SARS-CoV-2 infections. Asymptomatic individuals, as well as persons with low and moderate scores for self-reported likelihood of having had SARS-CoV-2, accounted for the majority of infections.ConclusionsOur findings support Rose's logic, which encourages public health measures among persons who self-identify as unlikely to have SARS-CoV-2. See video abstract at, http://links.lww.com/EDE/B860.

Background: Trachoma control programs use multiple approaches to identify individuals with trachomatous trichiasis (TT). Evidence is limited regarding which approaches are most effective and cost-efficient. Methods: We evaluated the effectiveness of two TT case-identification approaches in Ethiopia: community mobilization to encourage self-referral for centralized screening and house-to-house screenings conducted by case finders. We compared the number of true cases found per 1000 population and costs associated with case identification under each approach, stratified by villages that received one or multiple screening visits. Results: We conducted screenings in 396 villages. In villages receiving one house-to-house visit, case finders identified 14,229 suspected cases, of whom 10,513 (73.9%) presented for TT confirmation. A median of 17.2% (interquartile range [IQR]: 9.1%-27.8%) of those presenting truly had TT (positive predictive value). In single-visit villages, the community mobilization approach yielded higher rates of confirmed cases than the house-to-house approach (1.5 [IQR: 1.1, 2.6] vs. 1.1 [IQR: 0.5, 1.9] cases per 1000 population), and the median cost of identifying a TT case was less ($5.59 vs. $31.18) using community mobilization than house-to-house. In multiple-visit villages, additional screening visits increased the median rate of confirmed cases to 2.5 per 1000 population in community mobilization villages, but the rate remained unchanged in house-to-house villages. Conclusions: Community mobilization-based TT case finding had a higher yield than house-to-house, at a substantially lower cost. Future research should examine whether additional tools to aid case finders in their diagnosis increases case-finding efficiency and accuracy and whether TT prevalence and surgical program duration impact case-finding success. [ABSTRACT FROM AUTHOR]

Background: Little is known about the role of air quality in fatal asthma exacerbations among children. Methods: We collected information about 80 deaths that occurred in North Carolina from 2001 through 2016, among children aged 5-17 years, with asthma identified as the primary cause of death. We linked information about each death with county-level estimates of particulate matter =2.5 µm (PM2.5) and ozone (O3). Using the linked data, we conducted a case-crossover analysis of associations between PM2.5 and O3 lagged by 3-5 days with the odds of fatal asthma exacerbations. Results: In the highest tertile of PM2.5 lag(3-5), the odds of a fatal exacerbation of asthma were more than twice the odds in the lowest tertile (odds ratio = 2.2; 95% confidence interval = 1.1, 4.6). Conclusion: These findings from North Carolina provide evidence to support the hypothesis that ambient air pollution increases the risk of fatal exacerbations of asthma among children. [ABSTRACT FROM AUTHOR]

Background: The Berlin-based B_PROUD study was designed to assess the effect of mobile stroke unit (MSU) dispatch among ischemic stroke and transient ischemic attack (TIA) patients without contraindications to reperfusion treatments. However, a large proportion of patients for whom the MSU was dispatched did not ultimately receive MSU care. We estimated the causal effect of additional MSU care on 3-month functional outcomes among B_PROUD patients for whom an MSU was dispatched. Methods: We used data from the B_PROUD study (1 February 2017–8 May 2019). Given the presence of exposure–outcome unmeasured confounding, we used the front-door formula to identify the distribution of modified Rankin scale (mRS) outcomes under two hypothetical interventions: (1) receiving additional MSU care and (2) only receiving conventional care. We considered the time from dispatch to thrombolysis as the full mediator and adjusted for exposure–mediator and mediator–outcome confounding. We used a parametric estimator to estimate the common odds ratio (cOR) and 95% bootstrapped confidence intervals (CI). Results: We included in total 768 ischemic stroke/TIA patients with MSU dispatch. The MSU was canceled for 180 (23%) patients, whereas 588 (77%) received MSU care. The unadjusted association between the care group and mRS favored conventional care (cOR = 1.7; 95% CI = 1.2, 2.3); however, after applying the front-door formula, the mRS distribution favored MSU care (cOR = 0.88; 95% CI = 0.81, 0.96). Conclusions: Receiving MSU care was associated with better functional outcomes than conventional care only, compatible with the hypothesized beneficial effect of MSU care on poststroke outcomes, among stroke and TIA patients without contraindications to reperfusion treatments. [ABSTRACT FROM AUTHOR]

BackgroundSocial isolation among HIV-positive persons might be an important barrier to care. Using data from the SEARCH Study in rural Kenya and Uganda, we constructed 32 community-wide, sociocentric networks and evaluated whether less socially connected HIV-positive persons were less likely to know their status, have initiated treatment, and be virally suppressed.MethodsBetween 2013 and 2014, 168,720 adult residents in the SEARCH Study were census-enumerated, offered HIV testing, and asked to name social contacts. Social networks were constructed by matching named contacts to other residents. We characterized the resulting networks and estimated risk ratios (aRR) associated with poor HIV care outcomes, adjusting for sociodemographic factors and clustering by community with generalized estimating equations.ResultsThe sociocentric networks contained 170,028 residents (nodes) and 362,965 social connections (edges). Among 11,239 HIV-positive persons who named ≥1 contact, 30.9% were previously undiagnosed, 43.7% had not initiated treatment, and 49.4% had viral nonsuppression. Lower social connectedness, measured by the number of persons naming an HIV-positive individual as a contact (in-degree), was associated with poorer outcomes in Uganda, but not Kenya. Specifically, HIV-positive persons in the lowest connectedness tercile were less likely to be previously diagnosed (Uganda-West aRR: 0.89 [95% confidence interval (CI): 0.83, 0.96]; Uganda-East aRR: 0.85 [95% CI: 0.76, 0.96]); on treatment (Uganda-West aRR: 0.88 [95% CI: 0.80, 0.98]; Uganda-East aRR: 0.81 [0.72, 0.92]), and suppressed (Uganda-West aRR: 0.84 [95% CI: 0.73, 0.96]; Uganda-East aRR: 0.74 [95% CI: 0.58, 0.94]) than those in the highest connectedness tercile.ConclusionsHIV-positive persons named as a contact by fewer people may be at higher risk for poor HIV care outcomes, suggesting opportunities for targeted interventions.

BackgroundCollaborative research often combines findings across multiple, independent studies via meta-analysis. Ideally, all study estimates that contribute to the meta-analysis will be equally unbiased. Many meta-analyses require all studies to measure the same covariates. We explored whether differing minimally sufficient sets of confounders identified by a directed acyclic graph (DAG) ensures comparability of individual study estimates. Our analysis applied four statistical estimators to multiple minimally sufficient adjustment sets identified in a single DAG.MethodsWe compared estimates obtained via linear, log-binomial, and logistic regression and inverse probability weighting, and data were simulated based on a previously published DAG.ResultsOur results show that linear, log-binomial, and inverse probability weighting estimators generally provide the same estimate of effect for different estimands that are equally sufficient to adjust confounding bias, with modest differences in random error. In contrast, logistic regression often performed poorly, with notable differences in effect estimates obtained from unique minimally sufficient adjustment sets, and larger standard errors than other estimators.ConclusionsOur findings do not support the reliance of collaborative research on logistic regression results for meta-analyses. Use of DAGs to identify potentially differing minimally sufficient adjustment sets can allow meta-analyses without requiring the exact same covariates.

Host adaptive immune responses may protect against infection or disease when a pathogen is repeatedly encountered. The hazard ratio of infection or disease, given previous infection, is typically sought to estimate the strength of protective immunity. However, variation in individual exposure or susceptibility to infection may introduce frailty bias, whereby a tendency for infections to recur among individuals with greater risk confounds the causal association between previous infection and susceptibility. We introduce a self-matched "case-only" inference method to control for unmeasured individual heterogeneity, making use of negative-control endpoints not attributable to the pathogen of interest. To control for confounding, this method compares event times for endpoints due to the pathogen of interest and negative-control endpoints during counterfactual risk periods, defined according to individuals' infection history. We derive a standard Mantel-Haenszel (matched) odds ratio conveying the effect of prior infection on time to recurrence. We compare performance of this approach to several proportional hazards modeling frameworks and estimate statistical power of the proposed strategy under various conditions. In an example application, we use the proposed method to reestimate naturally acquired protection against rotavirus gastroenteritis using data from previously published cohort studies. This self-matched negative-control design may present a flexible alternative to existing approaches for analyzing naturally acquired immunity, as well as other exposures affecting the distribution of recurrent event times.

Transient exposures are difficult to measure in epidemiologic studies, especially when both the status of being at risk for an outcome and the exposure change over time and space, as when measuring built-environment risk on transportation injury. Contemporary "big data" generated by mobile sensors can improve measurement of transient exposures. Exposure information generated by these devices typically only samples the experience of the target cohort, so a case-control framework may be useful. However, for anonymity, the data may not be available by individual, precluding a case-crossover approach. We present a method called at-risk-measure sampling. Its goal is to estimate the denominator of an incidence rate ratio (exposed to unexposed measure of the at-risk experience) given an aggregated summary of the at-risk measure from a cohort. Rather than sampling individuals or locations, the method samples the measure of the at-risk experience. Specifically, the method as presented samples person-distance and person-events summarized by location. It is illustrated with data from a mobile app used to record bicycling. The method extends an established case-control sampling principle: sample the at-risk experience of a cohort study such that the sampled exposure distribution approximates that of the cohort. It is distinct from density sampling in that the sample remains in the form of the at-risk measure, which may be continuous, such as person-time or person-distance. This aspect may be both logistically and statistically efficient if such a sample is already available, for example from big-data sources like aggregated mobile-sensor data.

BackgroundWomen can be exposed to a multitude of hardships before and during pregnancy that may affect fetal growth, but previous approaches have not analyzed them jointly as social exposure mixtures.MethodsWe evaluated the independent, mutually adjusted, and pairwise joint associations between self-reported hardships and birthweight for gestational age z-scores in the Chemicals in Our Bodies-2 prospective birth cohort (N = 510) using G-computation. We examined financial hardship, food insecurity, job strain, poor neighborhood environment, low community standing, caregiving, high burden of stressful life events, and unplanned pregnancy collected via questionnaire administered in the second trimester of pregnancy. We used propensity scores to ensure our analyses had sufficient data support and estimated absolute differences in outcomes.ResultsFood insecurity was most strongly associated with reduced birthweight for gestational age z-scores individually, with an absolute difference of -0.16, 95% confidence interval (CI) -0.45, 0.14. We observed an unexpected increase in z-scores associated with poor perceived neighborhood environment (0.18, 95% CI -0.04, 0.41). Accounting for coexposures resulted in similar findings. The pairwise joint effects were strongest for food insecurity in combination with unplanned pregnancy (-0.45, 95% CI -0.93, 0.02) and stressful life events (-0.42, 95% CI -0.90, 0.05). Poor neighborhood environment in combination with caregiving was associated with an increase in z-scores (0.47, 95% CI -0.01, 0.95).ConclusionsOur results are consistent with the hypothesis that experiencing food insecurity during pregnancy, alone and in combination with stressful life events and unplanned pregnancy, may affect fetal growth.

Public health and the underlying disease processes are complex, often involving the interaction of biologic, social, psychologic, economic, and other processes that may be nonlinear and adaptive and have other features of complex systems. There is therefore a need to push the boundaries of public health beyond single-factor data analysis and expand the capacity of research methodology to tackle real-world complexities. This article sets out a way to operationalize complex systems thinking in public health, with a particular focus on how epidemiologic methods and data can contribute towards this end. Our proposed framework comprises three core dimensions-- patterns, mechanisms, and dynamics--along which complex systems may be conceptualized. These dimensions cover seven key features of complex systems--emergence, interactions, nonlinearity, interference, feedback loops, adaptation, and evolution. We relate this framework to examples of methods and data traditionally used in epidemiology. We conclude that systematic production of knowledge on complex health issues may benefit from: formulation of research questions and programs in terms of the core dimensions we identify, as a comprehensive way to capture crucial features of complex systems; integration of traditional epidemiologic methods with systems methodology such as computational simulation modeling; interdisciplinary work; and continued investment in a wide range of data types. We believe that the proposed framework can support the systematic production of knowledge on complex health problems, with the use of epidemiology and other disciplines. This will help us understand emergent health phenomena, identify vulnerable population groups, and detect leverage points for promoting public health. [ABSTRACT FROM AUTHOR]

BackgroundRecently, it has been suggested that environmental exposures from traffic sources including noise may play a role in cognitive impairment in the elderly. The objective of the study was to investigate the association between local traffic-related noise pollution and incident dementia or cognitive impairment without dementia (CIND) during a 10-year follow-up period.Methods1612 Mexican-American participants from the Sacramento Area Latino Study on Aging (SALSA) were followed every 12-15 months via home visits from 1998 to 2007. We used the SoundPLAN software package to estimate noise originating from local traffic with the input of Annual Average Daily Traffic (AADT) data from Metropolitan Planning Organizations (MPO) based on geocoded residential addresses at baseline (1998-1999). We estimated the risks of incident dementia or CIND from 24-hour and nighttime noise exposure using Cox proportional hazard models.ResultsDuring the follow-up, we identified 159 incident dementia or CIND cases in total. Per 11.6 dB (interquartile range width) increase in 24-hour noise, the hazard of developing dementia or CIND increased (hazard ratio = 1.3 [1.0, 1.6]) during follow-up; estimates were slightly lower (hazard ratio = 1.2 [0.97, 1.6]) when adjusting for modeled local air pollution exposure from traffic sources. Overall, the risk of dementia/CIND was elevated when 24-hour and nighttime noise were higher than 75 and 65 dB respectively. See video Abstract: http://links.lww.com/EDE/B728.ConclusionsIn our study, traffic-related noise exposure was associated with increased risk of dementia or CIND in elderly Mexican-Americans. Future studies taking into account other noise sources and occupational noise exposure before retirement are needed.

BackgroundPopulation-level estimates of disease prevalence and control are needed to assess prevention and treatment strategies. However, available data often suffer from differential missingness. For example, population-level HIV viral suppression is the proportion of all HIV-positive persons with suppressed viral replication. Individuals with measured HIV status, and among HIV-positive individuals those with measured viral suppression, likely differ from those without such measurements.MethodsWe discuss three sets of assumptions to identify population-level suppression in the intervention arm of the SEARCH Study (NCT01864603), a community randomized trial in rural Kenya and Uganda (2013-2017). Using data on nearly 100,000 participants, we compare estimates from (1) an unadjusted approach assuming data are missing-completely-at-random (MCAR); (2) stratification on age group, sex, and community; and (3) targeted maximum likelihood estimation to adjust for a larger set of baseline and time-updated variables.ResultsDespite high measurement coverage, estimates of population-level viral suppression varied by identification assumption. Unadjusted estimates were most optimistic: 50% (95% confidence interval [CI] = 46%, 54%) of HIV-positive persons suppressed at baseline, 80% (95% CI = 78%, 82%) at year 1, 85% (95% CI = 83%, 86%) at year 2, and 85% (95% CI = 83%, 87%) at year 3. Stratifying on baseline predictors yielded slightly lower estimates, and full adjustment reduced estimates meaningfully: 42% (95% CI = 37%, 46%) of HIV-positive persons suppressed at baseline, 71% (95% CI = 69%, 73%) at year 1, 76% (95% CI = 74%, 78%) at year 2, and 79% (95% CI = 77%, 81%) at year 3.ConclusionsEstimation of population-level disease burden and control requires appropriate adjustment for missing data. Even in large studies with limited missingness, estimates relying on the MCAR assumption or baseline stratification should be interpreted cautiously.

BackgroundThe observance of nonrandom space-time groupings of childhood cancer has been a concern of health professionals and the general public for decades. Many childhood cancers are suspected to have initiated in utero; therefore, we examined the spatial-temporal randomness of the birthplace of children who later developed cancer.MethodsWe performed a space-time cluster analysis using birth addresses of 5,896 cases and 23,369 population-based, age-, sex-, and race/ethnicity-matched controls in California from 1997 to 2007, evaluating 20 types of childhood cancer and three a priori designated subgroups of childhood acute lymphoblastic leukemia (ALL). We analyzed data using a newly designed semiparametric analysis program, ClustR, and a common algorithm, SaTScan.ResultsWe observed evidence for nonrandom space-time clustering for ALL diagnosed at 2-6 years of age in the South San Francisco Bay Area (ClustR P = 0.04, SaTScan P = 0.07), and malignant gonadal germ cell tumors in a region of Los Angeles (ClustR P = 0.03, SaTScan P = 0.06). ClustR did not identify evidence of clustering for other childhood cancers, although SaTScan suggested some clustering for Hodgkin lymphoma (P = 0.09), astrocytoma (P = 0.06), and retinoblastoma (P = 0.06).ConclusionsOur study provides evidence that childhood ALL diagnosed at 2-6 years and malignant gonadal germ cell tumors sporadically occurs in nonrandom space-time clusters. Further research is warranted to identify epidemiologic features that may inform the underlying etiology.

BackgroundPrescription drug monitoring programs (PDMPs) that collect and distribute information on dispensed controlled substances have been adopted by nearly all US states. We know little about program characteristics that modify PDMP impact on prescription opioid (PO) overdose deaths.MethodsWe measured associations between adoption of any PDMP and changes in fatal PO overdoses in 2002-2016 across 3109 counties in 49 states and D.C. We then measured changes related to the adoption of "proactive PDMPs," which report outlying prescribing/dispensing patterns and provide broader access to PDMP data by law enforcement. Comparisons were made within 3 time intervals that broadly represent the evolution of PDMPs (2002-2004, 2005-2009, and 2010-2016). We modeled overdoses using Bayesian space-time models.ResultsAdoption of electronic PDMP access was associated with 9% lower rates of fatal PO overdoses after three years (rate ratio [RR] = 0.91, 95% credible interval [CI]: 0.88-0.93) with well-supported effects for methadone (RR = 0.86,95% CI: 0.82-0.90) and other synthetic opioids (RR = 0.82, 95% CI: 0.77-0.86). Compared with states with no/weak PDMPs, proactive PDMPs were associated with fewer deaths attributed to natural/semi-synthetic opioids (2002-2004: RR = 0.72 [0.66-0.78]; 2005-2009: RR = 0.93 [0.90-0.97]; 2010-2016: 0.89 [0.86-0.92]) and methadone (2002-2004: RR = 0.77 [0.69-0.85]; 2010-2016: RR = 0.90 [0.86-0.94]). Unintended effects were observed for synthetic opioids other than methadone (2005-2009: RR = 1.29 [1.21-1.38]; 2010-2016: RR = 1.22 [1.16-1.29]).ConclusionsState adoption of PDMPs was associated with fewer PO deaths overall while proactive PDMPs alone were associated with fewer deaths related to natural/semisynthetic opioids and methadone, the specific targets of these programs. See video abstract at, http://links.lww.com/EDE/B619.

BackgroundThe test-negative design is an increasingly popular approach for estimating vaccine effectiveness (VE) due to its efficiency. This review aims to examine published test-negative design studies of VE and to explore similarities and differences in methodological choices for different diseases and vaccines.MethodsWe conducted a systematic search on PubMed, Web of Science, and Medline, for studies reporting the effectiveness of any vaccines using a test-negative design. We screened titles and abstracts and reviewed full texts to identify relevant articles. We created a standardized form for each included article to extract information on the pathogen of interest, vaccine(s) being evaluated, study setting, clinical case definition, choices of cases and controls, and statistical approaches used to estimate VE.ResultsWe identified a total of 348 articles, including studies on VE against influenza virus (n = 253), rotavirus (n = 48), pneumococcus (n = 24), and nine other pathogens. Clinical case definitions used to enroll patients were similar by pathogens of interest but the sets of symptoms that defined them varied substantially. Controls could be those testing negative for the pathogen of interest, those testing positive for nonvaccine type of the pathogen of interest, or a subset of those testing positive for alternative pathogens. Most studies controlled for age, calendar time, and comorbidities.ConclusionsOur review highlights similarities and differences in the application of the test-negative design that deserve further examination. If vaccination reduces disease severity in breakthrough infections, particular care must be taken in interpreting vaccine effectiveness estimates from test-negative design studies.

Background: With the increasing prevalence of risk factors for nephrotic syndrome, updated epidemiologic data on the syndrome are needed. We examined its age- and sex-specific incidence, histopathology, and mortality over 24 years. Methods: This nationwide cohort study included all adults with first-time-recorded nephrotic syndrome in Denmark during 1995–2018 using the Danish National Patient Registry. We obtained data on age, sex, hospital-diagnosed comorbidities, and histopathologic findings. We computed overall, and age- and sex-specific, incidence rates of nephrotic syndrome, 1- and 5-year mortality by calendar period, and 1-year hazard ratios (HRs) of death using Cox models. Results: We identified 3,970 adults with first-time nephrotic syndrome diagnosis. Incidence was highest in men and increased with age to 11.77 per 100,000 person-years (95% confidence interval [CI]: 10.21–13.32) in men aged 80+ years, and 6.56 per 100,000 person-years (95% CI: 5.71–7.41) in women aged 80+ years. Incidence of nephrotic syndrome increased from 3.35 per 100,000 person-years (95% CI: 3.12–3.58) in 1995–2000 to 4.30 per 100,000 person-years (95% CI: 4.05–4.54) in 2013–2018. Over time, 1-year mortality of nephrotic syndrome was stable at 13%–16%, but HR of death was 0.54 (95% CI: 0.42–0.69), adjusted for age, sex, and comorbidities, in 2013–2018 compared with 1995–2000. Subdistribution of glomerulopathies was stable over time with membranous nephropathy and minimal change disease being the most common. Conclusion: During 1995–2018, the incidence of recorded adult nephrotic syndrome increased slightly, and the adjusted mortality of nephrotic syndrome decreased markedly. Whether these findings reflect changes in epidemiology or awareness and coding of nephrotic syndrome, remains to be clarified. [ABSTRACT FROM AUTHOR]

Background: Existing evidence on the link between smoking and appendicitis is scarce and ambiguous. We therefore conducted a population-based cohort study in Denmark to investigate whether smoking during pregnancy is associated with an increased risk of appendicitis in offspring. Methods: We used the Danish Birth Registry to include all singletons born during 1991–2017 and to identify maternal smoking status during pregnancy. We followed the children from birth until date of appendicitis, emigration, death, or administrative end of study (31 December 2018), whichever came first. We calculated crude and adjusted hazard ratios (HRs) of appendicitis with 95% confidence intervals (CIs) comparing children of mothers who smoked during pregnancy to children of nonsmokers. Further, we conducted a bias analysis and sibling analysis. Results: We included 1,659,526 singletons of whom 19% were born to mothers who smoked during pregnancy. After maximum 28 years of follow-up, hazard rates for children of smokers were slightly higher than for children of nonsmokers [adjusted HR: 1.07 (95% CI = 1.04, 1.10)]. Stratification by sex revealed no association for males [adjusted HR: 1.02 (95% CI = 0.99, 1.06)], but a higher HR for females [adjusted HR: 1.13 (95% CI = 1.09, 1.18)]. This association increased with increasing length of follow-up, indicating that the association may be mediated by later-life exposures. The bias analysis indicated that misclassification of maternal smoking could attenuate a true association, while the sibling analysis showed no association. Conclusions: Maternal smoking during pregnancy and appendicitis in the offspring may be associated. [ABSTRACT FROM AUTHOR]

BackgroundCommunity violence is an understudied aspect of social context that may affect risk of preterm birth and small-for-gestational age (SGA).MethodsWe matched California mothers with live singleton births who were exposed to a homicide in their Census tract of residence in 2007-2011 to unexposed mothers within the same tract. We estimated risk differences with a weighted linear probability model, with weights corresponding to the matched data structure. We estimated the average treatment effect on the treated of homicide exposure on the risk of preterm birth and SGA during the preconception period and first and second trimester.ResultsWe found a small increase in risk of SGA associated with homicide exposure in the first trimester (0.14% [95% confidence interval (CI) = -0.01%, 0.30%]), but not for exposure during the preconception period (-0.01% [95% CI = -0.17%, 0.15%]) or the second trimester (-0.06% [95% CI = -0.23%, 0.11%]). Risk of preterm birth was not affected by homicide exposure. When women were exposed to homicides during all three exposure windows, there was a larger increase in risk of SGA (1.09% [95% CI = 0.15%, 2.03%]) but not preterm birth (0.14% [95% CI = -0.74%, 1.01%]). Exposure to three or more homicides was also associated with greater risk of SGA (0.78% [95% CI = 0.15%, 1.40%]). Negative controls indicated that residual confounding by temporal patterning was unlikely.ConclusionsHomicide exposure during early pregnancy is associated with a small increased risk of SGA.

BackgroundIndustrial blue-collar workers face multiple work-related stressors, but evidence regarding the burden of mental illness among today's blue-collar men and women remains limited.MethodsIn this retrospective cohort study, we examined health and employment records for 37,183 blue- and white-collar workers employed by a single US aluminum manufacturer from 2003 to 2013. Using Cox proportional hazards regression, we modeled time to first episode of treated depression by gender and occupational class. Among cases, we modeled rates of depression-related service utilization with generalized gamma regression.ResultsCompared with their white-collar counterparts, blue-collar men were more likely to be treated for depression (hazard ratio [HR] = 1.3; 95% confidence interval [CI] = 1.1, 1.4) as were blue-collar women (HR = 1.4; 1.2, 1.6). Blue-collar women were most likely to be treated for depression as compared with white-collar men (HR = 3.2; 95% CI = 2.1, 5.0). However, blue-collar workers used depression-related services less frequently than their white-collar counterparts among both men (rate ratio = 0.91; 95% CI = 0.84, 0.98) and women (rate ratio = 0.82; 95% CI = 0.77, 0.88).ConclusionsBlue-collar women were more likely to be treated for depression than white-collar workers, and blue-collar women were most likely to be treated for depression compared with white-collar men. However, blue-collar men and women used depression-related healthcare services less frequently than white-collar workers. These findings underscore that blue-collar women may be uniquely susceptible to depression, and suggest that blue-collar workers may encounter barriers to care-seeking related mental illness other than their insurance status.

BACKGROUND:Prescription drug monitoring program are designed to reduce harms from prescription opioids; however, little is known about what populations benefit the most from these programs. We investigated how the relation between implementation of online prescription drug monitoring programs and rates of hospitalizations related to prescription opioids and heroin overdose changed over time, and varied across county levels of poverty and unemployment, and levels of medical access to opioids. METHODS:Ecologic county-level, spatiotemporal study, including 990 counties within 16 states, in 2001-2014. We modeled overdose counts using Bayesian hierarchical Poisson models. We defined medical access to opioids as the county-level rate of hospital discharges for noncancer pain conditions. RESULTS:In 2010-2014, online prescription drug monitoring programs were associated with lower rates of prescription opioid-related hospitalizations (rate ratio 2014 = 0.74; 95% credible interval = 0.69, 0.80). The association between online prescription drug monitoring programs and heroin-related hospitalization was also negative but tended to increase in later years. Counties with lower rates of noncancer pain conditions experienced a lower decrease in prescription opioid overdose and a faster increase in heroin overdoses. No differences were observed across different county levels of poverty and unemployment. CONCLUSIONS:Areas with lower levels of noncancer pain conditions experienced the smallest decrease in prescription opioid overdose and the faster increase in heroin overdose following implementation of online prescription drug monitoring programs. Our results are consistent with the hypothesis that prescription drug monitoring programs are most effective in areas where people are likely to access opioids through medical providers.

BackgroundMelanoma, the deadliest form of skin cancer, is the second most common cancer diagnosed before age 30. Little is known about potentially modifiable or intervenable risk factors specific to developing melanoma at a young age. The objective was to determine if high birth weight or higher early-life ultraviolet (UV) radiation exposure would be associated with increased risk of melanoma in young patients.MethodsPopulation-based, case-control study of 1,396 cases of melanoma diagnosed before age 30 in 1988-2013 and 27,920 controls, obtained by linking cancer registry data to birth records in California.ResultsHigh birth weight (>4,000 g) was associated with 19% higher risk of melanoma (OR = 1.19; 95% CI = 1.02, 1.39), while low birth weight (

BackgroundPeople living with HIV are at risk of increased myocardial infarction (MI). Cumulative HIV viral load (VL) has been proposed as a better measure of HIV inflammation than other measures of VL, like baseline VL, but its associations with MI are not known.MethodsThe multisite Centers for AIDS Research Network of Integrated Clinical Systems (CNICS) cohort includes clinical data and centrally adjudicated MI with distinction between atheroembolic MI (type 1) and MI related to supply-demand mismatch (type 2). We examined CNICS participants who were not on antiretroviral therapy (ART) at enrollment. Cumulative VL (copy-days of virus) from 6 months after enrollment was estimated with a time-weighted sum using the trapezoidal rule. We modeled associations of cumulative and baseline VL with MI by type using marginal structural Cox models. We contrasted the 75% percentile of the VL distribution with the 25% percentile.ResultsAmong 11,324 participants, 218 MIs occurred between 1996 and 2016. Higher cumulative VL was associated with risk of all MI (hazard ratio [HR] = 1.72; 95% confidence interval [CI] = 1.26, 2.36), type 1 MI (HR = 1.23; 95% CI = 0.78, 1.96), and type 2 MI (HR = 2.52; 95% CI = 1.74, 3.66). While off ART, cumulative VL had a stronger association with type 1 MI (HR = 2.13; 95% CI = 1.15, 3.94) than type 2 MI (HR = 1.25; 95% CI = 0.70, 2.25). Baseline VL was associated with all MI (HR = 1.60; 95% CI = 1.28, 2.01), type 1 MI (HR = 1.73; 95% CI = 1.26, 2.38), and type 2 MI (HR = 1.51; 95% CI = 1.10, 2.08).ConclusionsHigher cumulative and baseline VL is associated with all MI, with a particularly strong association between cumulative VL and type 2 MI.

BackgroundA few papers have considered reproducibility of a posteriori dietary patterns across populations, as well as pattern associations with head and neck cancer risk when multiple populations are available.MethodsWe used individual-level pooled data from seven case-control studies (3844 cases; 6824 controls) participating in the International Head and Neck Cancer Epidemiology consortium. We simultaneously derived shared and study-specific a posteriori patterns with a novel approach called multi-study factor analysis applied to 23 nutrients. We derived odds ratios (ORs) and 95% confidence intervals (CIs) for cancers of the oral cavity and pharynx combined, and larynx, from logistic regression models.ResultsWe identified three shared patterns that were reproducible across studies (75% variance explained): the Antioxidant vitamins and fiber (OR = 0.57, 95% CI = 0.41, 0.78, highest versus lowest score quintile) and the Fats (OR = 0.80, 95% CI = 0.67, 0.95) patterns were inversely associated with oral and pharyngeal cancer risk. The Animal products and cereals (OR = 1.5, 95% CI = 1.1, 2.1) and the Fats (OR = 1.8, 95% CI = 1.4, 2.3) patterns were positively associated with laryngeal cancer risk, whereas a linear inverse trend in laryngeal cancer risk was evident for the Antioxidant vitamins and fiber pattern. We also identified four additional study-specific patterns, one for each of the four US studies examined. We named them all as Dairy products and breakfast cereals, and two were associated with oral and pharyngeal cancer risk.ConclusionMulti-study factor analysis provides insight into pattern reproducibility and supports previous evidence on cross-country reproducibility of dietary patterns and on their association with head and neck cancer risk. See video abstract at, http://links.lww.com/EDE/B430.

BackgroundThere is widespread concern about the use of body mass index (BMI) to define obesity status in postmenopausal women because it may not accurately represent an individual's true obesity status. The objective of the present study is to examine and adjust for exposure misclassification bias from using an indirect measure of obesity (BMI) compared with a direct measure of obesity (percent body fat).MethodsWe used data from postmenopausal non-Hispanic black and non-Hispanic white women in the Women's Health Initiative (n=126,459). Within the Women's Health Initiative, a sample of 11,018 women were invited to participate in a sub-study involving dual-energy x-ray absorptiometry scans. We examined indices of validity comparing BMI-defined obesity (≥30 kg/m), with obesity defined by percent body fat. We then used probabilistic bias analysis models stratified by age and race to explore the effect of exposure misclassification on the obesity-mortality relationship.ResultsValidation analyses highlight that using a BMI cutpoint of 30 kg/m to define obesity in postmenopausal women is associated with poor validity. There were notable differences in sensitivity by age and race. Results from the stratified bias analysis demonstrated that failing to adjust for exposure misclassification bias results in attenuated estimates of the obesity-mortality relationship. For example, in non-Hispanic white women 50-59 years of age, the conventional risk difference was 0.017 (95% confidence interval = 0.01, 0.023) and the bias-adjusted risk difference was 0.035 (95% simulation interval = 0.028, 0.043).ConclusionsThese results demonstrate the importance of using quantitative bias analysis techniques to account for nondifferential exposure misclassification of BMI-defined obesity. See video abstract at, http://links.lww.com/EDE/B385.

BackgroundSelf-harm is a leading cause of morbidity and mortality. Exposure to community violence is an important and potentially modifiable feature of the social environment that may affect self-harm, but studies to date are limited in the samples and outcomes examined.MethodsWe conducted a population-based, nested case-control study. Cases were all deaths and hospital visits due to self-harm in California, 2006-2013. We frequency-matched California resident population-based controls from the American Community Survey to cases on age, gender, race/ethnicity, and year of survey/injury. We assessed past-year community violence using deaths and hospital visits due to interpersonal violence in the community of residence. We estimated risk difference parameters that were defined to avoid extrapolation and to capture associations between changes in the distribution of community violence and the population-level risk of self-harm.ResultsAfter adjustment for confounders, setting past-year community violence to the lowest monthly levels observed within each community over the study period was associated with a 30.1 (95% confidence interval = 29.6, 30.5) per 100,000 persons per year lower risk of nonfatal self-harm but no difference in the risk of fatal self-harm. Associations for a parameter corresponding to a hypothetical violence prevention intervention targeting high-violence communities indicated a 5% decrease in nonfatal self-harm at the population level. In sensitivity analyses, results were robust.ConclusionsThis study strengthens evidence on the relationship between community violence and self-harm. Future research should investigate reasons for differential associations by age and gender and whether community violence prevention programs have meaningful impacts on self-harm.

BackgroundIn middle age, stroke incidence is higher among black than white Americans. For unknown reasons, this inequality decreases and reverses with age. We conducted simulations to evaluate whether selective survival could account for observed age patterning of black-white stroke inequalities.MethodsWe simulated birth cohorts of 20,000 blacks and 20,000 whites with survival distributions based on US life tables for the 1919-1921 birth cohort. We generated stroke incidence rates for ages 45-94 years using Reasons for Geographic and Racial Disparities in Stroke (REGARDS) study rates for whites and setting the effect of black race on stroke to incidence rate difference (IRD) = 20/10,000 person-years at all ages, the inequality observed at younger ages in REGARDS. We compared observed age-specific stroke incidence across scenarios, varying effects of U, representing unobserved factors influencing mortality and stroke risk.ResultsDespite a constant adverse effect of black race on stroke risk, the observed black-white inequality in stroke incidence attenuated at older age. When the hazard ratio for U on stroke was 1.5 for both blacks and whites, but U only directly influenced mortality for blacks (hazard ratio for U on mortality =1.5 for blacks; 1.0 for whites), stroke incidence rates in late life were lower among blacks (average observed IRD = -43/10,000 person-years at ages 85-94 years versus causal IRD = 20/10,000 person-years) and mirrored patterns observed in REGARDS.ConclusionsA relatively moderate unmeasured common cause of stroke and survival could fully account for observed age attenuation of racial inequalities in stroke.

BackgroundEvidence suggests that aspects of the neighborhood environment may influence risk of problematic drug use among adolescents. Our objective was to examine mediating roles of aspects of the school and peer environments on the effect of receiving a Section 8 housing voucher and using it to move out of public housing on adolescent substance use outcomes.MethodsWe used data from the Moving to Opportunity (MTO) experiment that randomized receipt of a Section 8 housing voucher. Hypothesized mediators included school climate, safety, peer drug use, and participation in an after-school sport or club. We applied a doubly robust, semiparametric estimator to longitudinal MTO data to estimate stochastic direct and indirect effects of randomization on cigarette use, marijuana use, and problematic drug use. Stochastic direct and indirect effects differ from natural direct and indirect effects in that they do not require assuming no posttreatment confounder of the mediator-outcome relationship. Such an assumption would be at odds with any causal model that reflects an intervention affecting a mediator and outcome through adherence to treatment assignment.ResultsHaving friends who use drugs and involvement in after-school sports or clubs partially mediated the effect of housing voucher receipt on adolescent substance use (e.g., stochastic indirect effect 0.45% [95% confidence interval: 0.12%, 0.79%] for having friends who use drugs and 0.04% [95% confidence interval: -0.02%, 0.10%] for involvement in after-school sports or clubs mediating the relationship between housing voucher receipt and marijuana use among boys). However, these mediating effects were small, contributing only fractions of a percent to the effect of voucher receipt on probability of substance use. No school environment variables were mediators.ConclusionsMeasured school- and peer-environment variables played little role in mediating the effect of housing voucher receipt on subsequent adolescent substance use.

BackgroundResearchers have suggested that social media and online search data might be used to monitor and predict syphilis and other sexually transmitted diseases. Because people at risk for syphilis might seek sexual health and risk-related information on the internet, we investigated associations between internet state-level search query data (e.g., Google Trends) and reported weekly syphilis cases.MethodsWe obtained weekly counts of reported primary and secondary syphilis for 50 states from 2012 to 2014 from the US Centers for Disease Control and Prevention. We collected weekly internet search query data regarding 25 risk-related keywords from 2012 to 2014 for 50 states using Google Trends. We joined 155 weeks of Google Trends data with 1-week lag to weekly syphilis data for a total of 7750 data points. Using the least absolute shrinkage and selection operator, we trained three linear mixed models on the first 10 weeks of each year. We validated models for 2012 and 2014 for the following 52 weeks and the 2014 model for the following 42 weeks.ResultsThe models, consisting of different sets of keyword predictors for each year, accurately predicted 144 weeks of primary and secondary syphilis counts for each state, with an overall average R of 0.9 and overall average root mean squared error of 4.9.ConclusionsWe used Google Trends search data from the prior week to predict cases of syphilis in the following weeks for each state. Further research could explore how search data could be integrated into public health monitoring systems.

Occupational exposure guidelines are ideally based on estimated effects of static interventions that assign constant exposure over a working lifetime. Static effects are difficult to estimate when follow-up extends beyond employment because their identifiability requires additional assumptions. Effects of dynamic interventions that assign exposure while at work, allowing subjects to leave and become unexposed thereafter, are more easily identifiable but result in different estimates. Given the practical implications of exposure limits, we explored the drivers of the differences between static and dynamic interventions in a simulation study where workers could terminate employment because of an intermediate adverse health event that functions as a time-varying confounder. The two effect estimates became more similar with increasing strength of the health event and outcome relationship and with increasing time between health event and employment termination. Estimates were most dissimilar when the intermediate health event occurred early in employment, providing an effective screening mechanism.

BackgroundPneumococcal conjugate vaccines (PCVs) prevent invasive pneumococcal disease and pneumonia. However, some low-and middle-income countries have yet to introduce PCV into their immunization programs due, in part, to lack of certainty about the potential impact. Assessing PCV benefits is challenging because specific data on pneumococcal disease are often lacking, and it can be difficult to separate the effects of factors other than the vaccine that could also affect pneumococcal disease rates.MethodsWe assess PCV impact by combining Bayesian model averaging with change-point models to estimate the timing and magnitude of vaccine-associated changes, while controlling for seasonality and other covariates. We applied our approach to monthly time series of age-stratified hospitalizations related to pneumococcal infection in children younger 5 years of age in the United States, Brazil, and Chile.ResultsOur method accurately detected changes in data in which we knew true and noteworthy changes occurred, i.e., in simulated data and for invasive pneumococcal disease. Moreover, 24 months after the vaccine introduction, we detected reductions of 14%, 9%, and 9% in the United States, Brazil, and Chile, respectively, in all-cause pneumonia (ACP) hospitalizations for age group 0 to

BackgroundColiphages have been proposed as indicators of fecal contamination in recreational waters because they better mimic the persistence of pathogenic viruses in the environment and wastewater treatment than fecal indicator bacteria. We estimated the association between coliphages and gastrointestinal illness and compared it with the association with culturable enterococci.MethodsWe pooled data from six prospective cohort studies that enrolled coastal beachgoers in California, Alabama, and Rhode Island. Water samples were collected and gastrointestinal illness within 10 days of the beach visit was recorded. Samples were tested for enterococci and male-specific and somatic coliphages. We estimated cumulative incidence ratios (CIR) for the association between swimming in water with detectable coliphage and gastrointestinal illness when human fecal pollution was likely present, not likely present, and under all conditions combined. The reference group was unexposed swimmers. We defined continuous and threshold-based exposures (coliphage present/absent, enterococci >35 vs. ≤35 CFU/100 ml).ResultsUnder all conditions combined, there was no association between gastrointestinal illness and swimming in water with detectable coliphage or enterococci. When human fecal pollution was likely present, coliphage and enterococci were associated with increased gastrointestinal illness, and there was an association between male-specific coliphage level and illness that was somewhat stronger than the association between enterococci and illness. There were no substantial differences between male-specific and somatic coliphage.ConclusionsSomatic coliphage and enterococci had similar associations with gastrointestinal illness; there was some evidence that male-specific coliphage had a stronger association with illness than enterococci in marine waters with human fecal contamination.

BackgroundTobacco smoke contains known hormonally active chemicals and reproductive toxicants. Several studies have examined prenatal maternal smoking and offspring age at menarche, but few examined earlier pubertal markers, nor accounted for exposure during childhood. Our objective was to examine pre- and postnatal smoke exposure in relation to timing of early pubertal events.MethodsAn ethnically diverse cohort of 1239 girls was enrolled at age 6-8 years old for a longitudinal study of puberty at three US sites. Girls participated in annual or semi-annual exams to measure anthropometry and Tanner breast and pubic hair stages. Prenatal and current tobacco smoke exposures, as well as covariates, were obtained from parent questionnaire. Cotinine was measured in urine collected at enrollment. Using accelerated failure time models, we calculated adjusted time ratios for age at pubertal onset (maturation stages 2 or higher) and smoke exposure.ResultsGirls with higher prenatal (≥5 cigarettes per day) or secondhand smoke exposure had earlier pubic hair development than unexposed (adjusted time ratio: 0.92 [95% CI = 0.87, 0.97] and 0.94 [95% CI = 0.90, 0.97], respectively). Including both exposures in the same model yielded similar associations. Higher urinary cotinine quartiles were associated with younger age at breast and pubic hair onset in unadjusted models, but not after adjustment.ConclusionsGreater prenatal and childhood secondhand smoke exposure were associated with earlier onset of pubic hair, but not breast, development. These exposures represent modifiable risk factors for early pubertal development that should be considered for addition to the extensive list of adverse effects from tobacco smoke.

Background: Intersectionality theoretical frameworks have been increasingly incorporated into quantitative research. A range of methods have been applied to describing outcomes and disparities across large numbers of intersections of social identities or positions, with limited evaluation.Methods: Using data simulated to reflect plausible epidemiologic data scenarios, we evaluated methods for intercategorical intersectional analysis of continuous outcomes, including cross-classification, regression with interactions, multilevel analysis of individual heterogeneity (MAIHDA), and decision-tree methods (classification and regression trees [CART], conditional inference trees [CTree], random forest). The primary outcome was estimation accuracy of intersection-specific means. We applied each method to an illustrative example using National Health and Nutrition Examination Study (NHANES) systolic blood pressure data.Results: When studying high-dimensional intersections at smaller sample sizes, MAIHDA, CTree, and random forest produced more accurate estimates. In large samples, all methods performed similarly except CART, which produced less accurate estimates. For variable selection, CART performed poorly across sample sizes, although random forest performed best. The NHANES example demonstrated that different methods resulted in meaningful differences in systolic blood pressure estimates, highlighting the importance of selecting appropriate methods.Conclusions: This study evaluates some of a growing toolbox of methods for describing intersectional health outcomes and disparities. We identified more accurate methods for estimating outcomes for high-dimensional intersections across different sample sizes. As estimation is rarely the only objective for epidemiologists, we highlight different outputs each method creates, and suggest the sequential pairing of methods as a strategy for overcoming certain technical challenges. [ABSTRACT FROM AUTHOR]