OBJECTIVES: Thrombophilia is associated with an elevated risk of thrombosis, which may be further elevated with use of hormonal contraception. Our objective was to update a previously published systematic review on thrombosis risk with use of hormonal contraception among women with thrombophilia. STUDY DESIGN: We conducted a systematic review of five databases from database inception through December 8, 2022. We searched for articles that examined risk of venous thromboembolism (VTE) or arterial thromboembolism (ATE) in women with thrombophilia using hormonal contraception compared with women using non-hormonal or no contraception. We assessed risk of bias for each study and certainty of evidence for all outcomes. RESULTS: Eighteen articles met inclusion criteria; four had moderate risk of bias and 14 had high risk of bias. Odds of VTE in women with factor V Leiden (FVL) mutation or prothrombin (PT) gene mutation were elevated in combined oral contraception (COC) users vs non-users. Odds of VTE were elevated in COC users with FVL mutation, PT gene mutation, both FVL and PT mutations, antithrombin (AT) deficiency, or protein C deficiency compared with non-users without the mutation. Odds of stroke were elevated in COC users with FVL mutation compared with non-users without the mutation. Evidence was mixed on whether risk was elevated in women with protein S deficiency using COC compared with non-use. One study found elevated odds of VTE in women with FVL mutation but not women with PT gene mutation using progestin-only contraception (POC), compared with non-users without the mutation. CONCLUSIONS: Overall, studies found elevated odds of VTE and ATE in women with thrombophilia using COC compared with non-users without thrombophilia. The certainty of evidence for all outcomes is low. Evidence is also limited by small numbers of women and minimal evidence on use of patch, ring, or POC, and is insufficient to assess differential risk by all thrombophilia types. IMPLICATIONS: Use of estrogen-containing hormonal contraception might further elevate risk of thrombosis among women with thrombophilia. Further study is needed on safety of POC use in women with thrombophilia.

Background: Performance of thyroid function assays can vary significantly. To address this issue, the Centers for Disease Control and Prevention (CDC) Clinical Standardization Programs conducted an interlaboratory comparison of free thyroxine (fT4) immunoassays (IAs) and laboratory-developed tests (LDTs). This assessment aimed to determine the current performance characteristics of these assays as a first step toward measurement standardization. Thyrotropin (TSH) IAs were also evaluated. Methods: Assays measured 41 blinded individual-donor sera, including a sample from a pregnant woman (for fT4 analysis only) and three serum pools, with 11.3-32.1 pmol/L (0.881-2.49 ng/dL) fT4 and 0.337-21.6 mIU/L TSH in duplicate over 2 days. Passing-Bablok regression analysis performed pre-recalibration compared assays performance to the CDC fT4 reference measurement procedure (RMP) or TSH all-lab mean (ALM). Additionally, the impact of linear regression-based recalibration of assays to the CDC fT4 RMP or TSH ALM was estimated. Inter-assay agreement of sample classification according to the assay-specific reference interval (RI) was assessed pre- and post-recalibration. Results: A total of 21 fT4 and 17 TSH assays participated. Pre-recalibration, median biases of TSH measurements to the ALM were -1.2% [confidence interval or CI -1.8% to -0.4%], and good classification agreement among TSH assays was observed. fT4 assays all showed a negative median bias to the RMP, with higher bias among IAs (median: -20.3%, CI [-21.5% to -19.4%]) than LDTs (median: -4.5%, [CI -6.1% to -3.2%]). Of the individual-donor sera, only 21 out of 40 samples were classified uniformly by all fT4 assays, indicating poor inter-assay agreement. Post-recalibration, agreement improved to 33 out of 40 individual-donor sera correctly classified by all tested IAs and LDTs. Similar improvement in post-recalibration median percent bias was observed for fT4 IAs (median: -0.2, [CI -1.2% to 0.6%]) and LDTs (median: -0.3%, [CI -2.5% to 1.4%]). Conclusions: The comparison among fT4 assays emphasizes the need for measurement standardization to improve accuracy and comparability. This and previous studies demonstrate the possibility to develop common fT4 RIs via standardization, enabling the use of evidence-based clinical guidelines universally in patient care. Recalibration can effectively address high variability in fT4 assays, ensuring consistent diagnostic classification.

Introduction: Diabetes is the seventh leading cause of death in the United States, impacting over 37 million people. Accurate glucose measurements are critical for effective diabetes management. A reliable candidate reference measurement procedure (cRMP) for assessing the analytical performance of glucose tests performed in patient care is essential for ensuring measurement accuracy. Methods: We have developed a gas chromatography-mass spectrometry (GC–MS)-based cRMP for glucose in human serum. In this procedure, glucose is measured as the aldononitrile acetate derivative and quantitated using a 13C6-glucose internal standard. Results: Analytical selectivity was achieved through chromatographic separation and monitoring the quantitation ion/confirmation ion ratios in samples. With bias ranging from −0.79 % to 0.67 % for eight levels of serum-based certified reference materials from the National Institute of Standards and Technology (NIST) and Laboratoire national de métrologie et d'essais (LNE) and total CVs of 1.11 %, 0.68 % and 0.74 % at the low, medium, and high glucose concentration levels, respectively, the cRMP provided excellent accuracy and precision. The calibration curve was linear throughout the 13.51–378.21 mg/dL [0.75–21 mmol/L] measurement range (R2 = 0.9999), with a mean slope of 270.73 (95 % CI, 270.19 to 271.27) and an intercept of 0.021 (95 % CI, −0.157 to 0.199). The limit of detection was 0.25 mg/dL (0.014 mmol/L) and the limit of quantitation was 0.83 mg/dL (0.046 mmol/L). Conclusion: The described GC–MS method, with metrological traceability to the International System of Units (SI), provides highly accurate and precise measurements of glucose in human serum. © 2025

OBJECTIVES: To systematically review literature on whether hormonal contraception following ulipristal acetate (UPA) for emergency contraception decreases the effectiveness of either drug. STUDY DESIGN: We searched multiple databases through December 2022 for studies assessing the interaction between UPA and hormonal contraception. The primary outcome was contraceptive effectiveness, measured by pregnancy rates or proxy measures (e.g., ovarian activity). We extracted and summarized findings from identified studies, assessed risk of bias for each study, and determined certainty of evidence for all outcomes. RESULTS: Four studies met inclusion criteria; all had low risk of bias. Two studies assessed whether UPA use affected the ability of oral contraceptives (OCs) to inhibit ovulation; no differences were observed in ovarian activity when starting OCs one day after UPA compared with starting OCs one day after placebo. Two studies assessed whether OC use affected the ability of UPA to delay ovulation; both studies observed higher proportions of ovulation when UPA was followed by OC use versus delayed or no OC use. One study assessed ovulation risk when UPA was taken after missed OCs, followed by immediate versus delayed OC resumption; no ovulations occurred within the first five days after UPA administration in either group, but there was greater risk of ovulation beyond five days with delayed versus immediate OC resumption. CONCLUSIONS: While there is no evidence that UPA affects the ability of hormonal contraception to inhibit ovulation, hormonal contraception use immediately or soon after UPA may decrease UPA's ability to delay ovulation. The certainty of evidence ranged from moderate to very low. IMPLICATIONS: If a patient desires hormonal contraception after UPA, an interaction with UPA can be avoided by delaying initiation or resumption of hormonal contraception. The risk of pregnancy due to decreased effectiveness of UPA with immediate hormonal contraception should be balanced against the risk of subsequent pregnancy due to delay or non-start of hormonal contraception.

OBJECTIVES: To explore how artificial intelligence (AI) can enhance infodemiology, which distributes and scans information in the electronic medium, to process social media posts for HIV pre-exposure prophylaxis (PrEP). DESIGN: Systematic Review. METHODS: We searched in the U.S. Centers for Disease Control and Prevention's Prevention Research Synthesis database through June 2024 (PROSPERO: CRD42023458870). We included infodemiology studies published in English and reported using AI to process social media posts on PrEP. Two reviewers independently screened citations, extracted data, and conducted a risk of bias assessment using the Joanna Briggs Institute Critical Appraisal Checklist for Prevalence Studies. Findings are narratively summarized. RESULTS: Of the 135 citations screened, eight infodemiology studies were identified, analyzing over 58.9 million posts. Infodemiology studies found the PrEP topics commonly discussed in communities (e.g., barriers of uptake), rumors that may raise public health concerns (e.g., PrEP is a prevention method against COVID-19 infection), geographic locations where concerns regarding risk of acquiring HIV were raised (e.g., most HIV-related posts were from the 10 states with the highest numbers of new HIV diagnoses), and predicted HIV trends (e.g., HIV-related tweets were negatively correlated with the county-level HIV incidence rate in the following year). CONCLUSIONS: Despite the limitations of this review including a small number of studies reviewed, our review suggests social media posts may provide information on real-time PrEP-related concerns, and AI can accelerate and enhance the processing of mass data to identify the information that communities need and the areas/locations that may need HIV prevention intervention.

BACKGROUND: Medicare claims are frequently used to study Clostridioides difficile infection (CDI) epidemiology. However, they lack specimen collection and diagnosis dates to assign location of onset. Algorithms to classify CDI onset location using claims data have been published, but the degree of misclassification is unknown. METHODS: We linked patients with laboratory-confirmed CDI reported to four Emerging Infections Program (EIP) sites from 2016-2021 to Medicare beneficiaries with fee-for-service Part A/B coverage. We calculated sensitivity of ICD-10-CM codes in claims within ±28 days of EIP specimen collection. CDI was categorized as hospital, long-term care facility, or community-onset using three different Medicare claims-based algorithms based on claim type, ICD-10-CM code position, duration of hospitalization, and ICD-10-CM diagnosis code presence-on-admission indicators. We assessed concordance of EIP case classifications, based on chart review and specimen collection date, with claims case classifications using Cohen's kappa statistic. RESULTS: Of 12,671 CDI cases eligible for linkage, 9,032 (71%) were linked to a single, unique Medicare beneficiary. Compared to EIP, sensitivity of CDI ICD-10-CM codes was 81%; codes were more likely to be present for hospitalized patients (93.0%) than those who were not (56.2%). Concordance between EIP and Medicare claims algorithms ranged from 68% to 75%, depending on the algorithm used (κ = 0.56-0.66). CONCLUSION: ICD-10-CM codes in Medicare claims data had high sensitivity compared to laboratory-confirmed CDI reported to EIP. Claims-based epidemiologic classification algorithms had moderate concordance with EIP classification of onset location. Misclassification of CDI onset location using Medicare algorithms may bias findings of claims-based CDI studies.

INTRODUCTION: Non-survey-based data sources (e.g. electronic health records, administrative claims) have been used to estimate vaccination coverage among US adults. However, these data sources were not collected for research or surveillance purposes and may have substantial limitations. The objectives of this narrative review were to: 1) identify published studies that used non-survey-based data sources to estimate adult vaccination coverage for one or more routinely recommended vaccines; and 2) summarize the strengths and limitations of these data sources for coverage assessments. AREAS COVERED: Thirty-four publications derived from 9 data sources were reviewed: 16 publications were in a general population (i.e. defined by age), 12 were among pregnant women, and 6 were among individuals with chronic health conditions. While several data sources used continuous health insurance enrollment to define the study population, doing so limited generalizability to stably insured populations. Methods for obtaining race and ethnicity data were complex and potentially subject to bias. None of the reviewed studies presented any formal assessment of vaccine data validity. EXPERT OPINION: While multiple non-survey-based data sources have been used to assess adult vaccination coverage in the United States, important limitations exist, including related to generalizability, data validity, and risk of bias.

Objectives This report describes the development and operations of the 2016 National Culturally and Linguistically Appropriate Services Survey for Office-based Physicians (National CLAS Physician Survey). The survey was developed to understand awareness, adoption, and implementation of the National CLAS Standards in health and health care among office-based physicians. Methods Survey development included a literature review of survey and assessment instruments that evaluated cultural and linguistic appropriateness in health care. Survey questions were pretested during a cognitive interview study of 20 office-based physicians in the District of Columbia metropolitan area. The cognitive interviews were analyzed using a grounded theory approach. The final survey was administered via web, mail, and computer-assisted telephone interview to 2, 400 sampled physicians between August 2016 and December 2016. A nonresponse bias assessment was conducted. Results The literature review identified five survey and assessment instruments. Collectively, survey content included: cultural competency training, cultural awareness, and adoption of the National CLAS Standards. Cognitive interviews showed respondent difficulty in question interpretation and survey completion of some items. Survey revisions addressed these issues. The final overall weighted survey response rate was 33.8%. Final weights produced a lower standardized bias than base weights. Conclusions The National CLAS Physician Survey is the first nationally representative survey to describe the use and implementation of culturally and linguistically appropriate services by office-based physicians. Data can serve as a baseline for future studies and as a benchmark for meeting the key objectives of the National CLAS Standards. © 2025, null. All rights reserved.

BACKGROUND: Parent-reported surveys are commonly used in child health research. However, few national surveys have examined concordance between parent- and teen-reported Adverse Childhood Experiences (ACEs). OBJECTIVE: To examine concordance between parent- and teen-reported ACEs among a nationally representative sample of teens and understand sociodemographic correlates of concordance. PARTICIPANTS AND SETTING: Data were collected as part of the National Health Interview Survey (NHIS), an annual nationally representative survey of the United States, with parent-reported interviews and its follow-back survey, the National Health Interview Survey-Teen (NHIS-Teen), a self-administered web survey of youth aged 12-17 years. METHODS: Parent- and teen-reported ACEs, and several measures of concordance (e.g. weighted Cohen's kappa, prevalence adjusted bias adjusted kappa (PABAK)) are presented. Unadjusted logistic regressions tested associations between sociodemographic characteristics and likelihood of concordance for each ACE. RESULTS: ACEs prevalence based on parent report were lower compared to teen report (e.g. victim of or witnessed violence in the neighborhood, 6.5 % parent-reported vs. 15.5 %, teen-reported). Weighted Cohen's kappa statistics showed fair to moderate agreement (ks ranging from 0.27 to 0.53), with PABAK statistics showing slightly higher levels (ks ranging from 0.41 to 0.88). CONCLUSIONS: There is discordance in ACEs reporting between teens and their parents, with parents less likely to report that their teen experience an ACE than teens are. This emphasizes the importance of collecting information directly from teenagers, which may help inform intervention work.

BACKGROUND: Real-world COVID-19 vaccine effectiveness (VE) studies are investigating exposures of increasing complexity accounting for time since vaccination. These studies require methods that adjust for the confounding that arises when morbidities and demographics are associated with vaccination and the risk of outcome events. Methods based on propensity scores (PS) are well-suited to this when the exposure is dichotomous, but present challenges when the exposure is multinomial. OBJECTIVE: This simulation study aimed to investigate alternative methods to adjust for confounding in VE studies that have a test-negative design. METHODS: Adjustment for a disease risk score (DRS) is compared with multivariable logistic regression. Both stratification on the DRS and direct covariate adjustment of the DRS are examined. Multivariable logistic regression with all the covariates and with a limited subset of key covariates is considered. The performance of VE estimators is evaluated across a multinomial vaccination exposure in simulated datasets. RESULTS: Bias in VE estimates from multivariable models ranged from -5.3% to 6.1% across 4 levels of vaccination. Standard errors of VE estimates were unbiased, and 95% coverage probabilities were attained in most scenarios. The lowest coverage in the multivariable scenarios was 93.7% (95% CI 92.2%-95.2%) and occurred in the multivariable model with key covariates, while the highest coverage in the multivariable scenarios was 95.3% (95% CI 94.0%-96.6%) and occurred in the multivariable model with all covariates. Bias in VE estimates from DRS-adjusted models was low, ranging from -2.2% to 4.2%. However, the DRS-adjusted models underestimated the standard errors of VE estimates, with coverage sometimes below the 95% level. The lowest coverage in the DRS scenarios was 87.8% (95% CI 85.8%-89.8%) and occurred in the direct adjustment for the DRS model. The highest coverage in the DRS scenarios was 94.8% (95% CI 93.4%-96.2%) and occurred in the model that stratified on DRS. Although variation in the performance of VE estimates occurred across modeling strategies, variation in performance was also present across exposure groups. CONCLUSIONS: Overall, models using a DRS to adjust for confounding performed adequately but not as well as the multivariable models that adjusted for covariates individually.

Data science is an emerging field that provides new analytical methods. It incorporates novel data sources (eg, internet data) and methods (eg, machine learning) that offer valuable and timely insights into public health issues, including injury and violence prevention. The objective of this research was to describe ethical considerations for public health data scientists conducting injury and violence prevention-related data science projects to prevent unintended ethical, legal, and social consequences, such as loss of privacy or loss of public trust. We first reviewed foundational bioethics and public health ethics literature to identify key ethical concepts relevant to public health data science. After identifying these ethics concepts, we held a series of discussions to organize them under broad ethical domains. Within each domain, we examined relevant ethics concepts from our review of the primary literature. Lastly, we developed questions for each ethical domain to facilitate the early conceptualization stage of the ethical analysis of injury and violence prevention projects. We identified 4 ethical domains: privacy, responsible stewardship, justice as fairness, and inclusivity and engagement. We determined that each domain carries equal weight, with no consideration bearing more importance than the others. Examples of ethical considerations are clearly identifying project goals, determining whether people included in projects are at risk of reidentification through external sources or linkages, and evaluating and minimizing the potential for bias in data sources used. As data science methodologies are incorporated into public health research to work toward reducing the effect of injury and violence on individuals, families, and communities in the United States, we recommend that relevant ethical issues be identified, considered, and addressed.

Annual surveys of refugees in Gambella, Ethiopia suggest that anemia is a persistent public health problem among non-pregnant women of reproductive age (NP-WRA, 15-49 years). Measurement of anemia in most refugee camp settings is conducted using an invasive HemoCue 301. We assessed the accuracy and precision of a non-invasive, pulse CO-oximeter in measuring anemia among NP-WRA in four Gambella refugee camps. We conducted a population-representative household survey between November 7 and December 4, 2022. Hemoglobin (Hb) concentration was measured by HemoCue 301, using capillary blood, and Rad-67, a novel non-invasive device. We collected four measurements per participant: two per device. We calculated Rad-67 bias and precision of Hb measurements and sensitivity and specificity of detecting anemia. Of the 812 NP-WRAs selected, 807 (99%) participated in the study. Anemia was detected in 39% of NP-WRA as classified by the Rad-67 compared with 47% of NP-WRA as classified by the HemoCue 301. Average bias of Rad-67 measurements was 1.1 ± 1.0 SD g/dL, using HemoCue 301 as a comparator. Absolute mean difference between the first and second measurements was 0.9 g/dL (95% CI 0.8, 0.9) using the Rad-67, compared with 0.6 g/dL (95% CI 0.5, 0.6) using the HemoCue 301. The Rad-67 had 49% sensitivity and 70% specificity for detecting anemia, compared with the HemoCue 301. The Rad-67 can be a useful tool for anemia screening; however, lower accuracy and precision, and poor sensitivity suggest it cannot immediately replace the HemoCue 301 in the study area.

Commutability is where the measurement response for a reference material (RM) is the same as for an individual patient sample with the same concentration of analyte measured using two or more measurement systems. Assessment of commutability is essential when the RM is used in a calibration hierarchy or to ensure that clinical measurements are comparable across different measurement procedures and at different times. The commutability of three new Standard Reference Materials(®) (SRMs) for determining serum total 25-hydroxyvitamin D [25(OH)D], defined as the sum of 25-hydroxyvitamin D(2) [25(OH)D(2)] and 25-hydroxyvitamin D(3) [25(OH)D(3)], was assessed through an interlaboratory study. The following SRMs were assessed: (1) SRM 2969 Vitamin D Metabolites in Frozen Human Serum (Total 25-Hydroxyvitamin D Low Level), (2) SRM 2970 Vitamin D Metabolites in Frozen Human Serum (25-Hydroxyvitamin D(2) High Level), and (3) SRM 1949 Frozen Human Prenatal Serum. These SRMs represent three clinically relevant situations including (1) low levels of total 25(OH)D, (2) high level of 25(OH)D(2), and (3) 25(OH)D levels in nonpregnant women and women during each of the three trimesters of pregnancy with changing concentrations of vitamin D-binding protein (VDBP). Twelve laboratories using 17 different ligand binding assays and eight laboratories using nine commercial and custom liquid chromatography-tandem mass spectrometry (LC-MS/MS) assays provided results in this study. Commutability of the SRMs with patient samples was assessed using the Clinical and Laboratory Standards Institute (CLSI) approach based on 95% prediction intervals or a pre-set commutability criterion and the recently introduced International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) approach based on differences in bias for the clinical and reference material samples using a commutability criterion of 8.8%. All three SRMs were deemed as commutable with all LC-MS/MS assays using both CLSI and IFCC approaches. SRM 2969 and SRM 2970 were deemed noncommutable for three and seven different ligand binding assays, respectively, when using the IFCC approach. Except for two assays, one or more of the three pregnancy levels of SRM 1949 were deemed noncommutable or inconclusive using different ligand binding assays and the commutability criterion of 8.8%. Overall, a noncommutable assessment for ligand binding assays is determined for these SRMs primarily due to a lack of assay selectivity related to 25(OH)D(2) or an increasing VDBP in pregnancy trimester materials rather than the quality of the SRMs. With results from 17 different ligand binding and nine LC-MS/MS assays, this study provides valuable knowledge for clinical laboratories to inform SRM selection when assessing 25(OH)D status in patient populations, particularly in subpopulations with low levels of 25(OH)D, high levels of 25(OH)D(2), women only, or women who are pregnant.

Rotavirus vaccine appears to perform sub-optimally in countries with higher rotavirus burden. We hypothesized that differences in the magnitude of rotavirus exposures may bias vaccine efficacy (VE) estimates, so true differences in country-specific rotavirus VE would be exaggerated without accommodating differences in exposure. We estimated VE against any-severity and severe rotavirus gastroenteritis (RVGE) using Poisson regression models fit to pooled individual-level data from Phase II and III monovalent rotavirus vaccine trials conducted between 2000 and 2012. The standard approach model included terms for vaccination, country, and a vaccination-country interaction. Other models used proxies for exposure magnitude like severe RVGE rate or age at severe RVGE instead of country. Country-specific proxies were calculated from placebo group data or extracted from an external meta-analysis. Analyses included 83,592 infants from 23 countries in the Americas, Europe, Africa, and Asia. Using the standard approach, VE against severe RVGE substantially varied (10-100%). Using the severe RVGE rate proxy brought VE from all but two countries between 80% and 86%. Heterogeneity for VE against any-severity RVGE was similarly attenuated. Adjusting for exposure proxies reduced heterogeneity in country-specific rotavirus VE estimates. This phenomenon may extend to other vaccines against partially immunizing pathogens with global disparities in burden.

INTRODUCTION: Occupational injuries have been associated with increased suicide mortality, but prior studies have not accounted for pre-injury depression. METHODS: We linked injuries that occurred from 1994 to 2000 in the Washington State workers' compensation system with Social Security Administration data on earnings and mortality through 2018. We estimated the subdistribution hazard ratio (sHR) and 95% confidence interval using competing risks regression of suicide deaths with lost time compared with medical-only injuries separately for men and women, adjusting for age, pre-injury annual earnings, and industry. We further adjusted for pre-injury diagnosis of major depressive disorder by using a quantitative bias analysis (QBA), with the prevalence of this disorder in workers derived from an external health insurance claims data set. RESULTS: Elevated suicide mortality was observed following lost-time injuries compared with medical-only injuries for men (sHR = 1.49, 95% CI [1.14, 1.93]) and women (sHR = 1.30, 95% CI [1.00, 1.69]), adjusting for age, pre-injury earnings, and industry. Adjusted for pre-injury depression using a QBA, elevated suicide risk in men remained statistically significant (median sHR = 1.33, simulation interval [1.18, 1.47]) but not for women. DISCUSSION: Workplace injury requiring time off work appeared to remain influential in increasing suicide risk among men, even after controlling for pre-injury depression. The relationship between mental health before and after occupational injury is complex and studies should better integrate mental health pre-injury. CONCLUSIONS: Though many questions remain on the complex relationship between work, depression, injuries, and suicide, employers should work to prevent injuries and consider implementing mental health programs, which could be helpful in reducing suicide risk.

Respirable dust mass is a prevalent occupational health hazard to the mining workforce. Mineral matrices observed in the mine environment are complex, time varying, and heterogeneous. This poses a challenge to assessing dust exposure using Fourier transform infrared (FT-IR) spectrometry as calibrations for constituent dust species (e.g., crystalline silica) have historically been trained using homogeneous standards or simple mixtures therein. Investigations have considered direct-on-filter analysis, which collects FT-IR spectra directly from sampling filters for calibration, as an alternative. Direct-on-filter analysis using a partial least squares (PLS) method has gained particular interest recently due to the potential to rapidly quantify multiple species from a single filter at the mine site. By design, heterogeneity, and its presumed impact on method accuracy, cannot be addressed in the laboratory when using a direct-on-filter approach motivating the need for more advanced calibration approaches. When heterogeneity is present, mixture of experts (MoE) finite mixture models offer a promising and novel alternative to PLS direct-on-filter analysis as MoE incorporates cluster discovery, regression, and outlier identification into model fitting. Three MoE models of increasing complexity were tasked with determining respirable dust mass in 243 field samples from thirteen active coal, limestone, sandstone, and silver mines. All MoE models, including those using only "expert" spectroscopic predictors or a combination of expert and categorical "gate" variables (e.g., mine type), significantly outperform PLS in terms of accuracy (α = 0.05). Decomposing bias by mine type shows that accuracy generally improves across all types considered when MoE models are not overfitted. The MoE method's effectiveness was linked to its ability to endogenously classify outliers as well as possibly to the use of an additional cluster model for mass predictions. Overall, MoE methods appear as a capable and novel tool to addressing problems of heterogeneity for direct-on-filter quantitative analysis.

BACKGROUND: We aimed to develop a method for assessing occupational styrene exposures for application in epidemiological studies on risks of lymphohematopoietic neoplasms and other malignant and non-malignant diseases in the European and the US glass reinforced plastics industries. METHOD: We estimated a linear mixed effects model based on individual airborne personal measurements of styrene from the glass reinforced plastics industry in Denmark, Norway, Sweden, UK, and the US. The most suitable model was chosen based on its predictive power as assessed using cross validation with different combinations of predictors; and by comparing their prediction errors. RESULTS: We created a database containing 21,201 personal and area measurements but a subset of 14,440 personal measurements that spanned a period from 1962 to 2018, were used in the analysis. The selected model included fixed effects for year, sampling duration, measurement reason, product, process and random effects for country and worker. There was strong agreement between the model's predictions and actual exposure values indicating a good fit (Lin's CCC: 0.85 95% CI 0.84, 0.85). There were regional differences in exposure levels, with the UK and the US having comparable exposures that were higher than those in the Nordic countries. Higher exposures were consistent with measurements collected for inspection purposes, the lamination process, and specific products. Styrene exposure levels have decreased annually on average by 7%. CONCLUSION: Our exposure model and the resulting exposure predictions will enable estimation of lifetime occupational exposure for individual workers in the European and the US glass reinforced plastics industry and possibly related health risks among employees. The approach facilitates understanding of the uncertainty in our prediction model and can inform analysis of the bias that application of our exposure assessment approach can produce in epidemiologic analyses of exposure-response associations. Addressing systematic sources of bias can increase confidence in the conclusions of the epidemiologic analysis.

A significant portion of the work of developing and validating methods for volatile organic compound (VOC) sampling in workplace atmospheres involves the use of laboratory-generated atmospheres. The sample variability was evaluated from the dynamic atmosphere generation system used for VOC atmosphere generation and sampling. Characterization of the bias and variability of samples was done for a variety of atmospheres containing neat n-heptane and mixtures of VOCs sampled on activated coconut shell charcoal. Estimates of sampling variability ranged from 2% for neat n-heptane to 12% for a component in the 10 VOC mix. Sample variability increased for lower concentration samples and for mixtures of VOCs compared to single component atmospheres. This study can serve as a baseline for future atmosphere sampling experiments evaluating performance at lower concentrations and mixed VOC environments.

Introduction: The avoidance of asthma triggers, like tobacco smoke, facilitates asthma management. Reliance upon caregiver report of their child’s environmental tobacco smoke (ETS) exposure may result in information bias and impaired asthma management. This analysis aimed to characterize the chronicity of ETS exposure, assess the validity of caregiver report of ETS exposure, and investigate the relationship between ETS exposure and asthma attack. Methods: A secondary data analysis was performed on data from a longitudinal study of 162 children aged 7–12 years with asthma living in federally subsidized housing in three US cities (Boston, Cincinnati, and New Orleans). Data were collected at three time points over 1 year. Results: Over 90% of children were exposed to ETS (≥0.25 ng/ml of urine cotinine (UC)). Exposure was consistent over 1 year. Questionnaire data had a sensitivity of 28–34% using UC ≥0.25 ng/ml as the gold standard. High ETS exposure (UC ≥ 30 ng/ml) was significantly associated with asthma attack (aOR 2.97, 0.93–9.52, p = 0.07). Lower levels (UC 0.25–30 ng/ml) were not statistically significant (aOR 1.76, 0.71– 4.38, p = 0.22). No association was found using caregiver-reported ETS exposure. Conclusion: Relying on questionnaire data to assess children’s exposure to tobacco smoke may lead to substantial information bias. For children with asthma, incorrect characterization may substantially impact asthma morbidity. © The Author(s), 2024.

Unmeasured confounding is a major concern in many epidemiologic studies that are not randomized. Negative control methods can detect and reduce confounding by leveraging the proxies of the unmeasured confounders, including negative control outcomes (NCO) and exposures (NCE). An NCO is presumably unaffected by the exposure of interest but would be associated with unmeasured confounders; an NCE presumably does not affect the outcome of interest but would be associated with unmeasured confounders. A recently proposed double negative control method leverages both NCO and NCE for unmeasured confounding bias. To demonstrate this relatively new methodology in pharmacoepidemiologic studies, we re-analyzed data from a prior safety study of Recombinant Zoster Vaccine (RZV). The prior study compared risk of safety outcomes of RZV versus unvaccinated comparators, using logistic regression with propensity score adjustment. We identified NCOs and NCEs that could be used to adjust for unmeasured confounding bias that could arise if RZV recipients are incomparable to the comparators due to unmeasured factors. The double negative control analysis yielded relative risk estimates slightly closer to 1.0 than those reported previously, providing additional evidence of RZV safety that is less vulnerable to unmeasured confounding.

PURPOSE: To investigate differences in teen-reported and parent-reported lifetime prevalence estimates of traumatic brain injury (TBI) symptoms, TBI evaluation, and TBI diagnosis among a nationally representative sample of teenagers aged 12-17 years old and their parents. METHODS: Parent-reported data from the 2021 to 2022 National Health Interview Survey linked with teen-reported data from the National Health Interview Survey-Teen July 2021-December 2022 (n = 1,153) were analyzed. Lifetime prevalence estimates for TBI symptoms (e.g., selected symptoms as a result of a blow or jolt to the head), history of evaluation by health professional for TBI (i.e., TBI evaluation), and TBI diagnosis stratified by sociodemographic characteristics and reporter type were produced, and z-tests were conducted to test for differences. Concordance measures were calculated to assess agreement between teen and parent survey responses to TBI measures. RESULTS: Lifetime prevalence of TBI symptoms varied by reporter type across all sociodemographic characteristics with teen-report consistently producing higher estimates. Estimates of TBI evaluation varied by reporter type only among older teens, non-Hispanic teens, and teens who participated in sports; there was no difference for TBI diagnosis. Percent agreement between the 2 reporters ranged from 73% to 95%, prevalence-adjusted bias-adjusted kappa ranged from 0.45 to 0.90, and Cohen's kappa ranged from 0.22 to 0.63. DISCUSSION: There was general agreement for observable outcomes TBI evaluation and TBI diagnosis, but discordance existed in reports of TBI symptoms. These findings suggest that youth self-report of TBI symptoms may enhance surveillance efforts.

Test-negative designs (TNDs) are used to assess vaccine effectiveness (VE). Protection from infection-induced immunity may confound the association between case and vaccination status, but collecting reliable infection history can be challenging. If vaccinated individuals have less infection-induced protection than unvaccinated individuals, failure to account for infection history could underestimate VE, though the bias is not well understood. We simulated individual-level SARS-CoV-2 infection and COVID-19 vaccination histories and a TND. VE against symptomatic infection and VE against severe disease estimates unadjusted for infection history underestimated VE compared to estimates adjusted for infection history, and unadjusted estimates were more likely to be below 0%, which could lead to an incorrect interpretation that COVID-19 vaccines are harmful. TNDs assessing VE immediately following vaccine rollout introduced the largest bias and potential for negative VE against symptomatic infection. Despite the potential for bias, VE estimates from TNDs without prior infection information are useful because underestimation is rarely more than 8 percentage points.

BACKGROUND: In test-negative studies of vaccine effectiveness (VE), including patients with co-circulating, vaccine-preventable, respiratory pathogens in the control group for the pathogen of interest can introduce a downward bias on VE estimates. METHODS: A multicenter sentinel surveillance network in the US prospectively enrolled adults hospitalized with acute respiratory illness from September 1, 2022-March 31, 2023. We evaluated bias in estimates of VE against influenza-associated and COVID-19-associated hospitalization based on: inclusion vs exclusion of patients with a co-circulating virus among VE controls; observance of VE against the co-circulating virus (rather than the virus of interest), unadjusted and adjusted for vaccination against the virus of interest; and observance of influenza or COVID-19 against a sham outcome of respiratory syncytial virus (RSV). RESULTS: Overall VE against influenza-associated hospitalizations was 6 percentage points lower when patients with COVID-19 were included in the control group, and overall VE against COVID-19-associated hospitalizations was 2 percentage points lower when patients with influenza were included in the control group. Analyses of VE against the co-circulating virus and against the sham outcome of RSV showed that downward bias was largely attributable the correlation of vaccination status across pathogens, but also potentially attributable to other sources of residual confounding in VE models. CONCLUSION: Excluding cases of confounding respiratory pathogens from the control group in VE analysis for a pathogen of interest can reduce downward bias. This real-world analysis demonstrates that such exclusion is a helpful bias mitigation strategy, especially for measuring influenza VE, which included a high proportion of COVID-19 cases among controls.

Epidemiological delays are key quantities that inform public health policy and clinical practice. They are used as inputs for mathematical and statistical models, which in turn can guide control strategies. In recent work, we found that censoring, right truncation, and dynamical bias were rarely addressed correctly when estimating delays and that these biases were large enough to have knock-on impacts across a large number of use cases. Here, we formulate a checklist of best practices for estimating and reporting epidemiological delays. We also provide a flowchart to guide practitioners based on their data. Our examples are focused on the incubation period and serial interval due to their importance in outbreak response and modeling, but our recommendations are applicable to other delays. The recommendations, which are based on the literature and our experience estimating epidemiological delay distributions during outbreak responses, can help improve the robustness and utility of reported estimates and provide guidance for the evaluation of estimates for downstream use in transmission models or other analyses.

Test-negative design (TND) studies are cornerstones of vaccine effectiveness (VE) monitoring for influenza. The introduction of SARS-CoV-2 and RSV vaccines complicate the analysis of this design, with control selection restriction based on other pathogen diagnosis proposed as a solution. We conducted a simulation study and secondary analysis of 2017-18 and 2018-19 TND estimates from a Southeast Michigan ambulatory population to evaluate RSV-status-based control restriction. Simulations suggest that with vaccine-preventable RSV, influenza VE could be moderately biased with RSV prevalence ≥25 % of controls. Real-world analysis showed 151 influenza-negative adults (10.4 %) had RSV detected from the enrollment nasal swab. There were minimal differences in results of adjusted models with or without RSV exclusion from control groups. Findings suggest that inclusion of RSV cases in the control group of TND studies for influenza VE, particularly where RSV is not vaccine preventable, does not currently pose a major concern for bias in VE estimates.

Mumps outbreaks among fully vaccinated young adults have raised questions about potential waning of immunity over time and need for a third dose of the measles, mumps, rubella (MMR) vaccine. However, there are currently limited data on real-life effectiveness of the third-dose MMR vaccine in preventing mumps. Here, we used a deterministic compartmental model to infer the effectiveness of the third-dose MMR vaccine in preventing mumps cases by analyzing the mumps outbreak that occurred at the University of Iowa between August 24, 2015, and May 13, 2016. The modeling approach further allowed us to evaluate the population-level impact of vaccination by different timing in relation to the start of the outbreak and varied coverage levels, and to account for potential sources of bias in estimating vaccine effectiveness. We found large uncertainty in vaccine effectiveness estimates; however, our models showed that early introduction of a third dose of MMR vaccine during a mumps outbreak can be effective in preventing transmission. School holidays, such as the winter break, likely played important roles in preventing mumps transmission.

OBJECTIVES: To estimate the impact of occupational injury and illness on opioid-related mortality while accounting for confounding by preinjury opioid use. METHODS: We employed a retrospective cohort study design using Washington State workers' compensation data for 1994-2000 injuries linked to US Social Security Administration earnings and mortality data and National Death Index (NDI) cause of death data from 1994 to 2018. We categorised injuries as lost-time versus medical-only, where the former involved more than 3 days off work or permanent disability. We determined death status and cause of death from NDI records. We modelled separate Fine and Gray subdistribution hazard ratios (sHRs) and 95% CIs for injured men and women for opioid-related and all drug-related mortality through 2018. We used quantitative bias analysis to account for unmeasured confounding by preinjury opioid use. RESULTS: The hazard of opioid-related mortality was elevated for workers with lost-time relative to medical-only injuries: sHR for men: 1.53, 95% CI 1.41 to 1.66; for women: 1.31, 95% CI 1.16 to 1.48. Accounting for preinjury opioid use, effect sizes were reduced but remained elevated: sHR for men was 1.43, 95% simulation interval (SI) 1.20 to 1.69; for women: 1.27, 95% SI 1.10 to 1.45. CONCLUSIONS: Occupational injuries and illnesses severe enough to require more than 3 days off work are associated with an increase in the hazard of opioid-related mortality. The estimated increase is reduced when we account for preinjury opioid use, but it remains substantial. Reducing work-related injuries and postinjury opioid prescribing and improving employment and income security may decrease opioid-related mortality.

As part of the response to the highly pathogenic avian influenza A(H5N1) virus outbreak in U.S. cattle and poultry and the associated human cases, CDC and partners are monitoring influenza A virus levels and detection of the H5 subtype in wastewater. Among 48 states and the District of Columbia that performed influenza A testing of wastewater during May 12-July 13, 2024, a weekly average of 309 sites in 38 states had sufficient data for analysis, and 11 sites in four states reported high levels of influenza A virus. H5 subtype testing was conducted at 203 sites in 41 states, with H5 detections at 24 sites in nine states. For each detection or high level, CDC and state and local health departments evaluated data from other influenza surveillance systems and partnered with wastewater utilities and agriculture departments to investigate potential sources. Among the four states with high influenza A virus levels detected in wastewater, three states had corresponding evidence of human influenza activity from other influenza surveillance systems. Among the 24 sites with H5 detections, 15 identified animal sources within the sewershed or adjacent county, including eight milk-processing inputs. Data from these early investigations can help health officials optimize the use of wastewater surveillance during the upcoming respiratory illness season.

OBJECTIVES: CA 15-3 and CEA are tumor markers used in routine clinical care for breast cancer and colorectal cancer, among others. Current measurement procedures (MP) for these tumor markers are considered to be insufficiently harmonized. This study investigated the achievable harmonization for CA 15-3 and CEA by using an in silico simulation of external quality assessment (EQA) data from multiple EQA programs using patient-pool based samples. METHODS: CA 15-3 and CEA data from SKML (2021), UK NEQAS (2020-2021) and KEQAS (2020-2021) were used. A harmonization protocol was defined in which MPs that were considered equivalent were used to value assign EQA samples, and recalibration was only required if the MP had a bias of >5 % with value assigned EQA. Harmonization status was assessed by determining the mean level of agreement and residual variation by CV (%). RESULTS: Only MPs from Abbott, Beckman, Roche and Siemens were available in all EQA programs. For CA 15-3, recalibration was proposed for Beckman MP only and for CEA, recalibration was proposed for Siemens MP only. When the harmonization procedures were applied, for CA 15-3 the pre-harmonization mean bias range per MP was reduced from -29.28 to 9.86 %, into -0.09-0.12 % after harmonization. For CEA, the mean bias range per MP was reduced from -23.78 to 2.00 % pre-harmonization to -3.13-1.42 % post-harmonization. CONCLUSIONS: The present study suggests that a significant improvement in the harmonization status of CA 15-3 and CEA may be achieved by recalibration of a limited number of MPs.

Understanding whether influenza vaccine promotion strategies produce community-wide indirect effects is important for establishing vaccine coverage targets and optimizing vaccine delivery. Empirical epidemiologic studies and mathematical models have been used to estimate indirect effects of vaccines but rarely for the same estimand in the same dataset. Using these approaches together could be a powerful tool for triangulation in infectious disease epidemiology because each approach is subject to distinct sources of bias. We triangulated evidence about indirect effects from a school-located influenza vaccination program using two approaches: a difference-in-difference (DID) analysis, and an age-structured, deterministic, compartmental model. The estimated indirect effect was substantially lower in the mathematical model than in the DID analysis (2.1% (95% Bayesian credible intervals 0.4 - 4.4%) vs. 22.3% (95% CI 7.6% - 37.1%)). To explore reasons for differing estimates, we used sensitivity analyses and probabilistic bias analyses. When we constrained model parameters such that projections matched the DID analysis, results only aligned with the DID analysis with substantially lower pre-existing immunity among school-age children and older adults. Conversely, DID estimates corrected for potential bias only aligned with mathematical model estimates under differential outcome misclassification. We discuss how triangulation using empirical and mathematical modelling approaches could strengthen future studies.