Search

Julie Barberio,1 Thomas P Ahern,2 Richard F MacLehose,3 Lindsay J Collin,1 Deirdre P Cronin-Fenton,4 Per Damkier,5 Henrik Toft Sørensen,4 Timothy L Lash1 1Rollins School of Public Health, Emory University, Atlanta, GA, USA; 2The Robert Larner College of Medicine, University of Vermont, Burlington, VT, USA; 3University of Minnesota, Minneapolis, MN, USA; 4Aarhus University Hospital, Aarhus, Denmark; 5Odense University Hospital, Odense, DenmarkCorrespondence: Julie BarberioRollins School of Public Health, Emory University, Atlanta, GA, USAEmail julie.barberio@emory.eduPurpose: To compare the magnitude of bias due to unmeasured confounding estimated from various techniques in an applied example.Patients and Methods: We examined the association between dibutyl phthalate (DBP) and incident estrogen receptor (ER)-positive breast cancer in a Danish nationwide cohort (N=1,122,042). Cox regression analyses were adjusted for age and active drug compounds contributing to DBP exposure. We estimated the hazard ratios (HRs) that would have been observed had one of the DBP sources been unmeasured and calculated the strength of confounding by comparing to the fully adjusted HR. We performed a quantitative bias analysis (QBA) of the “unmeasured” confounder, using external information to specify the bias parameters. Upper bounds on the bias were estimated and E-values were calculated.Results: The adjusted HR for incident ER-positive breast cancer among women with high-level (≥ 10,000 cumulative milligrams) versus no DBP exposure was 2.12 (95% confidence interval 1.12 to 4.05). Removing each DBP source in isolation resulted in negligible change in the HR. The bias estimates from the QBA ranged from 1.00 to 1.01. The estimated maximum impact of unmeasured confounding ranged from 1.01 to 1.51. E-values ranged from 3.46 to 3.68.Conclusion: The impact of bias due to simulated unmeasured confounding was negligible, in part, because the unmeasured variable was not independent of controlled variables. When a suspected confounder cannot be measured in the study data, our exercise suggests that QBA is the most informative method for assessing the impact. E-values may best be reserved for situations where uncontrolled confounding emanates from an unknown confounder.Keywords: bias analysis, unmeasured confounding, the E-value

Lindsay J Collin,1,2 Anders H Riis,2 Richard F MacLehose,3 Thomas P Ahern,4 Rune Erichsen,2,5 Ole Thorlacius-Ussing,6 Timothy L Lash1 1Department of Epidemiology, Emory University, Atlanta, GA, USA; 2Department of Clinical Epidemiology, Aarhus University Hospital, Aarhus, Denmark; 3Division of Epidemiology and Community Health, University of Minnesota, Minneapolis, MN, USA; 4Department of Surgery, The Robert Larner, M.D. College of Medicine at the University of Vermont, Burlington, VT, USA; 5Department of Surgery, Randers Regional Hospital, Randers, Denmark; 6Department of Gastrointestinal Surgery, Aalborg University Hospital, Aalborg, DenmarkCorrespondence: Lindsay J CollinDepartment of Epidemiology, Emory University, 1518 Clifton RoadNE, Atlanta, GA 30322, USATel +1 530-386-3341Email lindsay.jane.collin@emory.eduBackground: Among men and women diagnosed with colorectal cancer (CRC), 20– 50% will develop a cancer recurrence. Cancer recurrences are not routinely captured by most population-based registries; however, linkage across Danish registries allows for the development of predictive models to detect recurrence. Successful application of such models in population-based settings requires validation against a gold standard to ensure the accuracy of recurrence identification.Objective: We apply a recently developed validation study design for prospectively collected validation data to validate predicted CRC recurrences against gold standard diagnoses from medical records in an actively followed cohort of CRC patients in Denmark.Methods: We use a Bayesian monitoring framework, traditionally used in clinical trials, to iteratively update classification parameters (positive and negative predictive values, and sensitivity and specificity) in an adaptive validation substudy design. This design allows determination of the sample size necessary to estimate the corresponding parameters and to identify when validation efforts can cease based on predefined criteria for parameter values and levels of precision.Results: Among 355 men and women diagnosed with CRC in Denmark and actively followed semi-annually, there were 63 recurrences diagnosed by active follow-up and 70 recurrences identified by a predictive algorithm. The adaptive validation design met stopping criteria for the classification parameters after 120 patients had their recurrence information validated. This stopping point yielded parameter estimates for the classification parameters similar to those obtained when the entire cohort was validated, with 66% less patients needed for the validation study.Conclusion: In this proof of concept application of the adaptive validation study design for outcome misclassification, we demonstrated the ability of the method to accurately determine when sufficient validation data have been collected. This method serves as a novel validation substudy design for prospectively collected data with simultaneous implementation of a validation study.Keywords: validation study design, colorectal cancer recurrence

Background: Several cardiovascular risk factors have been associated with the risk of atrial fibrillation (AF). Limited and inconsistent evidence exists on the association of blood lipid levels and lipid-lowering medication use with AF risk.Methods and Results: We analyzed 13 969 participants (25% African American, 45% men) free of AF at baseline from the Atherosclerosis Risk in Communities study. Fasting high-density lipoprotein cholesterol (HDLc), low-density lipoprotein cholesterol (LDLc), triglycerides, and total cholesterol were measured at baseline (1987-1989) and each of 3 follow-up visits. The incidence of AF was ascertained through 2007. The association of the use of statins and other lipid-lowering medications with AF was estimated in 13 044 Atherosclerosis Risk in Communities participants attending visit 2 (1990-1992), adjusting for covariates from the previous visit. During a median follow-up of 18.7 years, there were 1433 incident AF cases. Multivariable hazard ratios (HRs) and 95% CIs of AF associated with a 1-SD increase in lipid levels were as follows: HDLc, 0.97 (0.91-1.04); LDLc, 0.90 (0.85-0.96); total cholesterol, 0.89 (0.84-0.95); and triglycerides, 1.00 (0.96-1.04). Participants taking lipid-lowering medications had an adjusted HR (95% CI) of AF of 0.96 (0.82-1.13) compared with those not taking medications, whereas those taking statins had an adjusted HR of 0.91 (0.66-1.25) compared with those taking other lipid-lowering medications.Conclusions: Higher levels of LDLc and total cholesterol were associated with a lower incidence of AF. However, HDLc and triglycerides were not independently associated with AF incidence. No association was found between the use of lipid-lowering medications and incident AF. [ABSTRACT FROM AUTHOR]

Case-control studies are particularly susceptible to differential exposure misclassification when exposure status is determined following incident case status. Probabilistic bias analysis methods have been developed as ways to adjust standard effect estimates based on the sensitivity and specificity of exposure misclassification. The iterative sampling method advocated in probabilistic bias analysis bears a distinct resemblance to a Bayesian adjustment; however, it is not identical. Furthermore, without a formal theoretical framework (Bayesian or frequentist), the results of a probabilistic bias analysis remain somewhat difficult to interpret. We describe, both theoretically and empirically, the extent to which probabilistic bias analysis can be viewed as approximately Bayesian. Although the differences between probabilistic bias analysis and Bayesian approaches to misclassification can be substantial, these situations often involve unrealistic prior specifications and are relatively easy to detect. Outside of these special cases, probabilistic bias analysis and Bayesian approaches to exposure misclassification in case-control studies appear to perform equally well. [ABSTRACT FROM AUTHOR]

OBJECTIVE: We evaluated a series of questions pertaining to vulvar pain symptoms to determine their association with a localized vulvodynia (vestibulodynia) diagnosis in women from the general population. METHODS: A sample of 12,435 women completed a self-administered screening questionnaire for the presence of specific types and characteristics of vulvar pain lasting 3 months or longer. Sensitivity, specificity, and predictive values were calculated for each cross-classification of vulvar pain type and characteristic, using as the gold standard 121 subjects with a clinically confirmed vestibulodynia diagnosis. RESULTS: Relative to women with clinically confirmed vestibulodynia, 83% reported >10 episodes of pain on contact at the time of tampon insertion, intercourse, or pelvic examination, and 83% also reported pain on contact that limited or prevented sexual intercourse. These strong associations with a vestibulodynia diagnosis were not observed with respect to women who reported vulvar pain symptoms of burning or knifelike pain, or vulvar pain characteristics of continuous versus intermittent pain, or provoked versus spontaneous pain. CONCLUSIONS: Our findings suggest that a small number of symptoms may be suitable for identifying a large proportion of women suffering from vestibulodynia which may be ideal for the development of an effective screening test in the future. However, we also recognize that a large proportion of women experiencing vulvar pain symptoms will not meet the diagnostic criteria for vestibulodynia. Thus, implementing such a screening procedure as part of a routine examination or testing would require a subsequent pelvic examination to confirm a vestibulodynia diagnosis and to rule out other known explanations for vulvar pain. [ABSTRACT FROM AUTHOR]

Study Objectives: To examine the association of empiric inpatient antibiotic treatment of community-acquired pneumonia (CAP) with mortality, and whether this association varies from year to year.Design: Population-based, retrospective study adjusting for demographics, comorbidities, and clinical characteristics.Setting: Acute-care hospitals in 10 western states.Patients: A group of 10,069 Medicare beneficiaries aged > or = 65 years who were hospitalized with CAP during fiscal years 1993, 1995, and 1997.Measurements and Results: We examined the risk for mortality during the 30 days after admission to the hospital. The impact of specific antibiotic regimens varied greatly from year to year. In 1993, therapy with a macrolide plus a beta-lactam was associated with significantly lower mortality than therapy with either a beta-lactam alone (adjusted odds ratio [AOR], 0.42; 95% confidence interval [CI], 0.25 to 0.69) or other regimens that did not include a macrolide, beta-lactam, or fluoroquinolone (AOR, 0.35; 95% CI, 0.20 to 0.62). Those associations were not observed in 1995 or 1997. Lower mortality was associated with fluoroquinolone monotherapy compared with beta-lactam monotherapy in 1997 (AOR, 0.27; 95% CI, 0.07 to 0.96) and with macrolide monotherapy compared with other regimens in 1995 (AOR, 0.24; 95% CI, 0.06 to 0.93), but the number of patients who received these regimens was small.Conclusions: The inclusion of a macrolide or a fluoroquinolone in initial empiric CAP treatment was associated with improved survival, but this association varied from year to year, perhaps as a result of a temporal variation in the incidence of atypical pathogen pneumonia. Improved testing and surveillance for atypical pathogen pneumonia are needed to guide empiric therapy. [ABSTRACT FROM AUTHOR]

Background: To account for misclassification of dichotomous variables using probabilistic bias analysis, beta distributions are often assigned to bias parameters (e.g., positive and negative predictive values) based on data from an internal validation substudy. Due to the small sample size of validation substudies, zero-cell frequencies can occur. In these scenarios, it may be helpful to assign prior distributions or apply continuity corrections to the predictive value estimates., Methods: We simulated cohort studies of varying sizes, with a binary exposure and outcome and a true risk ratio (RR) = 2.0, as well as internal validation substudies, to account for exposure misclassification. We conducted bias adjustment under five approaches assigning prior distributions to the positive and negative predictive value parameters: (1) conventional method (i.e., no prior), (2) uniform prior beta ( α = 1, β = 1), (3) Jeffreys prior beta ( α = 0.5, β = 0.5), (4) using Jeffreys prior as a continuity correction only when zero cells occurred, and (5) using the uniform prior as a continuity correction only when zero cells occurred. We evaluated performance by measuring coverage probability, bias, and mean squared error., Results: For sparse validation data, methods (2)-(5) all had better coverage and lower mean squared error than the conventional method, with the uniform prior (2) yielding the best performance. However, little difference between methods was observed when the validation substudy did not contain zero cells., Conclusion: If sparse data are expected in a validation substudy, using a uniform prior for the beta distribution of bias parameters can improve the validity of bias-adjusted measures., Competing Interests: Disclosure: T.L.L. is a member of the Amgen Epidemiology Methods Advisory Council, for which he receives consulting fees and travel support. Q.Z.’s doctoral training is supported by a grant from Amgen Inc. to Emory University. L.J.C. and Q.Z. report personal fees from Epidemiologic Research & Methods, LLC outside of the submitted work. The remaining authors report no conflicts of interest., (Copyright © 2024 Wolters Kluwer Health, Inc. All rights reserved.)

Objectives: An excess of mesothelioma has been previously observed in iron ore miners in Northeastern Minnesota. This study explored the potential association between mesothelioma and elongate mineral particle (EMP) exposures in the Minnesota taconite mining industry, examining both regulated and non-regulated EMP dimensions., Methods: A nested case-control study design within the Mineral Resources Health Assessment Programme cohort analysed 104 mesothelioma cases and 410 controls. Cumulative EMP exposures were assessed across various dimensional definitions, including regulated (National Institute for Occupational Safety and Health, NIOSH) and non-regulated EMPs. Conditional logistic regression models were applied to estimate mesothelioma risk linked to employment duration and cumulative EMP exposure while adjusting for potential confounding variables., Results: Consistent with earlier assessments, mesothelioma was associated with the number of years employed in the taconite industry (rate ratio (RR) 1.02, 95% CI 1.00 to 1.05) and cumulative NIOSH EMP exposure ((EMP/cc)×years) in taconite mining and processing (RR 1.20, 95% CI 0.99 to 1.46). Positive associations were also observed with mesothelioma and cumulative exposure to non-regulated EMPs., Conclusions: This study supported the main conclusions in the previous study that the data were consistent with an association between mesothelioma and employment duration and with NIOSH EMP exposure in taconite mining and processing. However, the 95% CI indicates the data are also compatible with a null effect. Given the limitations we cite in the manuscript, additional study is needed to clarify the effect. Additionally, this study found possible evidence of a positive association between mesothelioma and cumulative exposure to Chatfield EMP, Suzuki EMP and Cleavage Fragments in the Minnesota taconite worker population., Competing Interests: Competing interests: None declared., (© Author(s) (or their employer(s)) 2025. No commercial re-use. See rights and permissions. Published by BMJ Group.)

Background: We evaluated the impact of systematic bias due to loss of heterozygosity (LOH) and incomplete phenotyping in studies examining the relationship between CYP2D6 variants and breast cancer recurrence among women treated with tamoxifen., Methods: We performed a systematic review of the literature on tamoxifen, CYP2D6 variants, and breast cancer recurrence. A quantitative bias analysis was performed to adjust for LOH and incomplete phenotyping. Bias-adjusted results were then combined in a meta-analysis., Results: Thirty-three studies informed the bias analysis and meta-analysis on CYP2D6 variants and breast cancer recurrence and/or mortality. An unadjusted meta-analysis suggested increased risk of recurrence and/or mortality for poor relative to normal metabolizers [RR = 1.28; 95% simulation interval (SI), 1.04-1.58] with substantial heterogeneity (I2 = 27%; P for heterogeneity = 0.07). Adjusting for LOH and incomplete genotyping resulted in a slight change in the effect estimate and a decrease in heterogeneity (RR = 1.34; 95% SI, 1.10-1.63; I2 = 0%; P for heterogeneity = 0.17). Intermediate metabolizers had a slightly increased risk of recurrence and/or mortality relative to normal metabolizers (RR = 1.15; 95% SI, 1.00-1.34; I2 = 0%; P for heterogeneity = 0.89)., Conclusions: Adjusting for biases such as LOH and incomplete genotyping reduced observed heterogeneity between studies. Individuals with poor CYP2D6 phenotypes were at increased risk for breast cancer outcomes compared with those with normal phenotypes., Impact: Reduction in CYP2D6 activity was associated with an increased risk of breast cancer recurrence and/or mortality, and results underscore the importance of quantitatively adjusting for biases when aggregating study results. See related In the Spotlight, p. 221., (©2024 American Association for Cancer Research.)

Little is known about the relationships between demographic and economic social determinants of health and the probability of contracting COVID-19 in American Indian and Alaska Native (AI/AN) peoples. In addition, we do not know if and how tribal payments, unique to AI/AN peoples, are associated with the probability of contracting COVID-19. We surveyed 767 AI/AN patients of five geographically disparate health organizations that primarily served AI/AN peoples in urban settings between January and May of 2021. We used univariate modified Poisson regressions to estimate the influence of age, gender, household composition, education, household income, and tribal payments on risk of contracting COVID-19, with results presented as both risk and risk difference. Fifteen percent of the sample contracted COVID-19, and individuals who lived in households with two or more generations had an 11-percentage point elevated risk of contracting COVID-19 compared to those who lived alone. Twenty-seven percent of participants received tribal payments; receipt was associated with seven percentage points (change from 18% probability to 11% probability) lower risk of contracting COVID-19. Our findings showed interventions specifically designed to reduce the spread of COVID-19 in multigenerational households, and regular tribal payments may help improve health outcomes in urban AI/AN populations., Competing Interests: Declaration of Conflicting InterestsThe authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Background: Empirically evaluating the potential impact of recall bias on observed associations of prenatal medication exposure is crucial., Objective: We sought to assess the effects of exposure misclassification on previous studies of the use of nonsteroidal anti-inflammatory drugs (NSAIDs) in early pregnancy and increased risk of amniotic band syndrome (ABS)., Methods: Using data from the National Birth Defects Prevention Study (NBDPS) on births from 1997 to 2011, we included 189 mothers of infants with ABS and 11,829 mothers of infants without congenital anomalies. We identified external studies of medication use during pregnancy to obtain validity parameters for a probabilistic bias analysis to adjust for exposure misclassification. Due to uncertainty about the transportability of these parameters, we conducted multidimensional bias analyses to explore combinations of values on the results., Results: When we assumed higher specificity in cases or higher sensitivity in controls, misclassification-adjusted estimates suggested confounding-adjusted estimates were attenuated. However, in a few instances, when we assumed greater sensitivity in the cases than the controls (and Sp ≥ 0.9), the misclassification-adjusted estimates suggested upward bias in the confounding-adjusted estimates., Conclusions: Results from our bias analysis highlighted that the magnitude of bias depended on the mechanism and the extent of misclassification. However, the parameters available from the validation studies were not directly applicable to our study. In the absence of reliable validation studies, considering mechanisms of bias and simulation studies to outline combinations of plausible scenarios to better inform conclusions on the effects of these medications on pregnancy outcomes remains important., (© 2024 John Wiley & Sons Ltd.)

Background: A 2013 meta-analysis observed a protective association between overweight BMI (versus normal BMI) and all-cause mortality that was particularly strong in people aged ≥65. Estimates informing this meta-analysis were highly heterogeneous, and critics raised insufficient or inappropriate confounder adjustment in many studies as an explanation for the protective summary association. Using this topic as an example, we demonstrate a novel approach for external adjustment of individual studies for a uniform and sufficient confounder set before meta-analysis., Methods: We abstracted summary data on the 33 associations comprising the age ≥65 stratum of the 2013 meta-analysis. Using an external dataset (NHANES III), we derived covariates used in each study's multivariable model of the overweight-mortality association. We then calculated a bias factor to quantify the direction and magnitude of displacement of the ratio measure of association after changing from the original adjustment set to a sufficient adjustment set. After applying bias factors to adjust original associations, we compared summary results from random effects meta-analyses with and without such adjustment., Results: We reproduced the original meta-analysis of overweight-mortality estimates among older participants and found a protective association similar to that reported in 2013 (summary RR=0.88, 95% CI: 0.84, 0.92, I 2 =38.4%). After we simulated uniform adjustment of all 33 associations for a minimally sufficient confounder set (age, sex, and smoking status), the meta-analysis showed a similar summary association (summary RR=0.90, 95% CI: 0.86, 0.94), but with reduced heterogeneity (I 2 =34.6%)., Conclusion: Simulated uniform adjustment for a sufficient confounder set may improve rigor and promote consensus in meta-analysis., Competing Interests: Conflict of interest: No author declares a conflict of interest., (Copyright © 2024 Wolters Kluwer Health, Inc. All rights reserved.)

Background: Childhood maltreatment is associated with elevated adult weight. It is unclear whether this association extends to pregnancy, a critical window for the development of obesity., Methods: We examined associations of childhood maltreatment histories with prepregnancy body mass index (BMI) and gestational weight gain among women who had participated for >20 years in a longitudinal cohort. At age 26-35 years, participants reported childhood maltreatment (physical, sexual, and emotional abuse; emotional neglect) and, 5 years later, about prepregnancy weight and gestational weight gain for previous pregnancies (n = 656). Modified Poisson regression models were used to estimate associations of maltreatment history with prepregnancy BMI and gestational weight gain z -scores, adjusting for sociodemographics. We used multivariate imputation by chained equations to adjust outcome measures for misclassification using data from an internal validation study., Results: Before misclassification adjustment, results indicated a higher risk of prepregnancy BMI ≥30 kg/m 2 in women with certain types of maltreatment (e.g., emotional abuse risk ratio = 2.4; 95% confidence interval: 1.5, 3.7) compared with women without that maltreatment type. After misclassification adjustment, estimates were attenuated but still modestly elevated (e.g., emotional abuse risk ratio = 1.7; 95% confidence interval: 1.1, 2.7). Misclassification-adjusted estimates for maltreatment associations with gestational weight gain z -scores were close to the null and imprecise., Conclusions: Findings suggest an association of maltreatment with prepregnancy BMI ≥30 kg/m 2 but not with high gestational weight gain. Results suggest a potential need for equitable interventions that can support all women, including those with maltreatment histories, as they enter pregnancy., Competing Interests: Disclosure: The authors report no conflicts of interest., (Copyright © 2024 Wolters Kluwer Health, Inc. All rights reserved.)

Mismeasurement of a dichotomous outcome yields an unbiased risk ratio estimate when there are no false positive cases (perfect specificity) and when sensitivity is non-differential with respect to exposure status. In studies where these conditions are expected, quantitative bias analysis may be considered unnecessary. We conducted a simulation study to explore the robustness of this special case to small departures from perfect specificity and stochastic departures from non-differential sensitivity. We observed substantial bias of the risk ratio with specificity values as high at 99.8%. The magnitude of bias increased directly with the true underlying risk ratio and was markedly stronger at lower baseline risk. Stochastic departure from non-differential sensitivity also resulted in substantial bias in most simulated scenarios; downward bias prevailed when sensitivity was higher among unexposed compared with exposed, and upward bias prevailed when sensitivity was higher among exposed compared with unexposed. Our results show that seemingly innocuous departures from perfect specificity (e.g., 0.2%) and from non-differential sensitivity can yield substantial bias of the risk ratio under outcome misclassification. We present a web tool permitting easy exploration of this bias mechanism under user-specifiable study scenarios., (© The Author(s) 2024. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.)

Severe asthma has been shown to occur in the combined presence of high pollen and thunderstorm conditions, also known as 'thunderstorm asthma.' First studied as severe epidemic events, recent longitudinal work studied less dramatic but more frequent occurrences. We explore thunderstorm asthma-related emergency department visits in the Minneapolis-St. Paul metropolitan area and evaluated risk differences by sex and age. We define a thunderstorm asthma exposure event as the daily occurrence of 2 or more lightning strikes during high pollen periods, and use daily counts of asthma-related emergency department visits to estimate relative and absolute risk of severe asthma during thunderstorm asthma events for the full population and for sex and age subgroups. The overall population had a 1.06 (95 % CI: 1.02, 1.09) times higher risk of asthma-related ED visits during thunderstorm asthma events compared to days without thunderstorm asthma events. Children under 18 show no higher risk (RR 1.02; 95 % CI: 0.97 1.08), but adults 18-44 years (RR 1.08; 95 % CI: 1.02, 1.13) and 45 and up (RR 1.08; 95 % CI 1.02, 1.15) show higher relative risk. Absolute risk measures show similar patterns to the age and sex results, but age-sex subgroups show more variation in absolute vs relative risk. Our results support an association between ED visits and thunderstorm asthma and provide evidence of varying risks by sex across the life course. These differences in risk have implications for clinical treatment of this allergic type of asthma and for future research into this poorly recognized environmental exposure. Plain Language Summary: Recent research has highlighted the existence of Thunderstorm asthma events, a phenomenon in which pollen grains rupture in the conditions that occur with a thunderstorm, releasing subpollen particles that are capable of triggering severe asthma in susceptible populations. Where severe asthma is a disease that usually impacts children, we find in this study that asthma ED visits associated with thunderstorm asthma events more frequently impact adults, particularly males 18-44 and females 45 and up., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Direct oral anticoagulants (DOACs; rivaroxaban, apixaban) and warfarin are approved for venous thromboembolism (VTE) treatment. Few direct comparisons of DOACs on risk of mortality among VTE patients exist, and for patients with concomitant conditions (e.g., kidney and liver disease) clinical guidelines are unclear. We evaluated 6-month all-cause mortality by anticoagulant prescribed for primary treatment of VTE. Using Medicare 20% sample data, we created a propensity score matched analytic dataset of 47,860 beneficiaries with non-cancer incident VTE. We used Cox regression to estimate adjusted hazard ratios (HRs) of OACs with 6-month mortality, and tested interactions by liver/kidney disease. There were 3,422 deaths over 6 months of follow-up. In adjusted models, patients prescribed rivaroxaban [HR: 0.82 (95% CI: 0.76-0.90)] had lower rates of mortality versus warfarin. There was no association comparing apixaban to warfarin [HR: 0.96 (95% CI: 0.87-1.07)]. In head-to-head comparisons of apixaban versus rivaroxaban the HR was 1.14 (95% CI: 1.01-1.28)]. Findings were similar among patients with liver and kidney disease. Overall, risk of death was similar by OAC prescribed. Though it is possible residual confounding remained, there was some suggestion of lower risk with rivaroxaban than warfarin. DOACs appear safe among VTE patients with concomitant kidney or liver disease., (© The Author(s) 2024. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.)

Background: Prepregnancy body mass index (BMI) and gestational weight gain (GWG) are determinants of maternal and child health. However, many studies of these factors rely on error-prone self-reported measures., Methods: Using data from Life-course Experiences And Pregnancy (LEAP), a US-based cohort, we assessed the validity of prepregnancy BMI and GWG recalled on average 8 years postpartum against medical record data treated as alloyed gold standard ("true") values. We calculated probabilities of being classified into a self-reported prepregnancy BMI or GWG category conditional on one's true category (analogous to sensitivities and specificities) and probabilities of truly being in each prepregnancy BMI or GWG category conditional on one's self-reported category (analogous to positive and negative predictive values)., Results: There was a tendency toward under-reporting prepregnancy BMI. Self-report misclassified 32% (95% confidence interval [CI] = 19%, 48%) of those in LEAP with truly overweight and 13% (5%, 27%) with obesity into a lower BMI category. Self-report correctly predicted the truth for 72% (55%, 84%) with self-reported overweight to 100% (90%, 100%) with self-reported obesity. For GWG, both under- and over-reporting were common; self-report misclassified 32% (15%, 55%) with truly low GWG as having moderate GWG and 50% (28%, 72%) with truly high GWG as moderate or low GWG. Self-report correctly predicted the truth for 45% (25%, 67%) with self-reported high GWG to 85% (76%, 91%) with self-reported moderate GWG. Misclassification of BMI and GWG varied across maternal characteristics., Conclusion: Findings can be used in quantitative bias analyses to estimate bias-adjusted associations with prepregnancy BMI and GWG., Competing Interests: The authors report no conflicts of interest., (Copyright © 2024 Wolters Kluwer Health, Inc. All rights reserved.)

Background: Food insecurity is an important social determinant of health that was exacerbated by the COVID-19 pandemic. Both food insecurity and COVID-19 infection disproportionately affect racial and ethnic minority groups, particularly American Indian and Alaska Native communities; however, there is little evidence as to whether food insecurity is associated with COVID-19 infection or COVID-19 preventive behaviors such as vaccination uptake. The purpose of this study was to evaluate associations between food insecurity, COVID-19 infection, and vaccination status among urban American Indian and Alaska Native adults seen at 5 clinics serving urban Native people., Methods: In partnership with health organizations in Alaska, Colorado, Kansas, Minnesota, and New Mexico, the study team conducted a cross-sectional survey in 2021 to assess food security status and attitudes, barriers, and facilitators for COVID-19 testing and vaccination. Logistic regression was used to examine the association of food security status with sociodemographic factors and COVID-19 infection and vaccination status. Marginal standardization was applied to present results as prevalence differences., Results: Among 730 American Indian and Alaska Native adults, the prevalence of food insecurity measured during the pandemic was 38%. For participants who reported persistent food security status before and during the pandemic (n = 588), the prevalence of food insecurity was 25%. Prevalence of COVID-19 infection and vaccination did not vary by food security status after adjustment for confounders., Conclusions: High rates of food insecurity among American Indian and Alaska Native communities likely increased during the COVID-19 pandemic. However, despite the high prevalence of food insecurity, community-led efforts to reduce COVID-19 infection and increase vaccination uptake across Indian Health Service and Tribal healthcare facilities may have mitigated the negative impacts of the pandemic for families experiencing food insecurity. These successful approaches serve as an important reference for future public health efforts that require innovative strategies to improve overall health in American Indian and Alaska Native communities., (© 2024. The Author(s).)

Background: Food insecurity is an important social determinant of health that was exacerbated by the COVID-19 pandemic. Both food insecurity and COVID-19 infection disproportionately affect racial and ethnic minority groups, particularly American Indian and Alaska Native communities; however, there is little evidence as to whether food insecurity is associated with COVID-19 infection or COVID-19 preventive behaviors such as vaccination uptake. The purpose of this study was to evaluate associations between food insecurity, COVID-19 infection, and vaccination status among urban American Indian and Alaska Native adults seen at 5 clinics serving urban Native people., Methods: In partnership with health organizations in Alaska, Colorado, Kansas, Minnesota, and New Mexico, the study team conducted a cross-sectional survey in 2021 to assess food security status and attitudes, barriers, and facilitators for COVID-19 testing and vaccination. Logistic regression was used to examine the association of food security status with sociodemographic factors and COVID-19 infection and vaccination status. Marginal standardization was applied to present results as prevalence differences., Results: Among 730 American Indian and Alaska Native adults, the prevalence of food insecurity measured during the pandemic was 38%. For participants who reported persistent food security status before and during the pandemic (n=588), the prevalence of food insecurity was 25%. Prevalence of COVID-19 infection and vaccination did not vary by food security status after adjustment for confounders., Conclusions: High rates of food insecurity among American Indian and Alaska Native communities likely increased during the COVID-19 pandemic. However, despite the high prevalence of food insecurity, community-led efforts to reduce COVID-19 infection and increase vaccination uptake across Indian Health Service and Tribal healthcare facilities may have mitigated the negative impacts of the pandemic for families experiencing food insecurity. These successful approaches serve as an important reference for future public health efforts that require innovative strategies to improve overall health in AIAN communities., Competing Interests: Competing interests The authors declare that they have no competing interests.

Systematic error from selection bias, uncontrolled confounding, and misclassification is ubiquitous in epidemiologic research but is rarely quantified using quantitative bias analysis (QBA). This gap may in part be due to the lack of readily modifiable software to implement these methods. Our objective is to provide computing code that can be tailored to an analyst's dataset. We briefly describe the methods for implementing QBA for misclassification and uncontrolled confounding and present the reader with example code for how such bias analyses, using both summary-level data and individual record-level data, can be implemented in both SAS and R. Our examples show how adjustment for uncontrolled confounding and misclassification can be implemented. Resulting bias-adjusted point estimates can then be compared to conventional results to see the impact of this bias in terms of its direction and magnitude. Further, we show how 95% simulation intervals can be generated that can be compared to conventional 95% confidence intervals to see the impact of the bias on uncertainty. Having easy to implement code that users can apply to their own datasets will hopefully help spur more frequent use of these methods and prevent poor inferences drawn from studies that do not quantify the impact of systematic error on their results., (© The Author(s) 2023; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association.)

Venous thromboembolism (VTE) affects 1.2 million people per year in the United States. With several clinical changes in diagnosis and treatment approaches in the past decade, we evaluated contemporary post-VTE mortality risk profiles and trends. Incident VTE cases were identified from the 2011-2019 Medicare 20% Sample, which is representative of nearly all Americans aged 65 and older. The social deprivation index was linked from public data; race/ethnicity and sex were self-reported. The all-cause mortality risk 30 days and 1 year after incident VTE was calculated in demographic subgroups and by prevalent cancer diagnosis status using model-based standardization. Risks for major cancer types, risk differences by age, sex, race/ethnicity, and socio-economic status (SES), and trends over time are also reported. The all-cause mortality risk among older US adults following incident VTE was 3.1% (95% CI 3.0-3.2) at 30 days and 19.6% (95% CI 19.2-20.1) at 1 year. For cancer-related VTE events, the age-sex-race-standardized risk was 6.0% at 30 days and 34.7% at 1 year. The standardized 30-day and 1-year risks were higher among non-White beneficiaries and among those with low SES. One-year mortality risk decreased 0.28 percentage points per year (95% CI 0.16-0.40) on average across the study period, with no trend observed for 30-day mortality risk. In sum, all-cause mortality risk following incident VTE has decreased slightly in the last decade, but racial and socio-economic disparities persist. Understanding patterns of mortality among demographic subgroups and in cancer-associated events is important for targeting efforts to improve VTE management., (© 2023 Wiley Periodicals LLC.)

People living with HIV face multiple psychosocial challenges. In a large, predominantly rural Ethiopian region, 1799 HIV patients new to care were enrolled from 32 sites in a cluster randomized trial using trained community support workers with HIV to provide individual health education, counseling and social support. Participants received annual surveys through 36 months using items drawn from the Centre for Epidemiologic Studies Depression Scale-10, Medical Outcome Study Social Support Survey, and HIV/AIDS Stigma Instrument-PLWA. At 12 months (using linear mixed effects regression models controlling for enrollment site clustering), intervention participants had greater emotional/informational and tangible assistance social support scores, and lower scores assessing depression symptoms and negative self-perception due to HIV status. A significant treatment effect at 36 months was also seen on scores assessing emotional/informational social support, depression symptoms, and internalized stigma. An intervention using peer community support workers with HIV to provide individualized informational and psychological support had a positive impact on the emotional health of people living with HIV who were new to care.(ClinicalTrials.gov protocol ID: 1410S54203, May 19, 2015)., (© 2023. The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.)

Objective: To estimate the prevalence of dementia among American Indian and Alaska Native (AIAN) Medicare beneficiaries and compare the prevalence of dementia in AIANs and other racial and ethnic minority groups-non-Hispanic Blacks (NHBs) and Hispanics-with non-Hispanic Whites (NHWs) Medicare beneficiaries., Methods: We used survey data from Round 5 of the National Health and Aging Trends Study (NHATS, 2015) (N = 7,449), a nationally representative study of Medicare beneficiaries ages 65 years and older. We used logistic regression to estimate the age- and sex-adjusted prevalence of dementia with 95% confidence intervals by race/ethnicity as well as prevalence differences between groups., Results: The majority of participants were between 65 and 74 years old. Approximately half of them were female. NHWs had a prevalence of 5% for dementia, and AIAIs had a prevalence of 9%, four percentage points higher than NHWs (95%CI 1%, 11%). We estimated a similar difference in the prevalence of dementia in AIAN Medicare beneficiaries compared to NHBs., Conclusion: While previous research has reported that AIANs share a similar or lower prevalence of dementia than NHWs, our findings suggest a potentially higher prevalence of dementia among AIAN Medicare beneficiaries. A relatively small number of AIAN resulted in wide confidence intervals for many of our associations. Future research should focus on recruiting a larger sample and on dementia prevalence and unique risk factors within and among AIAN tribes., (© 2022. W. Montague Cobb-NMA Health Institute.)

Background: Opioid-related mortality is an important public health problem in the United States. Incidence estimates rely on death certificate data generated by health care providers and medical examiners. Opioid overdoses may be underreported when other causes of death appear plausible. We applied physician-elicited death certificate bias parameters to quantitative bias analyses assessing potential age-related differential misclassification in US opioid-related mortality estimates., Methods: We obtained cause-of-death data (US, 2017) from the National Center for Health Statistics and calculated crude opioid-related outpatient death counts by age category (25-54, 55-64, 65+). We elicited beliefs from 10 primary care physicians on sensitivity of opioid-related death classification from death certificates. We summarized elicited sensitivity estimates, calculated plausible specificity values, and applied resulting parameters in a probabilistic bias analysis., Results: Physicians estimated wide sensitivity ranges for classification of opioid-related mortality by death certificates, with lower estimated sensitivities among older age groups. Probabilistic bias analyses adjusting for physician-estimated misclassification indicated 3.1 times more (95% uncertainty interval: 1.2-23.5) opioid-related deaths than the observed death count in the 65+ age group. All age groups had substantial increases in bias-adjusted death counts., Conclusions: We developed and implemented a feasible method of eliciting physician expert opinion on bias parameters for sensitivity of a medical record-based death indicator and applied findings in quantitative bias analyses adjusting for differential misclassification. Our findings are consistent with the hypothesis that opioid-related mortality rates may be substantially underestimated, particularly among older adults, due to misclassification in cause-of-death data from death certificates., Competing Interests: The authors declare no financial arrangements that may represent possible conflicts of interest. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States Government. This manuscript has not previously been published and is not under consideration for publication by another journal. Portions of the included analyses were presented as part of Elizabeth Goldsmith’s epidemiology PhD dissertation defense in December 2021. The written dissertation will be released by ProQuest in December 2023 or later., (Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved.)

Background: The risk of pregnancy-related mortality in the United States has nearly doubled since 1990, with venous thromboembolism (VTE) accounting for approximately 10% of these deaths., Objectives: The objective of this study was to assess whether preexisting autoimmune disease is a risk factor for postpartum VTE., Methods: Using the MarketScan Commercial and Medicare Supplemental administrative databases, a retrospective cohort study analyzed whether postpartum persons with autoimmune disease had a higher risk of postpartum VTE incidence than postpartum persons without autoimmune disease. Using International Classification of Diseases codes, we identified 757,303 individuals of childbearing age who had a valid delivery date with at least 12 weeks of follow-up., Results: Individuals were, on average, 30.7 years old (SD, 5.4), and 3.7% ( N  = 27,997 of 757,303) of them had evidence of preexisting autoimmune disease. In covariate-adjusted models, postpartum persons with preexisting autoimmune disease had higher rates of postpartum VTE than postpartum persons without autoimmune disease (hazard ratio [HR], 1.33; 95% CI, 1.07-1.64). When analyzed by individual autoimmune disease, those with systemic lupus erythematosus (HR, 2.49; 95% CI, 1.47-4.21) and Crohn's disease (HR, 2.49; 95% CI, 1.34-4.64) were at an elevated risk of postpartum VTE compared with those without autoimmune disease., Conclusion: Autoimmune disease was associated with a higher rate of postpartum VTE, with evidence that the association was most pronounced among individuals with systemic lupus erythematosus and Crohn's disease. These findings suggest that postpartum persons of childbearing age with autoimmune disease may require more monitoring and prophylactic care after delivery to prevent potentially fatal VTE events., (© 2023 The Authors.)

Background: Binge drinking can result in various types of harms including traffic crashes. Bars and restaurants that serve alcohol to patrons who are obviously intoxicated (i.e., overservice) contribute to these crashes. One strategy to address overservice is place of last drink (POLD) where law enforcement officers responding to alcohol-related incidents inquire about where the individuals last drank alcohol. This information may then be used to identify bars and restaurants that frequently overserve alcohol. There is limited evaluation of the effectiveness of POLD in reducing overservice, traffic crashes, and other harms., Methods: We evaluated the effects of a POLD initiative, developed by some law enforcement agencies in Minnesota (USA), on alcohol-related traffic crashes from 2010 to 2019. Among 89 intervention (POLD) vs. comparison communities, we fit regression models with participation in POLD as the predictor. As secondary analyses, we fit models with POLD implementation level as the predictor (implementation levels were assessed via a survey of law enforcement agencies). We controlled for relevant community and agency characteristics., Results: In the model with participation in POLD as a predictor, there was little difference in the rate of total alcohol-related crashes (rate ratio [RR] = 1.07, 95% CI: 0.85-1.34). In the model with level of implementation as a predictor, the rate of total alcohol-related crashes was comparable between communities with high implementation and those with no implementation (RR = 0.89; 95% CI: 0.71-1.10). Similar results were seen for alcohol-related crashes with nonfatal injury and property damage outcomes., Conclusions: This study found little evidence that the POLD initiative, as currently implemented, was associated with reductions in traffic crashes across communities in Minnesota. Further research could explore whether specific characteristics of POLD are particularly important and whether POLD could be combined with other strategies to reduce traffic crashes and other alcohol-related harms., (© 2022 Research Society on Alcohol.)

Retention in care is a major challenge for global AIDS control, including sub-Saharan Africa. In a large Ethiopian region, we evaluated an intervention where HIV positive community support workers (CSWs) provided HIV health education, personal counseling and social support for HIV patients new to care. We enrolled 1,799 patients recently entering care from 32 hospitals and health centers, randomized to intervention or control sites. Dates of all clinic visits, plus deaths or transfers were abstracted from HIV medical records. Primary outcomes were gap in clinical care (>90 days from a missed clinical or drug pickup appointment) and death. For 36 months of follow-up, and for the first 12 months after enrollment, weighted risk differences [RD] between treatment arms were modest and non-significant for gap in clinical care, death or either outcome. Through 36 months, 624 of 980 controls and 469 of 819 intervention participants had gaps in clinical care (RD = -5.5%, 95% confidence interval [CI] = -17.9%, 7.0%); 79 controls and 82 intervention participants died (RD = 2.5% 95% CI = -1.7%, 6.8%). Factors including HIV stigma and a volatile political climate may have attenuated the advantages we anticipated, demonstrating how benefits of CSW interventions may depend upon psychosocial, clinical and structural factors particular to specific community settings.

Background: Previous research has shown an association between individual thunderstorm events in the presence of high pollen, commonly called thunderstorm asthma, and acute severe asthma events, but little work has studied risk over long periods of time, using detailed measurements of storms and pollen., Methods: We estimated change in the risk of asthma-related emergency room visits related to thunderstorm asthma events in the Minneapolis-St. Paul metropolitan area over the years 2007-2018. We defined thunderstorm asthma events as daily occurrence of two or more lightning strikes during high pollen periods interpolating weather and pollen monitor data and modeling lightning counts. We acquired daily counts of asthma-related emergency department visits from the Minnesota Hospital Association and used a quasi-Poisson time-series regression to estimate overall relative risk of emergency department visits during thunderstorm asthma events., Results: We observed a 1.047 times higher risk (95% confidence interval = 1.012, 1.083) of asthma-related emergency department visits on the day of thunderstorm asthma event. Our findings are robust to adjustment for temperature, humidity, wind, precipitation, ozone, PM 2.5 , day of week, and seasonal variation in asthma cases. Occurrence of lightning alone or pollen alone showed no association with the risk of severe asthma. A two-stage analysis combining individual zip code-level results shows similar RR, and we see no evidence of spatial correlation or spatial heterogeneity of effect., Discussion: Our results support an association between co-occurrence of lightning and pollen and risk of severe asthma events. Our approach incorporates lightning and pollen data and small-spatial area exposure and outcome counts., Competing Interests: The authors report no conflicts of interest., (Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.)

Background Pulmonary hypertension (PH) is a devastating potential complication of pulmonary embolism, a manifestation of venous thromboembolism (VTE). The incidence of and risk factors for PH in those with prior VTE are poorly characterized. Methods and Results International Classification of Diseases ( ICD ) codes from inpatient and outpatient medical claims from MarketScan administrative databases for years 2011 to 2018 were used to identify cases of VTE, comorbidities before the VTE event, and PH occurring subsequent to the VTE event. Cumulative incidence and hazard ratios (HR), and their 95% CI, were calculated. The 170 021 VTE cases included in the analysis were on average (±SD) 57.5±15.8 years old and 50.5% were female. A total of 5943 PH cases accrued over an average follow-up of 1.94 years. Two years after incident VTE, the cumulative incidence (95% CI) of PH was 3.5% (3.4%-3.7%) overall. It was higher among older individuals, among women (3.9% [3.8%-4.1%]) than men (3.2% [3.0%-3.3%]), and among patients presenting with pulmonary embolism (6.2% [6.0%-6.5%]) than those presenting with deep vein thrombosis only (1.1% [1.0%-1.2%]). Adjusting for age and sex, risk of PH was higher among patients with VTE with underlying comorbidities. Using the Charlson comorbidity index, there was a dose-response relationship, whereby greater scores were associated with increased PH risk (score ≥5 versus 0: HR, (2.50 [2.30-2.71])). When evaluating individual comorbidities, the strongest associations were observed with concomitant heart failure (HR, 2.17 [2.04-2.31]), chronic pulmonary disease (2.01 [1.90-2.14]), and alcohol abuse (1.66 [1.29-2.13]). Conclusions In this large, real-world population of insured people with VTE, 3.5% developed PH in the 2 years following their initial VTE event. Risk was higher among women, with increasing age, and in those with additional comorbidities at the time of the VTE event. These data provide insights into the burden of PH and risk factors for PH among patients with VTE.

Although dementia and atrial fibrillation (AF) are common in older adults, risk factors for dementia have not been sufficiently characterized in patients with AF. We studied 621,773 patients with AF without dementia at the time of AF diagnosis who were enrolled in the MarketScan Commercial and Medicare Supplemental databases from 2007 to 2015. Dementia incidence and presence of predictors at the time of AF diagnosis (cardiometabolic conditions, mental and neurologic disorders, and other chronic conditions) were based on International Classification of Diseases, Ninth Revision, Clinical Modification codes in outpatient and inpatient claims, whereas medication usage was based on outpatient pharmacy claims. A frailty score was calculated using a previously established algorithm. The associations between the predictors of interest and dementia were assessed with multivariable Cox models. Patients had a mean age of 68 years (SD 14 years) and 41% were women. During a mean follow-up of 2.0 years, there were 16,073 cases of dementia. The strongest predictors of dementia were frailty (hazard ratio [HR] 1.43, 95% confidence interval [CI] 1.40 to 1.45, per 1-SD increase in the score), cognitive impairment (HR 1.50, 95% CI 1.36 to 1.65), mood disorders (HR 1.49, 95% CI 1.32 to 1.70), schizophrenia (HR 1.86, 95% CI 1.75 to 1.98), and substance abuse (HR 1.58, 95% CI 1.39 to 1.80). Among cardiometabolic conditions, only stroke (HR 1.17, 95% CI 1.13 to 1.22) and diabetes mellitus (HR 1.14, 95% CI 1.11 to 1.18) were associated with small increases in dementia risk after adjusting for demographics, frailty, co-morbidities, and medications. We have identified several risk factors for dementia in patients with AF., (Copyright © 2022 Elsevier Inc. All rights reserved.)

Data collected from a validation substudy permit calculation of a bias-adjusted estimate of effect that is expected to equal the estimate that would have been observed had the gold standard measurement been available for the entire study population. In this paper, we develop and apply a framework for adaptive validation to determine when sufficient validation data have been collected to yield a bias-adjusted effect estimate with a prespecified level of precision. Prespecified levels of precision are decided a priori by the investigator, based on the precision of the conventional estimate and allowing for wider confidence intervals that would still be substantively meaningful. We further present an applied example of the use of this method to address exposure misclassification in a study of transmasculine/transfeminine youth and self-harm. Our method provides a novel approach to effective and efficient estimation of classification parameters as validation data accrue, with emphasis on the precision of the bias-adjusted estimate. This method can be applied within the context of any parent epidemiologic study design in which validation data will be collected and modified to meet alternative criteria given specific study or validation study objectives., (© The Author(s) 2022. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.)

The goal of this study was to reconstruct the historical respirable silica (RS) and respirable dust (RD) exposures of workers in the Minnesota taconite industry from 1955 to 2010 as part of several epidemiological studies for assessing the association between exposure to components of taconite dusts and the development of respiratory diseases. A job-exposure matrix (JEM) was developed that uses 9127 RS and 19 391 RD occupational hygiene historical measurements. Historical RS and RD data were extracted from several sources and were grouped into seven mines and then into eight departments [Concentrating, Crushing, Janitor, Mining, Office/control room, Pelletizing, Shop (mobile), and Shop (stationary)]. Within each department, we applied a two-level random-intercept regression model which assumes that the natural log of Y (RD or RS concentration) changes over time at a constant rate. Among all predicted RD and RS values, we found that larger RD values were located in the following departments: Crushing, Concentrating, Pelletizing, and Shop (mobile). Larger RS values were located only in either Crushing or Shop (mobile). The annual rates of change for historical RD and RS exposures were between -3.3 and 3.2%. The silica percentage in the dust varied by mine/department with the highest value of 29.3% in Mine F (Crushing) and the lowest value of 2.1% in Mine B (Pelletizing). The predicted historical RD and RS arithmetic mean exposures ranged between <0.075 and 3.14 mg m-3, and between <0.005 and 0.36 mg m-3, respectively. The result of this study is a JEM by mine, department, and year for RD and RS for epidemiological studies., (© The Author(s) 2021. Published by Oxford University Press on behalf of the British Occupational Hygiene Society.)

Objective: Recently Doi et al. argued that risk ratios should be replaced with odds ratios in clinical research. We disagreed, and empirically documented the lack of portability of odds ratios, while Doi et al. defended their position. In this response we highlight important errors in their position., Study Design and Setting: We counter Doi et al.'s arguments by further examining the correlations of odds ratios, and risk ratios, with baseline risks in 20,198 meta-analyses from the Cochrane Database of Systematic Reviews., Results: Doi et al.'s claim that odds ratios are portable is invalid because 1) their reasoning is circular: they assume a model under which the odds ratio is constant and show that under such a model the odds ratio is portable; 2) the method they advocate to convert odds ratios to risk ratios is biased; 3) their empirical example is readily-refuted by counter-examples of meta-analyses in which the risk ratio is portable but the odds ratio isn't; and 4) they fail to consider the causal determinants of meta-analytic inclusion criteria: Doi et al. mistakenly claim that variation in odds ratios with different baseline risks in meta-analyses is due to collider bias. Empirical comparison between the correlations of odds ratios, and risk ratios, with baseline risks show that the portability of odds ratios and risk ratios varies across settings., Conclusion: The suggestion to replace risk ratios with odds ratios is based on circular reasoning and a confusion of mathematical and empirical results. It is especially misleading for meta-analyses and clinical guidance. Neither the odds ratio nor the risk ratio is universally portable. To address this lack of portability, we reinforce our suggestion to report variation in effect measures conditioning on modifying factors such as baseline risk; understanding such variation is essential to patient-centered practice., (Copyright © 2021 Elsevier Inc. All rights reserved.)

Objectives: A recent paper by Doi et al. advocated completely replacing the relative risk (RR) with the odds ratio (OR) as the effect measure in clinical trials and meta-analyses with binary outcomes. Besides some practical advantages of RR over OR, Doi et al.'s key assumption that the OR is "portable" in the meta-analysis, that is, study-specific ORs are likely not correlated with baseline risks, was not well justified., Study Designs and Settings: We summarized Spearman's rank correlation coefficient between study-specific ORs and baseline risks in 40,243 meta-analyses from the Cochrane Database of Systematic Reviews., Results: Study-specific ORs tend to be higher in studies with lower baseline risks of disease for most meta-analyses in Cochrane Database of Systematic Reviews. Using an actual meta-analysis example, we demonstrate that there is a strong negative correlation between OR (RR or RD) with the baseline risk and the conditional effects notably vary with baseline risks., Conclusions: Replacing RR or RD with OR is currently unadvisable in clinical trials and meta-analyses. It is possible that no effect measure is "portable" in a meta-analysis. In addition to the overall (or marginal) effect, we suggest presenting the conditional effect based on the baseline risk using a bivariate generalized linear mixed model., (Copyright © 2021. Published by Elsevier Inc.)

Background: Women with atrial fibrillation (AF) experience greater symptomatology, worse quality of life, and have a higher risk of stroke as compared to men, but are less likely to receive rhythm control treatment. Whether these differences exist in elderly patients with AF, and whether sex modifies the effectiveness of rhythm versus rate control therapy has not been assessed., Methods: We studied 135,850 men and 139,767 women aged ≥ 75 years diagnosed with AF in the MarketScan Medicare database between 2007 and 2015. Anticoagulant use was defined as use of warfarin or a direct oral anticoagulant. Rate control was defined as use of rate control medication or atrioventricular node ablation. Rhythm control was defined by use of anti-arrhythmic medication, catheter ablation or cardioversion. We used multivariable Poisson and Cox regression models to estimate the association of sex with treatment strategy and to determine whether the association of treatment strategy with adverse outcomes (bleeding, heart failure and stroke) differed by sex., Results: At the time of AF, women were on average (SD) 83.8 (5.6) years old and men 82.5 (5.2) years, respectively. Compared to men, women were less likely to receive an anticoagulant or rhythm control treatment. Rhythm control (vs. rate) was associated with a greater risk for heart failure with a significantly stronger association in women (HR women = 1.41, 95% CI 1.34-1.49; HR men = 1.21, 95% CI 1.15-1.28, p < 0.0001 for interaction). No sex differences were observed for the association of treatment strategy with the risk of bleeding or stroke., Conclusion: Sex differences exist in the treatment of AF among patients aged 75 years and older. Women are less likely to receive an anticoagulant and rhythm control treatment. Women were also at a greater risk of experiencing heart failure as compared to men, when treated with rhythm control strategies for AF. Efforts are needed to enhance use AF therapies among women. Future studies will need to delve into the mechanisms underlying these differences., (© 2021. The Author(s).)

Increased transparency in study design and analysis is one proposed solution to the perceived reproducibility crisis facing science. Systematic review and meta-analysis-through which individual studies on a specific association are ascertained, assessed for quality and quantitatively combined-is a critical process for building consensus in medical research. However, the conventional publication model creates static evidence summaries that force the quality assessment criteria and analytical choices of a small number of authors onto all stakeholders, some of whom will have different views on the quality assessment and key features of the analysis. This leads to discordant inferences from meta-analysis results and delayed arrival at consensus. We propose a shift to interactive meta-analysis, through which stakeholders can take control of the evidence synthesis using their own quality criteria and preferred analytic approach-including the option to incorporate prior information on the association in question-to reveal how their summary estimate differs from that reported by the original analysts. We demonstrate this concept using a web-based meta-analysis of the association between genetic variation in a key tamoxifen-metabolising enzyme and breast cancer recurrence in tamoxifen-treated women. We argue that interactive meta-analyses would speed consensus-building to the degree that they reveal invariance of inferences to different study selection and analysis criteria. On the other hand, when inferences are found to differ substantially as a function of these choices, the disparities highlight where future research resources should be invested to resolve lingering sources of disagreement., Competing Interests: Competing interests: None declared., (© Author(s) (or their employer(s)) 2021. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.)

Background: Quantitative bias analysis (QBA) measures study errors in terms of direction, magnitude and uncertainty. This systematic review aimed to describe how QBA has been applied in epidemiological research in 2006-19., Methods: We searched PubMed for English peer-reviewed studies applying QBA to real-data applications. We also included studies citing selected sources or which were identified in a previous QBA review in pharmacoepidemiology. For each study, we extracted the rationale, methodology, bias-adjusted results and interpretation and assessed factors associated with reproducibility., Results: Of the 238 studies, the majority were embedded within papers whose main inferences were drawn from conventional approaches as secondary (sensitivity) analyses to quantity-specific biases (52%) or to assess the extent of bias required to shift the point estimate to the null (25%); 10% were standalone papers. The most common approach was probabilistic (57%). Misclassification was modelled in 57%, uncontrolled confounder(s) in 40% and selection bias in 17%. Most did not consider multiple biases or correlations between errors. When specified, bias parameters came from the literature (48%) more often than internal validation studies (29%). The majority (60%) of analyses resulted in >10% change from the conventional point estimate; however, most investigators (63%) did not alter their original interpretation. Degree of reproducibility related to inclusion of code, formulas, sensitivity analyses and supplementary materials, as well as the QBA rationale., Conclusions: QBA applications were rare though increased over time. Future investigators should reference good practices and include details to promote transparency and to serve as a reference for other researchers., (© The Author(s) 2021; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association.)

Background Acute outpatient management of venous thromboembolism (VTE), which includes pulmonary embolism (PE) and deep vein thrombosis (DVT), is perceived to be as safe as inpatient management in some settings. How widely this strategy is used is not well documented. Methods and Results Using MarketScan administrative claims databases for years 2011 through 2018, we identified patients with International Classification of Diseases ( ICD ) codes indicating incident VTE and trends in the use of acute outpatient management. We also evaluated healthcare utilization and hospitalized bleeding events in the 6 months following the incident VTE event. A total of 200 346 patients with VTE were included, of whom 50% had evidence of PE. Acute outpatient management was used for 18% of those with PE and 57% of those with DVT only, and for both DVT and PE its use increased from 2011 to 2018. Outpatient management was less prevalent among patients with cancer, higher Charlson comorbidity index scores, and whose primary treatment was warfarin as compared with a direct oral anticoagulant. Healthcare utilization in the 6 months following the incident VTE event was generally lower among patients managed acutely as outpatients, regardless of initial presentation. Acute outpatient management was associated with lower hazard ratios of incident bleeding risk for both patients who initially presented with PE (0.71 [95% CI, 0.61, 0.82]) and DVT only (0.59 [95% CI, 0.54, 0.64]). Conclusions Outpatient management of VTE is increasing. In the present analysis, it was associated with lower subsequent healthcare utilization and fewer bleeding events. However, this may be because healthier patients were managed on an outpatient basis.

Background Current scores for bleeding risk assessment in patients with venous thromboembolism (VTE) undergoing oral anticoagulation have limited predictive capacity. We developed and internally validated a bleeding prediction model using healthcare claims data. Methods and Results We selected patients with incident VTE initiating oral anticoagulation in the 2011 to 2017 MarketScan databases. Hospitalized bleeding events were identified using validated algorithms in the 180 days after VTE diagnosis. We evaluated demographic factors, comorbidities, and medication use before oral anticoagulation initiation as potential predictors of bleeding using stepwise selection of variables in Cox models run on 1000 bootstrap samples of the patient population. Variables included in >60% of all models were selected for the final analysis. We internally validated the model using bootstrapping and correcting for optimism. We included 165 434 patients with VTE and initiating oral anticoagulation, of whom 2294 had a bleeding event. After undergoing the variable selection process, the final model included 20 terms (15 main effects and 5 interactions). The c-statistic for the final model was 0.68 (95% CI, 0.67-0.69). The internally validated c-statistic corrected for optimism was 0.68 (95% CI, 0.67-0.69). For comparison, the c-statistic of the Hypertension, Abnormal Renal/Liver Function, Stroke, Bleeding History or Predisposition, Labile International Normalized Ratio, Elderly (>65 Years), Drugs/Alcohol Concomitantly (HAS-BLED) score in this population was 0.62 (95% CI, 0.61-0.63). Conclusions We have developed a novel model for bleeding prediction in VTE using large healthcare claims databases. Performance of the model was moderately good, highlighting the urgent need to identify better predictors of bleeding to inform treatment decisions.

Prior to the pandemic, our research team implemented a randomized controlled trial of an intervention to reduce risk for alcohol-exposed pregnancy (AEP) in American Indian women. When active recruitment for the in-person trial was paused due to COVID, the research team moved to conducting follow-up surveys with participants who had completed the intervention to better understand changes to their alcohol use during the pandemic. We collected surveys from 62 American Indian women who had completed the Native CHOICES intervention. Baseline data collected pre-COVID included demographics and scores on the Alcohol Use Disorders Identification Test (AUDIT). Follow-up surveys conducted during the active pandemic period included a self-reported questionnaire about changes in drinking patterns. At pre-COVID baseline, all participants were engaged in heavy or binge drinking. At follow-up during COVID, 24.2% reported drinking more, and over half had at least one binge drinking episode. Approximately half reported reduced drinking. We found that risky drinking remained an issue during the pandemic for many American Indian women who had engaged in this behavior pre-COVID, while others reported reducing their alcohol consumption. As the pandemic abates, concerted efforts must be made to reach those with identified alcohol use disorders to offer resources and intervention as needed.

Reports from Sub-Saharan Africa, with a large HIV-infected population, vary widely in how often HIV status is disclosed to others, including spouses and other partners. We surveyed 1799 Ethiopian HIV patients newly enrolled in care within the previous 3 months at one of 32 local hospitals and health centers about disclosure of HIV status and two perceived social support domains: emotional/informational (EI) and tangible assistance (TA) support. Disclosure to another person was reported by 1389 (77%) persons. Disclosure rates to specific persons were: spouses or other partners = 74%; mothers = 24%; fathers = 16%; children = 26%; other family members = 37%; friends = 19%, and neighbors/other community members = 13%. Disclosure to another person was associated with higher social support scores on both EI and TA domains, marriage, and a longer time knowing HIV status. In multivariate adjusted models, disclosure to any person, as well as disclosure specifically to a spouse or partner, were associated with higher EI and higher TA social support scores. Provision of knowledgeable and emotionally supportive assistance can be an important factor in facilitating HIV disclosure. Helping persons with HIV decide who to disclose to and how to do so in the most positive manner is an essential component of HIV care and support.

Quantitative bias analyses allow researchers to adjust for uncontrolled confounding, given specification of certain bias parameters. When researchers are concerned about unknown confounders, plausible values for these bias parameters will be difficult to specify. Ding and VanderWeele developed bounding factor and E-value approaches that require the user to specify only some of the bias parameters. We describe the mathematical meaning of bounding factors and E-values and the plausibility of these methods in an applied context. We encourage researchers to pay particular attention to the assumption made, when using E-values, that the prevalence of the uncontrolled confounder among the exposed is 100% (or, equivalently, the prevalence of the exposure among those without the confounder is 0%). We contrast methods that attempt to bound biases or effects and alternative approaches such as quantitative bias analysis. We provide an example where failure to make this distinction led to erroneous statements. If the primary concern in an analysis is with known but unmeasured potential confounders, then E-values are not needed and may be misleading. In cases where the concern is with unknown confounders, the E-value assumption of an extreme possible prevalence of the confounder limits its practical utility., (Copyright © 2021 The Author(s). Published by Wolters Kluwer Health, Inc.)

Background: Inconsistent evidence suggests that use of certain antidepressants, particularly selective serotonin reuptake inhibitors (SSRIs), in patients using oral anticoagulants (OACs) might be associated with an elevated risk of bleeding., Objective: This study aims to investigate the risk of bleeding associated with initiation of different types of antidepressants among atrial fibrillation (AF) patients on OAC therapy., Patients and Methods: A total of 30,336 AF patients (mean age 72.2 years; 54% female) on OAC therapy that started antidepressant treatment were identified from the Truven Health Analytics MarketScan Commercial and Medicare Databases for the period 2007-2015. Exposure was defined as filling a prescription for antidepressant, and categorized as SSRI, serotonin/norepinephrine reuptake inhibitors (SNRIs), serotonin reuptake inhibitors (SRIs), tricyclic antidepressants (TCAs), or other antidepressants. The primary outcome was incident hospitalized bleeding. Associations of antidepressant type with bleeding were assessed calculating hazard ratios (HRs) and 95% confidence intervals (CIs) with adjusted Cox models in pairwise propensity score-matched cohorts., Results: During a mean follow-up of 21 months, we identified 1612 bleeding episodes. In pairwise comparisons, SSRI use was associated with an increased risk of bleeding when compared to most other antidepressants (HR 1.22, 95% CI 0.96-1.54 vs SNRI; HR 1.10, 95% CI 0.90-1.35 vs SRI; HR 1.03, 95% CI 0.82-1.30 vs TCA). SNRI use was associated with the lowest bleeding risk. Results did not differ by OAC type, age, and sex., Conclusions: Among AF patients on OAC initiating antidepressants, risk of bleeding varied across antidepressant type. This information can inform treatment choices among patients receiving OAC., (© 2021. The Author(s).)

Background: Venous thromboembolism (VTE) affects nearly 1 million Americans annually, and many benefit from continued anticoagulation after the initial 3- to 6-month treatment period (secondary prevention)., Objectives: To determine whether warfarin, apixaban, or rivaroxaban is associated with reduced recurrent VTE hospitalization in the secondary prevention of VTE., Patients/methods: We performed a retrospective cohort study of participants enrolled in the MarketScan Insurance Database between 2013 and 2017 in those with an incident VTE. In those individuals who continued oral anticoagulation (warfarin, apixaban, or rivaroxaban) beyond 6 months, we determined the relative rate of recurrent VTE hospitalization., Results: Among 119 964 individuals with VTE, 25 419 remained on anticoagulation after 6 months and were matched successfully by age, sex, and date. After adjusting for a propensity score, apixaban versus rivaroxaban (hazard ratio [HR], 0.65; 95% confidence interval [CI], 0.45-0.94) and apixaban versus warfarin (HR, 0.68; 95% CI, 0.47-1.00) had a reduced risk of recurrent VTE hospitalization, and rivaroxaban versus warfarin (HR, 1.12; 95% CI, 0.94-1.33) had equivalent rates. For the rivaroxaban versus warfarin comparison there was a significant interaction by renal function ( P  < .01) where rivaroxaban was associated with a lower risk of recurrent VTE hospitalization (HR, 0.65; 95% CI, 0.41-1.03) in those with kidney disease and increased risk in those without kidney disease (HR, 1.24; 95% CI, 1.02-1.50)., Conclusions: These data suggest that apixaban has a lower recurrent VTE hospitalization rate than rivaroxaban during the secondary prevention of VTE, and further study of diverse patient populations, especially by kidney function, is warranted., (© 2021 The Authors. Research and Practice in Thrombosis and Haemostasis published by Wiley Periodicals LLC on behalf of International Society on Thrombosis and Haemostasis (ISTH).)

Quantitative bias analysis comprises the tools used to estimate the direction, magnitude, and uncertainty from systematic errors affecting epidemiologic research. Despite the availability of methods and tools, and guidance for good practices, few reports of epidemiologic research incorporate quantitative estimates of bias impacts. The lack of familiarity with bias analysis allows for the possibility of misuse, which is likely most often unintentional but could occasionally include intentional efforts to mislead. We identified 3 examples of suboptimal bias analysis, one for each common bias. For each, we describe the original research and its bias analysis, compare the bias analysis with good practices, and describe how the bias analysis and research findings might have been improved. We assert no motive to the suboptimal bias analysis by the original authors. Common shortcomings in the examples were lack of a clear bias model, computed example, and computing code; poor selection of the values assigned to the bias model's parameters; and little effort to understand the range of uncertainty associated with the bias. Until bias analysis becomes more common, community expectations for the presentation, explanation, and interpretation of bias analyses will remain unstable. Attention to good practices should improve quality, avoid errors, and discourage manipulation., (© The Author(s) 2021. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.)