Real-World Risk of Severe Cytopenias in Multiple Myeloma Patients Sequentially Treated with Immunomodulatory Drugs.

INTRODUCTION: Most multiple myeloma (MM) patients experience cytopenias, likely driven by both disease and treatment-related factors. Immunomodulatory agents (IMiDs), which form the backbone of most anti-myeloma regimens, are known to cause higher grade cytopenias. In this context, the impact of sequential IMiD treatments on cytopenia risk is unknown. METHODS: We evaluated the cumulative risks of severe cytopenias following second line of therapy (LOT) initiation in 5573 MM patients in the Flatiron Health database. Patients for whom both LOTs 1 and 2 contained IMiDs were considered "sequentially exposed"; those for whom neither contained IMiDs were "never exposed." RESULTS: For the neutropenia outcome, compared to the never exposed, the sequentially exposed had the highest 1-year risk (risk difference [RD] 12%), followed by those only recently exposed during LOT 2 (RD 8%), then by those with only past exposure during LOT 1 (RD 5%). A similar pattern was observed for leukopenia, but no meaningful differences were observed for anemia or thrombocytopenia. The associations between sequential exposure, versus never, with neutropenia and leukopenia were even stronger among those with a recent cytopenia history. CONCLUSION: Results suggest that sequential exposure to IMiDs is a risk factor for higher grade cytopenias. These findings have profound clinical implications in choosing newer LOTs with potential risks of cytopenia.

Use of quantitative bias analysis to evaluate single-arm trials with real-world data external controls.

PURPOSE: Use of real-world data (RWD) for external controls added to single-arm trials (SAT) is increasingly prevalent in regulatory submissions. Due to inherent differences in the data-generating mechanisms, biases can arise. This paper aims to illustrate how to use quantitative bias analysis (QBA). METHODS: Advanced non-small cell lung cancer (NSCLC) serves as an example, where many small subsets of patients with molecular tumor subtypes exist. First, some sources of bias that may occur in oncology when comparing RWD to SAT are described. Second, using a hypothetical immunotherapy agent, a dataset is simulated based on expert input for survival analysis of advanced NSCLC. Finally, we illustrate the impact of three biases: missing confounder, misclassification of exposure, and outcome evaluation. RESULTS: For each simulated scenario, bias was induced by removing or adding data; hazard ratios (HRs) were estimated applying conventional analyses. Estimating the bias-adjusted treatment effect and uncertainty required carefully selecting the bias model and bias factors. Although the magnitude of each biased and bias-adjusted HR appeared moderate in all three hypothetical scenarios, the direction of bias was variable. CONCLUSION: These findings suggest that QBA can provide an intuitive framework for bias analysis, providing a key means of challenging assumptions about the evidence. However, the accuracy of bias analysis is itself dependent on correct specification of the bias model and bias factors. Ultimately, study design should reduce bias, but QBA allows us to evaluate the impact of unavoidable bias to assess the quality of the evidence.

Clinical factors associated with patterns of endocrine therapy adherence in premenopausal breast cancer patients.

INTRODUCTION: Patients with hormone receptor positive breast cancer are recommended at least five years of adjuvant endocrine therapy, but adherence to this treatment is often suboptimal. We investigated longitudinal trends in adjuvant endocrine therapy (AET) adherence among premenopausal breast cancer patients and identified clinical characteristics, including baseline comorbidities and non-cancer chronic medication use, associated with AET adherence. METHODS: We included stage I-III premenopausal breast cancer patients diagnosed during 2002-2011 and registered in the Danish Breast Cancer Group clinical database who initiated AET. We used group-based trajectory modeling to describe AET adherence patterns. We also linked patients to Danish population-based registries and fit multinomial logistic models to compute odds ratios (ORs) and 95% confidence intervals (95% CIs) associating clinical characteristics with AET adherence patterns. RESULTS: We identified three adherence patterns among 4,353 women-high adherers (57%), slow decliners (36%), and rapid decliners (6.9%). Women with stage I disease (vs. stage II; OR: 1.9, 95% CI 1.5, 2.5), without chemotherapy (vs. chemotherapy; OR: 4.3, 95% CI 3.0, 6.1), with prevalent comorbid disease (Charlson Comorbidity Index score ≥ 1 vs. 0; OR: 1.6, 95% CI 1.1, 2.3), and with a history of chronic non-cancer medication use (vs. none; OR: 1.3, 95% CI 1.0, 1.8) were more likely to be rapid decliners compared with high adherers. CONCLUSIONS: Women with stage I cancer, no chemotherapy, higher comorbidity burden, and history of chronic non-cancer medication use were less likely to adhere to AET. Taking steps to promote adherence in these groups of women may reduce their risk of recurrence.

Influence of Incomplete Death Information on Cumulative Risk Estimates in United States Claims Data.

Administrative claims databases often do not capture date or fact of death, so studies using these data may inappropriately treat death as a censoring event-equivalent to other withdrawal reasons-rather than a competing event. We examined 1-, 3-, and 5-year inverse-probability-of-treatment-weighted cumulative risks of a composite cardiovascular outcome among 34,527 initiators of telmisartan (exposure) and ramipril (referent) ages ≥55 in Optum claims from 2003 to 2020. Differences in cumulative risks of the cardiovascular endpoint due to censoring of death (cause-specific), as compared to treating death as a competing event (sub-distribution), increased with greater follow-up time and older age, where event and mortality risks were higher. Among ramipril users (selected results), 5-year cause-specific and sub-distribution cumulative risk estimates per 100, respectively, were 16.4 (95% CI 15.3, 17.5) and 16.2 (95% CI 15.1, 17.3) among ages 55-64 (difference=0.2) and were 43.2 (95% CI 41.3, 45.2) and 39.7 (95% CI 37.9, 41.4) among ages ≥75 (difference=3.6). Plasmode simulation results demonstrated the differences in cause-specific versus sub-distribution cumulative risks to increase with increasing mortality rate. We suggest researchers consider the cohort's baseline mortality risk when deciding whether real-world data with incomplete death information can be used without concern.

Staging and clean room: Constructs designed to facilitate transparency and reduce bias in comparative analyses of real-world data.

PURPOSE: We describe constructs designed to protect the integrity of the results from comparative analyses using real-world data (RWD): staging and clean room. METHODS: Staging involves performing sequential preliminary analyses and evaluating the population size available and potential bias before conducting comparative analyses. A clean room involves restricted access to data and preliminary results, policies governing exploratory analyses and protocol deviations, and audit trail. These constructs are intended to allow decisions about protocol deviations, such as changes to design or model specification, to be made without knowledge of how they might affect subsequent analyses. We describe an example for implementing staging with a clean room. RESULTS: Stage 1 may involve selecting a data source, developing and registering a protocol, establishing a clean room, and applying inclusion/exclusion criteria. Stage 2 may involve attempting to achieve covariate balance, often through propensity score models. Stage 3 may involve evaluating the presence of residual confounding using negative control outcomes. After each stage, check points may be implemented when a team of statisticians, epidemiologists and clinicians masked to how their decisions may affect study outcomes, reviews the results. This review team may be tasked with making recommendations for protocol deviations to address study precision or bias. They may recommend proceeding to the next stage, conducting additional analyses to address bias, or terminating the study. Stage 4 may involve conducting the comparative analyses. CONCLUSIONS: The staging and clean room constructs are intended to protect the integrity and enhance confidence in the results of analyses of RWD.

Sex Differences in Psychopathology Following Potentially Traumatic Experiences.

Importance: Various psychopathology may follow trauma; however, sex differences in these ranging manifestations of posttraumatic psychopathology remain understudied. Objective: To investigate sex-specific incidence of posttraumatic psychopathology. Design, Setting, and Participants: This population-based cohort study of Danish national health registries included a cohort of individuals who experienced a potentially traumatic event (PTE) from 1994 to 2016. Individuals were further categorized by presence of any pretrauma psychopathology. A comparison group of individuals who experienced a nontraumatic stressor (nonsuicide death of a first-degree relative) was examined as a reference cohort. Exposures: At least 1 of 8 PTEs (eg, physical assault, transportation accident) derived through health registry International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) codes, with additional qualifiers to improve classification accuracy. Main Outcomes and Measures: Incidence of 9 categories of ICD-10 psychiatric disorders recorded in registries within 5 years of PTEs. The standardized morbidity ratios (SMRs) for psychopathology outcomes were also calculated to compare individuals experiencing PTEs with those experiencing a nontraumatic stressor. Results: This study included 1 398 026 individuals who had been exposed to trauma (475 280 males [34.0%]; 922 750 females [66.0%]). The group of males who had been exposed to trauma were evenly distributed across age, while most females in the trauma-exposed group were aged 16 to 39 years (592 385 [64.2%]). Males and females were equally distributed across income quartiles and predominantly single. Following PTEs, the most common diagnosis was substance use disorders for males (35 160 [7.4%]) and depressive disorders for females (29 255 [3.2%]); incidence proportions for these and other disorders were higher among males and females with any pretrauma psychopathology. Certain PTEs had elevated onset of various psychiatric disorders and some sex differences emerged. Following physical assault, associations were found with schizophrenia or psychotic disorders for males (SMR, 17.5; 95% CI, 15.9-19.3) and adult personality disorders for females (SMR, 16.3; 95% CI, 14.6-18.3). For noninterpersonal PTEs, males had larger SMRs for substance use, schizophrenia or psychotic disorders, and adult personality disorders (SMR, 43.4; 95% CI, 41.9-45.0), and females had larger SMRs for depressive disorders (SMR, 19.0; 95% CI, 18.6-19.4). Sex differences were also observed, particularly when considering pretrauma psychopathology. For example, among interpersonal PTEs, males were most likely to develop substance use disorders after physical assault, whereas females were more likely to develop various disorders, with stronger associations seen for females without pretrauma psychiatric diagnoses. Among noninterpersonal PTEs, exposure to toxic substance showed robust associations with psychopathology, particularly in those without pretrauma psychopathology, with sex-specific differences across psychiatric categories. Conclusions and Relevance: Mental disorders after trauma were wide-ranging for males and females, and sex differences in patterns of posttraumatic psychopathology were more pronounced when accounting for pretrauma psychopathology. Findings provide new insights for sex-relevant PTEs and their mental health consequences. It also outlines future directions for advancing understanding of a constellation of posttraumatic psychopathology in males and females.

Characterizing Fit-for-Purpose Real-World Data: An Assessment of a Mother-Infant Linkage in the Japan Medical Data Center Claims Database.

Purpose: Observational postapproval safety studies are needed to inform medication safety during pregnancy. Real-world databases can be valuable for supporting such research, but fitness for regulatory purpose must first be vetted. Here, we demonstrate a fit-for-purpose assessment of the Japan Medical Data Center (JMDC) claims database for pregnancy safety regulatory decision-making. Patients and Methods: The Duke-Margolis framework considers a database's fitness for regulatory purpose based on relevancy (capacity to answer the research question based on variable availability and a sufficiently sized, representative population) and quality (ability to validly answer the research question based on data completeness and accuracy). To assess these considerations, we examined descriptive characteristics of infants and pregnancies among females ages 12-55 years in the JMDC between January 2005 and March 2022. Results: For relevancy, we determined that critical data fields (maternal medications, infant major congenital malformations, covariates) are available. Family identification codes permitted linkage of 385,295 total mother-infant pairs, 57% of which were continuously enrolled during pregnancy. The prevalence of specific congenital malformation subcategories and maternal medical conditions were representative of the general population, but preterm births were below expectations (3.6% versus 5.6%) in this population. For quality, our methods are expected to accurately identify the complete set of mothers and infants with a shared health insurance plan. However, validity of gestational age information was limited given the high proportion (60%) of missing live birth delivery codes coupled with suppression of infant birth dates and inaccessibility of disease codes with gestational week information. Conclusion: The JMDC may be well suited for descriptive studies of pregnant people in Japan (eg, comorbidities, medication usage). More work is needed to identify a method to assign pregnancy onset and delivery dates so that in utero medication exposure windows can be defined more precisely as needed for many regulatory postapproval pregnancy safety studies.

Biomarkers predictive of a response to extended endocrine therapy in breast cancer: a systematic review and meta-analysis.

PURPOSE: Extension of adjuvant endocrine therapy beyond five years confers only modest survival benefit in breast cancer patients and carries risk of toxicities. This systematic review investigates the role of biomarker tests in predicting the clinical response to an extension of endocrine therapy. METHODS: We searched Ovid MEDLINE, Ovid Embase, Global Index Medicus, and the Cochrane Central Register of Controlled Trials using an iterative approach to identify full-text articles related to breast cancer, endocrine therapy, and biomarkers. RESULTS: Of the 1,217 unique reports identified, five studies were deemed eligible. Four investigated the Breast Cancer Index (BCI) assay in three distinct study populations. These studies consistently showed that BCI score was predictive of response to extended endocrine therapy among 1,946 combined patients, who were predominately non-Hispanic white and postmenopausal. CONCLUSIONS: Evidence in the setting of predictive tests for extended endocrine therapy is sparse. Most relevant studies investigated the use of BCI, but these study populations were largely restricted to a single age, race, and ethnicity group. Future studies should evaluate a variety of biomarkers in diverse populations. Without sufficient evidence, physicians and patients face a difficult decision in balancing the benefits and risks of endocrine therapy extension.

Prediction of Suicide Attempts Among Persons with Depression: A Population-Based Case Cohort Study.

Studies have highlighted the potential importance of modeling interactions for suicide attempt prediction. This case-cohort study identified risk factors for suicide attempts among persons with depression in Denmark using statistical approaches that do (random forests) or do not model interactions (least absolute shrinkage and selection operator regression [LASSO]). Cases made a non-fatal suicide attempt (n = 6,032) between 1995 and 2015. The comparison subcohort was a 5% random sample of all persons in Denmark on January 1, 1995 (n = 11,963). We used random forests and LASSO for sex-stratified prediction of suicide attempts from demographic variables, psychiatric and somatic diagnoses, and treatments. Poisonings, psychiatric disorders, and medications were important predictors for both sexes. Area under the receiver operating characteristic curve (AUC) values were higher in LASSO models (0.85 [95% CI = 0.84, 0.86] in men; 0.89 [95% CI = 0.88, 0.90] in women) than random forests (0.76 [95% CI = 0.74, 0.78] in men; 0.79 [95% CI = 0.78, 0.81] in women). Automatic detection of interactions via random forests did not result in better model performance than LASSO models that did not model interactions. Due to the complex nature of psychiatric comorbidity and suicide, modeling interactions may not always be the optimal statistical approach to enhancing suicide attempt prediction in high-risk samples.

Census Tracts Are Not Neighborhoods: Addressing Spatial Misalignment in Studies Examining the Impact of Historical Redlining on Present-day Health Outcomes.

BACKGROUND: Research examining the effects of historical redlining on present-day health outcomes is often complicated by the misalignment of contemporary census boundaries with the neighborhood boundaries drawn by the US Home Owners' Loan Corporation (HOLC) in the 1930s. Previous studies have used different approaches to assign historical HOLC grades to contemporary geographies, but how well they capture redlining exposure is unknown. METHODS: Our analysis included 7711 residences identified in the Multiple Listing Service database in Atlanta, Georgia (2017-2022). We evaluated the classification of HOLC grade assignment (A, B, C, D, or ungraded) when assigning exposure under four area-level approaches (centroid, majority land area, weighted score, and highest HOLC) compared with using complete address data (gold standard). We additionally compared approaches across three 2020 census geographies (tract, block group, and block). RESULTS: When comparing the use of census tracts to complete address data, sensitivity was highest for the weighted score approach, which correctly identified 77% of residences in truly A-D graded neighborhoods as compared with the majority land area (44%), centroid (54%), and highest HOLC (59%) approaches. Regarding specificity, the majority land area approach best-classified residences in truly ungraded neighborhoods (93%) as compared with the weighted score (65%), centroid (81%), and highest HOLC (54%) approaches. Classification improved regardless of approach when using census block compared with the census tract. CONCLUSIONS: Misclassification of historical redlining exposure is inevitable when using contemporary census geographies rather than complete address data. This study provides a framework for assessing spatial misalignment and selecting an approach for classification.

Antihypertensive Medications and PTSD Incidence in a Trauma Cohort.

Objective: Antihypertensive medications have been examined as agents for posttraumatic stress disorder (PTSD) prevention in trauma-exposed individuals, given well-documented associations between PTSD and increased risk of cardiovascular disease and purported trauma-relevant mechanisms of action for these medications. Evidence regarding the effectiveness of such drugs for this purpose remains mixed.Methods: We conducted a national population-based cohort study using data from Danish national registries to assess whether 4 classes of antihypertensive drugs (beta-adrenoceptor blockers [beta blockers], angiotensin II receptor blockers [ARBs], angiotensin-converting enzyme [ACE] inhibitors, and calcium channel blockers) were associated with a decreased incidence of PTSD (diagnosed according to ICD-10) over a 22-year study period. Data for this study originated from a population-based cohort of over 1.4 million persons who experienced a traumatic event between 1994 and 2016 in Denmark. We calculated the incidence rate of PTSD per 100,000 person-years among persons who filled a prescription for each class of drug in the 60 days prior to a traumatic event and for corresponding unexposed comparison groups. We then used Cox proportional hazards regression to compare the rate of PTSD among persons who filled an antihypertensive medication prescription within 60 days before their trauma to the rate among persons who did not.Results: We found evidence that calcium channel blockers were associated with a decreased incidence of PTSD (adjusted hazard ratio = 0.63, 95% confidence interval [CI] = 0.34, 1.2); all other antihypertensive medication classes had null or near null associations.Conclusions: These findings lay a foundation for additional research focusing on antihypertensive medications that appear most effective in reducing PTSD incidence following trauma and for additional replication work aimed at continuing to clarify the disparate findings reported in the literature to date.

Risk of recurrent cancer-associated venous thromboembolism: A Danish nationwide cohort study.

BACKGROUND: Predictive factors for recurrent cancer-associated venous thromboembolism have been inconsistent across previous studies. To provide data for improved risk stratification, we described the risk of recurrent venous thromboembolism overall and across age, sex, calendar period, cancer type, Ottawa risk score, cancer stage, and cancer treatment in a nationwide cohort of patients with active cancer. METHODS: Using Danish administrative registries, we identified a cohort of all adult patients with active cancer and a first-time diagnosis of venous thromboembolism during 2003-2018. We accounted for the competing risk of death and calculated absolute risks of recurrent venous thromboembolism at six months. RESULTS: The population included 34,072 patients with active cancer and venous thromboembolism. Recurrence risks at six months were higher for patients with genitourinary cancer (6.5%), lung cancer (6.1%), gastrointestinal cancer (5.6%), brain cancer (5.2%), and hematological cancer (5.1%) than for patients with gynecological cancer (4.7%), breast cancer (4.1%), and other cancer types (4.8%). Recurrence risks were similar for men (5.2%) and women (4.9%), with and without chemotherapy (5.1%), across Ottawa risk score group (low: 5.0%; high: 5.1%) and across calendar periods but increased with increasing cancer stage. The overall six-month all-cause mortality risk was 26%, and highest for patients with lung cancer (49%) and lowest among breast cancer patients (4.1%). CONCLUSIONS: Six-month recurrence risk after first-time cancer-associated venous thromboembolism was high and varied by cancer type and patient characteristics. Refining risk stratification for recurrence may improve decision-making regarding treatment duration after cancer-associated thromboembolism.

Association of Early-, Middle-, and Late-Life Depression With Incident Dementia in a Danish Cohort.

Importance: Late-life depressive symptoms are associated with subsequent dementia diagnosis and may be an early symptom or response to preclinical disease. Evaluating associations with early- and middle-life depression will help clarify whether depression influences dementia risk. Objective: To examine associations of early-, middle-, and late-life depression with incident dementia. Design, Setting, and Participants: This was a nationwide, population-based, cohort study conducted from April 2020 to March 2023. Participants included Danish citizens from the general population with depression diagnoses who were matched by sex and birth year to individuals with no depression diagnosis. Participants were followed up from 1977 to 2018. Excluded from analyses were individuals followed for less than 1 year, those younger than 18 years, or those with baseline dementia. Exposure: Depression was defined using diagnostic codes from the International Classification of Diseases (ICD) within the Danish National Patient Registry (DNPR) and Danish Psychiatric Central Research Register (DPCRR). Main Outcomes and Measure: Incident dementia was defined using ICD diagnostic codes within the DPCRR and DNPR. Cox proportional hazards regression was used to examine associations between depression and dementia adjusting for education, income, cardiovascular disease, chronic obstructive pulmonary disease, diabetes, anxiety disorders, stress disorders, substance use disorders, and bipolar disorder. Analyses were stratified by age at depression diagnosis, years since index date, and sex. Results: There were 246 499 individuals (median [IQR] age, 50.8 [34.7-70.7] years; 159 421 women [64.7%]) with diagnosed depression and 1 190 302 individuals (median [IQR] age, 50.4 [34.6-70.0] years; 768 876 women [64.6%]) without depression. Approximately two-thirds of those diagnosed with depression were diagnosed before the age of 60 years (684 974 [67.7%]). The hazard of dementia among those diagnosed with depression was 2.41 times that of the comparison cohort (95% CI, 2.35-2.47). This association persisted when the time elapsed from the index date was longer than 20 to 39 years (hazard ratio [HR], 1.79; 95% CI, 1.58-2.04) and among those diagnosed with depression in early, middle, or late life (18-44 years: HR, 3.08; 95% CI, 2.64-3.58; 45-59 years: HR, 2.95; 95% CI, 2.75-3.17; ≥60 years: HR, 2.31; 95% CI, 2.25-2.38). The overall HR was greater for men (HR, 2.98; 95% CI, 2.84-3.12) than for women (HR, 2.21; 95% CI, 2.15-2.27). Conclusions and Relevance: Results suggest that the risk of dementia was more than doubled for both men and women with diagnosed depression. The persistent association between dementia and depression diagnosed in early and middle life suggests that depression may increase dementia risk.

Environmental, social and behavioral risk factors in association with spatial clustering of childhood cancer incidence.

Childhood cancer incidence is known to vary by age, sex, and race/ethnicity, but evidence is limited regarding external risk factors. We aim to identify harmful combinations of air pollutants and other environmental and social risk factors in association with the incidence of childhood cancer based on 2003-2017 data from the Georgia Cancer Registry. We calculated the standardized incidence ratios (SIR) of Central Nervous System (CNS) tumors, leukemia and lymphomas based on age, gender and ethnic composition in each of the 159 counties in Georgia, USA. County-level information on air pollution, socioeconomic status (SES), tobacco smoking, alcohol drinking and obesity were derived from US EPA and other public data sources. We applied two unsupervised learning tools (self-organizing map [SOM] and exposure-continuum mapping [ECM]) to identify pertinent types of multi-exposure combinations. Spatial Bayesian Poisson models (Leroux-CAR) were fit with indicators for each multi-exposure category as exposure and SIR of childhood cancers as outcomes. We identified consistent associations of environmental (pesticide exposure) and social/behavioral stressors (low socioeconomic status, alcohol) with spatial clustering of pediatric cancer class II (lymphomas and reticuloendothelial neoplasms), but not for other cancer classes. More research is needed to identify the causal risk factors for these associations.

Impact of hemoglobin A1c level on the association between non-steroidal anti-inflammatory drug use and cardiovascular events in patients with type 2 diabetes: A population-based cohort study.

OBJECTIVE: Non-steroidal anti-inflammatory drugs (NSAIDs) should be used cautiously in patients with type 2 diabetes. We examined whether the cardiovascular risks associated with NSAID use depended on HbA1c level in patients with type 2 diabetes. METHODS: We conducted a population-based cohort study of all adult Danes with a first-time HbA1c measurement ≥48 mmol/mol during 2012-2020 (n = 103 308). We used information on sex, age, comorbidity burden, and drug use to calculate time-varying inverse probability of treatment weights. After applying these weights in a pooled logistic regression, we estimated hazard ratios (HRs) of the association between use of NSAIDs (ibuprofen, naproxen, or diclofenac) and cardiovascular events (a composite of myocardial infarction, ischemic stroke, congestive heart failure, atrial fibrillation or flutter, and all-cause death). We stratified all analyses by HbA1c level (<53 or ≥53 mmol/mol). RESULTS: For ibuprofen use, the HR of a cardiovascular event was 1.53 (95% confidence interval [CI]: 1.34-1.75) in patients with HbA1c <53 and 1.24 (95% CI: 1.00-1.53) in patients with HbA1c ≥53 mmol/mol. For naproxen use, the HR was 1.14 (95% CI: 0.59-2.21) in patients with HbA1c <53 and 1.30 (95% CI: 0.49-3.49) in patients with HbA1c ≥53 mmol/mol. For diclofenac use, the HR was 2.40 (95% CI: 1.62-3.56) in patients with HbA1c <53 and 2.89 (95% CI: 1.65-5.04) in patients with HbA1c ≥53 mmol/mol. CONCLUSIONS: In patients with type 2 diabetes, glycemic dysregulation did not affect the cardiovascular risk associated with NSAID use.

Tailoring capture-recapture methods to estimate registry-based case counts based on error-prone diagnostic signals.

Surveillance research is of great importance for effective and efficient epidemiological monitoring of case counts and disease prevalence. Taking specific motivation from ongoing efforts to identify recurrent cases based on the Georgia Cancer Registry, we extend recently proposed "anchor stream" sampling design and estimation methodology. Our approach offers a more efficient and defensible alternative to traditional capture-recapture (CRC) methods by leveraging a relatively small random sample of participants whose recurrence status is obtained through a principled application of medical records abstraction. This sample is combined with one or more existing signaling data streams, which may yield data based on arbitrarily non-representative subsets of the full registry population. The key extension developed here accounts for the common problem of false positive or negative diagnostic signals from the existing data stream(s). In particular, we show that the design only requires documentation of positive signals in these non-anchor surveillance streams, and permits valid estimation of the true case count based on an estimable positive predictive value (PPV) parameter. We borrow ideas from the multiple imputation paradigm to provide accompanying standard errors, and develop an adapted Bayesian credible interval approach that yields favorable frequentist coverage properties. We demonstrate the benefits of the proposed methods through simulation studies, and provide a data example targeting estimation of the breast cancer recurrence case count among Metro Atlanta area patients from the Georgia Cancer Registry-based Cancer Recurrence Information and Surveillance Program (CRISP) database.

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.

Effect of a FIT-Based Colorectal Cancer Screening Program on Mortality Estimated by the Regression Discontinuity Design.

The fecal immunochemical test (FIT) has been implemented in colorectal cancer (CRC) screening programs, but effect evaluations are lacking. We evaluated the effect of a positive FIT on all-cause and CRC mortality using the regression discontinuity design. The Danish CRC screening program invites all residents 50-74 years old, using a 20-µg hemoglobin/g feces cutoff for colonoscopy referral. In this cohort study, we followed all first-time screening participants from 2014-2019 until 2020. We estimated the local effect of screening results, of just above the cutoff vs. just below, as hazard ratios (HRs) between models fitted at each side of the cutoff. We conducted the analysis within a narrow hemoglobin range (≥17 and <23, n = 16,428) and a wider range (≥14 and <26, n = 35,353). Those screened just above the cutoff had lower all-cause mortality compared with below (HR = 0.87, 95% confidence interval: 0.69; 1.10), estimated from the narrow range. The CRC mortality analysis had few outcomes. In the wider range, those with a FIT just above the cutoff had a lower hazard of CRC mortality compared with just below the cutoff (HR = 0.49, 95% confidence interval: 0.17; 1.41). A FIT result just above the cutoff, leading to referral to colonoscopy, pointed towards reduced all-cause and CRC mortality compared with just below the cutoff.

Longitudinal weight status and academic achievement in elementary schoolchildren in the United States.

BACKGROUND/OBJECTIVES: Childhood overweight and obesity have a well-established negative impact on children's health. Overweight and obesity might also negatively impact children's academic performance, but existing literature on this association is inconclusive. This study uses a longitudinal design in a large, diverse elementary school sample to rigorously test the association between longitudinal weight status and academic achievement. Analyses also investigate modification by sex, race/ethnicity, and cardiorespiratory fitness (CRF). SUBJECTS/METHODS: In a large suburban school district in the United States, 4936 Grade 4 students were recruited. Demographic, course grade, and standardized test data were collected from school records for Grades 3-5, and body mass index and CRF were assessed each year. Students wore accelerometers during the school day for up to 15 days across three semesters (Grade 4 Fall and Spring, Grade 5 Fall) to objectively measure physical activity. Multiple imputation addressed missing data and multilevel analyses controlled for student demographics and clustering within schools. RESULTS: Unadjusted multilevel models found small negative associations between persistent obesity and course grades and standardized test scores, but these associations largely disappeared when controlling for demographic characteristics. Residual associations for math and writing course grades were attenuated when controlling for CRF, though some marginal negative associations for math and writing remained for students who developed obesity during follow-up. There was also evidence of marginal negative associations with course grades for students who developed overweight/obesity. There was no evidence of modification by sex or race/ethnicity. CONCLUSIONS: Results suggest very small associations between weight status and academic achievement that were largely explained by sociodemographic factors and CRF. Evidence of an association between weight status and achievement was stronger among students who developed overweight/obesity. Interventions promoting healthy weight and high CRF remain critical for schools given the link between student health and achievement.

Prevalence of self-inflicted injuries among transgender and gender diverse adolescents and young adults compared to their peers: an examination of interaction with mental health morbidity.

PURPOSE: Compare occurrence of self-inflicted injuries among transgender and gender diverse (TGD) youth to that of their cisgender peers while accounting for mental health diagnoses. METHODS: Review of electronic health records from three integrated health care systems identified 1087 transfeminine and 1431 transmasculine adolescents and young adults. Poisson regression was used to calculate prevalence ratios comparing the proportion of TGD participants with at least one self-inflicted injury (a surrogate for suicide attempt) before index date (first evidence of TGD status) to the corresponding proportions in presumed cisgender male and female referents matched on age, race/ethnicity, and health plan. Interactions between gender identities and mental health diagnoses were assessed on multiplicative and additive scales. RESULTS: TGD adolescents and young adults were more likely to have a self-inflicted injury, various mental health diagnoses, and multiple mental health diagnoses than their cisgender peers. The prevalence of self-inflicted injuries among TGD adolescents and young adults was high even in the absence of mental health diagnoses. Results were consistent with positive additive interaction and negative multiplicative interaction. CONCLUSIONS: Universal suicide prevention efforts for all youth, including those with no mental health diagnoses, and more intensive suicide prevention efforts for TGD adolescents and young adults and those with at least one mental health diagnosis are warranted.