Validation of long-term recall of pregnancy-related weight in the Life-course Experiences And Pregnancy (LEAP) study.

BACKGROUND: Pre-pregnancy 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 pre-pregnancy 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 pre-pregnancy BMI or GWG category conditional on one's true category (analogous to sensitivities and specificities), and probabilities of truly being in each pre-pregnancy 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 pre-pregnancy BMI. Self-report misclassified 32% (95% 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. CONCLUSIONS: Findings can be used in quantitative bias analyses to estimate bias-adjusted associations with pre-pregnancy BMI and GWG.

Food insecurity in urban American Indian and Alaska Native populations during the COVID-19 pandemic.

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.

Venous thromboembolism mortality and trends in older US adults, 2011-2019.

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.

SAS and R code for probabilistic quantitative bias analysis for misclassified binary variables and binary unmeasured confounders.

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.

Autoimmune disease and risk of postpartum venous thromboembolism.

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.

Opioid-related Mortality in United States Death Certificate Data: A Quantitative Bias Analysis With Expert Elicitation of Bias Parameters.

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.

Impact of Community Support Workers in Rural Ethiopia on Emotional and Psychosocial Health of Persons Living with HIV: Results of a Three-Year Randomized Community Trial.

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).

Prevalence of Dementia in American Indians and Alaska Natives Compared to White, Black, and Hispanic Medicare Beneficiaries: Findings from the National Health and Aging Trends Study.

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.

Place of last drink enforcement: Effects on alcohol-related traffic crashes.

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.

Food Insecurity in Urban American Indian and Alaska Native Populations During the COVID-19 Pandemic.

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.

Incidence and Risk Factors of Pulmonary Hypertension After Venous Thromboembolism: An Analysis of a Large Health Care Database.

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.

Thunderstorms, Pollen, and Severe Asthma in a Midwestern, USA, Urban Environment, 2007-2018.

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.

Risk Factors for Dementia in Patients With Atrial Fibrillation.

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.

Sampling Validation Data to Achieve a Planned Precision of the Bias-Adjusted Estimate of Effect.

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.

A three-year randomized community trial of community support workers in rural Ethiopia to promote retention in HIV care.

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.

Reconstructing Historical Exposures to Respirable Dust and Respirable Silica in the Taconite Mining Industry for 1955-2010.

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.

Controversy and Debate: Questionable utility of the relative risk in clinical research: Paper 2: Is the Odds Ratio "portable" in meta-analysis? Time to consider bivariate generalized linear mixed model.

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.

Controversy and Debate : Questionable utility of the relative risk in clinical research: Paper 4 :Odds Ratios are far from "portable" - A call to use realistic models for effect variation in meta-analysis.

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.

Sex differences in treatment strategy and adverse outcomes among patients 75 and older with atrial fibrillation in the MarketScan database.

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.

Inpatient Versus Outpatient Acute Venous Thromboembolism Management: Trends and Postacute Healthcare Utilization From 2011 to 2018.

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.