High-dimensional propensity scores for empirical covariate selection in secondary database studies: Planning, implementation, and reporting.

Real-world evidence used for regulatory, payer, and clinical decision-making requires principled epidemiology in design and analysis, applying methods to minimize confounding given the lack of randomization. One technique to deal with potential confounding is propensity score (PS) analysis, which allows for the adjustment for measured preexposure covariates. Since its first publication in 2009, the high-dimensional propensity score (hdPS) method has emerged as an approach that extends traditional PS covariate selection to include large numbers of covariates that may reduce confounding bias in the analysis of healthcare databases. hdPS is an automated, data-driven analytic approach for covariate selection that empirically identifies preexposure variables and proxies to include in the PS model. This article provides an overview of the hdPS approach and recommendations on the planning, implementation, and reporting of hdPS used for causal treatment-effect estimations in longitudinal healthcare databases. We supply a checklist with key considerations as a supportive decision tool to aid investigators in the implementation and transparent reporting of hdPS techniques, and to aid decision-makers unfamiliar with hdPS in the understanding and interpretation of studies employing this approach. This article is endorsed by the International Society for Pharmacoepidemiology.

Visualizations throughout pharmacoepidemiology study planning, implementation, and reporting.

Transparency is increasingly promoted to instill trust in nonrandomized studies using real-world data. Graphics and data visualizations support transparency by aiding communication and understanding, and can inform study design and analysis decisions. However, other than graphical representation of a study design and flow diagrams (e.g., a Consolidated Standards of Reporting Trials [CONSORT] like diagram), specific standards on how to maximize validity and transparency with visualization are needed. This paper provides guidance on how to use visualizations throughout the life cycle of a pharmacoepidemiology study-from initial study design to final report-to facilitate rationalized and transparent decision-making about study design and implementation, and clear communication of study findings. Our intent is to help researchers align their practices with current consensus statements on transparency.

Categorization of COVID-19 severity to determine mortality risk.

PURPOSE: Algorithms for classification of inpatient COVID-19 severity are necessary for confounding control in studies using real-world data. METHODS: Using Healthverity chargemaster and claims data, we selected patients hospitalized with COVID-19 between April 2020 and February 2021, and classified them by severity at admission using an algorithm we developed based on respiratory support requirements (supplemental oxygen or non-invasive ventilation, O2/NIV, invasive mechanical ventilation, IMV, or NEITHER). To evaluate the utility of the algorithm, patients were followed from admission until death, discharge, or a 28-day maximum to report mortality risks and rates overall and by stratified by severity. Trends for heterogeneity in mortality risk and rate across severity classifications were evaluated using Cochran-Armitage and Logrank trend tests, respectively. RESULTS: Among 118 117 patients, the algorithm categorized patients in increasing severity as NEITHER (36.7%), O2/NIV (54.3%), and IMV (9.0%). Associated mortality risk (and 95% CI) was 11.8% (11.6-12.0%) overall and increased with severity [3.4% (3.2-3.5%), 11.5% (11.3-11.8%), 47.3% (46.3-48.2%); p < 0.001]. Mortality rate per 1000 person-days (and 95% CI) was 15.1 (14.9-15.4) overall and increased with severity [5.7 (5.4-6.0), 14.5 (14.2-14.9), 32.7 (31.8-33.6); p < 0.001]. CONCLUSION: As expected, we observed a positive association between the algorithm-defined severity on admission and 28-day mortality risk and rate. Although performance remains to be validated, this provides some assurance that this algorithm may be used for confounding control or stratification in treatment effect studies.

Real-world evidence of bariatric surgery and cardiovascular benefits using electronic health records data: A lesson in bias.

AIM: To reproduce and correct studies on bariatric surgery and the reduction in major adverse cardiovascular events (MACE) among patients with obesity and type 2 diabetes (T2D). METHODS: We used electronic healthcare records (EHR) from in and outpatient facilities around the United States to identify a cohort of patients with T2D, aged 18 to 80 years and with a body mass index (BMI) of 30 kg/m2 or higher undergoing bariatric surgery. We compared against hip/knee arthroplasty to establish an active comparison group that reduced bias from differential information and confounding. The main outcome was six-point MACE. Pre-exposure characteristics were adjusted in propensity score (PS) models with 1:2 matching plus high-dimensional PS matching. RESULTS: After a range of exclusions, the final cohort included 344 bariatric surgery patients (65% female; mean age 58 years) and 551 PS-matched patients undergoing arthroplasty (65% female; 59 years). Median follow-up was 2.5 years in both groups. Bariatric surgery patients showed a sustained 20% weight reduction and an HbA1c reduction by 1% point. We found no benefits of bariatric surgery for six-point MACE (HR = 0.99; 95% CI 0.76-1.30). We observed known increases in risks for vitamin B12 deficiency anaemia (HR = 3.06; 1.10-8.49) and cholelithiasis (HR = 1.72; 0.94-3.13). CONCLUSIONS: This real-world evidence study found reductions in HbA1c and BMI following bariatric surgery similar to trials, and no meaningful cardiovascular benefit compatible with the underpowered trials but in contrast to earlier EHR studies. We showed how information bias typical in EHR analyses and confounding may cause substantial bias.

Diagnosis-wide analysis of COVID-19 complications: an exposure-crossover study.

BACKGROUND: Many studies reporting coronavirus disease 2019 (COVID-19) complications have involved case series or small cohorts that could not establish a causal association with COVID-19 or provide risk estimates in different care settings. We sought to study all possible complications of COVID-19 to confirm previously reported complications and to identify potential complications not yet known. METHODS: Using United States health claims data, we compared the frequency of all International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) diagnosis codes occurring before and after the onset of the COVID-19 pandemic in an exposure-crossover design. We included patients who received a diagnosis of COVID-19 between Mar. 1, 2020, and Apr. 30, 2020, and computed risk estimates and odds ratios (ORs) of association with COVID-19 for every ICD-10-CM diagnosis code. RESULTS: Among 70 288 patients with COVID-19, 69 of 1724 analyzed ICD-10-CM diagnosis codes were significantly associated with COVID-19. Disorders showing both strong association with COVID-19 and high absolute risk included viral pneumonia (OR 177.63, 95% confidence interval [CI] 147.19-214.37, absolute risk 27.6%), respiratory failure (OR 11.36, 95% CI 10.74-12.02, absolute risk 22.6%), acute kidney failure (OR 3.50, 95% CI 3.34-3.68, absolute risk 11.8%) and sepsis (OR 4.23, 95% CI 4.01-4.46, absolute risk 10.4%). Disorders showing strong associations with COVID-19 but low absolute risk included myocarditis (OR 8.17, 95% CI 3.58-18.62, absolute risk 0.1%), disseminated intravascular coagulation (OR 11.83, 95% CI 5.26-26.62, absolute risk 0.1%) and pneumothorax (OR 3.38, 95% CI 2.68-4.26, absolute risk 0.4%). INTERPRETATION: We confirmed and provided risk estimates for numerous complications of COVID-19. These results may guide prognosis, treatment decisions and patient counselling.

Empirical assessment of case-based methods for identification of drugs associated with acute liver injury in the French National Healthcare System database (SNDS).

PURPOSES: Drug induced acute liver injury (ALI) is a frequent cause of liver failure. Case-based designs were empirically assessed and calibrated in the French National claims database (SNDS), aiming to identify the optimum design for drug safety alert generation associated with ALI. METHODS: All cases of ALI were extracted from SNDS (2009-2014) using specific and sensitive definitions. Positive and negative drug controls were used to compare 196 self-controlled case series (SCCS), case-control (CC), and case-population (CP) design variants, using area under the receiver operating curve (AUC), mean square error (MSE) and coverage probability. Parameters that had major impacts on results were identified through logistic regression. RESULTS: Using a specific ALI definition, AUCs ranged from 0.78 to 0.94, 0.64 to 0.92 and 0.48 to 0.85, for SCCS, CC and CP, respectively. MSE ranged from 0.12 to 0.40, 0.22 to 0.39 and 1.03 to 5.29, respectively. Variants adjusting for multiple drug use had higher coverage probabilities. Univariate regressions showed that high AUCs were achieved with SCCS using exposed time as the risk window. The top SCCS variant yielded an AUC = 0.93 and MSE = 0.22 and coverage = 86%, with 1/7 negative and 13/18 positive controls presenting significant estimates. CONCLUSIONS: SCCS adjusting for multiple drugs and using exposed time as the risk window performed best in generating ALI-related drug safety alert and providing estimates of the magnitude of the risk. This approach may be useful for ad-hoc pharmacoepidemiology studies to support regulatory actions.

Empirical assessment of case-based methods for drug safety alert identification in the French National Healthcare System database (SNDS): Methodology of the ALCAPONE project.

OBJECTIVES: To introduce the methodology of the ALCAPONE project. BACKGROUND: The French National Healthcare System Database (SNDS), covering 99% of the French population, provides a potentially valuable opportunity for drug safety alert generation. ALCAPONE aimed to assess empirically in the SNDS case-based designs for alert generation related to four health outcomes of interest. METHODS: ALCAPONE used a reference set adapted from observational medical outcomes partnership (OMOP) and Exploring and Understanding Adverse Drug Reactions (EU-ADR) project, with four outcomes-acute liver injury (ALI), myocardial infarction (MI), acute kidney injury (AKI), and upper gastrointestinal bleeding (UGIB)-and positive and negative drug controls. ALCAPONE consisted of four main phases: (1) data preparation to fit the OMOP Common Data Model and select the drug controls; (2) detection of the selected controls via three case-based designs: case-population, case-control, and self-controlled case series, including design variants (varying risk window, adjustment strategy, etc.); (3) comparison of design variant performance (area under the ROC curve, mean square error, etc.); and (4) selection of the optimal design variants and their calibration for each outcome. RESULTS: Over 2009-2014, 5225 cases of ALI, 354 109 MI, 12 633 AKI, and 156 057 UGIB were identified using specific definitions. The number of detectable drugs ranged from 61 for MI to 25 for ALI. Design variants generated more than 50 000 points estimates. Results by outcome will be published in forthcoming papers. CONCLUSIONS: ALCAPONE has shown the interest of the empirical assessment of pharmacoepidemiological approaches for drug safety alert generation and may encourage other researchers to do the same in other databases.

Empirical assessment of case-based methods for identification of drugs associated with upper gastrointestinal bleeding in the French National Healthcare System database (SNDS).

PURPOSE: Upper gastrointestinal bleeding (UGIB) is a severe and frequent drug-related event. In order to enable efficient drug safety alert generation in the French National Healthcare System database (SNDS), we assessed and calibrated empirically case-based designs to identify drug associated with UGIB risk. METHODS: All cases of UGIB were extracted from SNDS (2009-2014) using two definitions. Positive and negative drug controls were used to compare 196 self-controlled case series (SCCS), case-control (CC) and case-population (CP) design variants. Each variant was evaluated in a 1/10th population sample using area under the receiver operating curve (AUC) and mean square error (MSE). Parameters that had major impacts on results were identified through logistic regression. Optimal designs were replicated in the unsampled population. RESULTS: Using a specific UGIB definition, AUCs ranged from 0.64 to 0.80, 0.44 to 0.61 and 0.50 to 0.67, for SCCS, CC and CP, respectively. MSE ranged from 0.07 to 0.39, 0.83 to 1.33 and 1.96 to 4.6, respectively. Univariate regressions showed that high AUCs were achieved with SCCS with multiple drug adjustment and a 30-day risk window starting at exposure. The top-performing SCCS variant in the unsampled population yielded an AUC = 0.84 and MSE = 0.14, with 10/36 negative controls presenting significant estimates. CONCLUSIONS: SCCS adjusting for multiple drugs and using a 30-day risk window has the potential to generate UGIB-related alerts in the SNDS and hypotheses on its potential population impact. Negative control implementation highlighted that low systematic error was generated but that protopathic bias and confounding by indication remained unaddressed issues.

Graphical Depiction of Longitudinal Study Designs in Health Care Databases.

Pharmacoepidemiologic and pharmacoeconomic analysis of health care databases has become a vital source of evidence to support health care decision making and efficient management of health care organizations. However, decision makers often consider studies done in nonrandomized health care databases more difficult to review than randomized trials because many design choices need to be considered. This is perceived as an important barrier to decision making about the effectiveness and safety of medical products. Design flaws in longitudinal database studies are avoidable but can be unintentionally obscured in the convoluted prose of methods sections, which often lack specificity. We propose a simple framework of graphical representation that visualizes study design implementations in a comprehensive, unambiguous, and intuitive way; contains a level of detail that enables reproduction of key study design variables; and uses standardized structure and terminology to simplify review and communication to a broad audience of decision makers. Visualization of design details will make database studies more reproducible, quicker to review, and easier to communicate to a broad audience of decision makers.

Reporting to Improve Reproducibility and Facilitate Validity Assessment for Healthcare Database Studies V1.0.

PURPOSE: Defining a study population and creating an analytic dataset from longitudinal healthcare databases involves many decisions. Our objective was to catalogue scientific decisions underpinning study execution that should be reported to facilitate replication and enable assessment of validity of studies conducted in large healthcare databases. METHODS: We reviewed key investigator decisions required to operate a sample of macros and software tools designed to create and analyze analytic cohorts from longitudinal streams of healthcare data. A panel of academic, regulatory, and industry experts in healthcare database analytics discussed and added to this list. CONCLUSION: Evidence generated from large healthcare encounter and reimbursement databases is increasingly being sought by decision-makers. Varied terminology is used around the world for the same concepts. Agreeing on terminology and which parameters from a large catalogue are the most essential to report for replicable research would improve transparency and facilitate assessment of validity. At a minimum, reporting for a database study should provide clarity regarding operational definitions for key temporal anchors and their relation to each other when creating the analytic dataset, accompanied by an attrition table and a design diagram. A substantial improvement in reproducibility, rigor and confidence in real world evidence generated from healthcare databases could be achieved with greater transparency about operational study parameters used to create analytic datasets from longitudinal healthcare databases.

Effects of expanding the look-back period to all available data in the assessment of covariates.

BACKGROUND: A fixed baseline period has been a common covariate assessment approach in pharmacoepidemiological studies from claims but may lead to high levels of covariate misclassification. Simulation studies have recommended expanding the look-back approach to all available data (AAD) for binary indicators of diagnoses, procedures, and medications, but there have been few real data analyses using this approach. OBJECTIVE: The objective of the study is to explore the impact on treatment effect estimates and covariate prevalence of expanding the look-back period within five validated studies in the Aetion system, a rapid cycle analytics platform. METHODS: We reran the five studies and assessed covariates using (i) a fixed window approach (usually 180 days before treatment initiation), (ii) AAD prior to treatment initiation, and (iii) AAD with a categorized by recency approach, where the most recent occurrence of a covariate was labeled as recent (occurring within the fixed window) or past (before the start of the fixed window). For each covariate assessment approach, we adjusted for covariates via propensity score matching. RESULTS: All studies had at least one covariate that had an increase in prevalence of 15% or higher from the fixed window to the AAD approach. However, there was little change in treatment effect estimates resulting from differing covariate assessment approaches. For example, in a study of acute coronary syndrome in high-intensity versus low-intensity statin users, the estimated hazard ratio from the fixed window approach was 1.11 (95% confidence interval 0.98, 1.25) versus 1.21 (1.07, 1.37) when using AAD and 1.19 (1.05, 1.35) using categorized by recency. CONCLUSION: Expanding the baseline period to AAD improved covariate sensitivity by capturing data that would otherwise be missed yet did not meaningfully change the overall treatment effect estimates compared with the fixed window approach. Copyright © 2017 John Wiley & Sons, Ltd.

Reporting to Improve Reproducibility and Facilitate Validity Assessment for Healthcare Database Studies V1.0.

PURPOSE: Defining a study population and creating an analytic dataset from longitudinal healthcare databases involves many decisions. Our objective was to catalogue scientific decisions underpinning study execution that should be reported to facilitate replication and enable assessment of validity of studies conducted in large healthcare databases. METHODS: We reviewed key investigator decisions required to operate a sample of macros and software tools designed to create and analyze analytic cohorts from longitudinal streams of healthcare data. A panel of academic, regulatory, and industry experts in healthcare database analytics discussed and added to this list. CONCLUSION: Evidence generated from large healthcare encounter and reimbursement databases is increasingly being sought by decision-makers. Varied terminology is used around the world for the same concepts. Agreeing on terminology and which parameters from a large catalogue are the most essential to report for replicable research would improve transparency and facilitate assessment of validity. At a minimum, reporting for a database study should provide clarity regarding operational definitions for key temporal anchors and their relation to each other when creating the analytic dataset, accompanied by an attrition table and a design diagram. A substantial improvement in reproducibility, rigor and confidence in real world evidence generated from healthcare databases could be achieved with greater transparency about operational study parameters used to create analytic datasets from longitudinal healthcare databases.

Analyse d'ensemble des complications de la COVID-19 : étude d'auto-appariement pré- et post-exposition.

CONTEXTE: De nombreuses études sur les complications de la maladie à coronavirus 2019 (COVID-19) ont reposé sur des séries de cas et de petites cohortes qui ne permettaient pas d'établir un lien causal avec la COVID-19 ni d'estimer les risques dans les différents milieux de soins. Nous avons voulu étudier toutes les complications possibles de la COVID-19 afin de confirmer les complications précédemment déclarées et d'identifier de potentielles complications encore inconnues. MÉTHODES: À partir des données sur les demandes de remboursement de frais médicaux aux États-Unis, nous avons comparé la fréquence de tous les codes de diagnostic de la Classification internationale des maladies, 10 e révision, modification clinique (CIM-10-MC) enregistrés avant et après le déclenchement de la pandémie de COVID-19 dans un modèle d'auto-appariement pré- et post-exposition. Nous avons inclus les patients ayant reçu un diagnostic de COVID-19 entre le 1er mars 2020 et le 30 avril 2020, et calculé les estimations de risque et les rapports de cotes (RC) pour le lien avec la COVID-19 de chaque code de diagnostic de la CIM-10-MC. RÉSULTATS: Sur les 1724 codes de diagnostic de la CIM-10-MC attribués à 70 288 patients atteints de COVID-19, 69 étaient significativement liés à la COVID-19. Les diagnostics étroitement liés à la COVID-19 et comportant un risque absolu élevé comprenaient la pneumonie virale (RC 177,63; intervalle de confiance [IC] à 95 % 147,19­214,37; risque absolu 27,6 %), l'insuffisance respiratoire (RC 11,36; IC à 95 % 10,74­12,02; risque absolu 22,6 %), l'insuffisance rénale aiguë (RC 3,50; IC à 95 % 3,34­3,68; risque absolu 11,8 %) et la sepsie (RC 4,23; IC à 95 % 4,01­4,46; risque absolu 10,4 %). Les diagnostics étroitement liés à la COVID-19, mais comportant un risque absolu faible comprenaient la myocardite (RC 8,17; IC à 95 % 3,58­18,62; risque absolu 0,1 %), la coagulation intravasculaire disséminée (RC 11,83; IC à 95 % 5,26­26,62; risque absolu 0,1 %) et le pneumothorax (RC 3,38; IC à 95 % 2,68­4,26; risque absolu 0,4 %). INTERPRÉTATION: Nous avons confirmé et établi les estimations du risque de plusieurs complications de la COVID-19. Ces résultats pourraient orienter le pronostic, les décisions thérapeutiques et les conseils aux patients.

High-dimensional propensity score algorithm in comparative effectiveness research with time-varying interventions.

The high-dimensional propensity score (hdPS) algorithm was proposed for automation of confounding adjustment in problems involving large healthcare databases. It has been evaluated in comparative effectiveness research (CER) with point treatments to handle baseline confounding through matching or covariance adjustment on the hdPS. In observational studies with time-varying interventions, such hdPS approaches are often inadequate to handle time-dependent confounding and selection bias. Inverse probability weighting (IPW) estimation to fit marginal structural models can adequately handle these biases under the fundamental assumption of no unmeasured confounders. Upholding of this assumption relies on the selection of an adequate set of covariates for bias adjustment. We describe the application and performance of the hdPS algorithm to improve covariate selection in CER with time-varying interventions based on IPW estimation and explore stabilization of the resulting estimates using Super Learning. The evaluation is based on both the analysis of electronic health records data in a real-world CER study of adults with type 2 diabetes and a simulation study. This report (i) establishes the feasibility of IPW estimation with the hdPS algorithm based on large electronic health records databases, (ii) demonstrates little impact on inferences when supplementing the set of expert-selected covariates using the hdPS algorithm in a setting with extensive background knowledge, (iii) supports the application of the hdPS algorithm in discovery settings with little background knowledge or limited data availability, and (iv) motivates the application of Super Learning to stabilize effect estimates based on the hdPS algorithm.

Prospective cohort studies of newly marketed medications: using covariate data to inform the design of large-scale studies.

BACKGROUND: Nonrandomized safety and effectiveness studies are often initiated immediately after the approval of a new medication, but patients prescribed the new medication during this period may be substantially different from those receiving an existing comparator treatment. Restricting the study to comparable patients after data have been collected is inefficient in prospective studies with primary collection of outcomes. METHODS: We discuss design and methods for evaluating covariate data to assess the comparability of treatment groups, identify patient subgroups that are not comparable, and decide when to transition to a large-scale comparative study. We demonstrate methods in an example study comparing Cox-2 inhibitors during their postmarketing period (1999-2005) with nonselective nonsteroidal anti-inflammatory drugs (NSAIDs). RESULTS: Graphical checks of propensity score distributions in each treatment group showed substantial problems with overlap in the initial cohorts. In the first half of 1999, >40% of patients were in the region of nonoverlap on the propensity score, and across the study period this fraction never dropped below 10% (the a priori decision threshold for transitioning to the large-scale study). After restricting to patients with no prior NSAID use, <1% of patients were in the region of nonoverlap, indicating that a large-scale study could be initiated in this subgroup and few patients would need to be trimmed from analysis. CONCLUSIONS: A sequential study design that uses pilot data to evaluate treatment selection can guide the efficient design of large-scale outcome studies with primary data collection by focusing on comparable patients.

Metrics for covariate balance in cohort studies of causal effects.

Inferring causation from non-randomized studies of exposure requires that exposure groups can be balanced with respect to prognostic factors for the outcome. Although there is broad agreement in the literature that balance should be checked, there is confusion regarding the appropriate metric. We present a simulation study that compares several balance metrics with respect to the strength of their association with bias in estimation of the effect of a binary exposure on a binary, count, or continuous outcome. The simulations utilize matching on the propensity score with successively decreasing calipers to produce datasets with varying covariate balance. We propose the post-matching C-statistic as a balance metric and found that it had consistently strong associations with estimation bias, even when the propensity score model was misspecified, as long as the propensity score was estimated with sufficient study size. This metric, along with the average standardized difference and the general weighted difference, outperformed all other metrics considered in association with bias, including the unstandardized absolute difference, Kolmogorov-Smirnov and Lévy distances, overlapping coefficient, Mahalanobis balance, and L1 metrics. Of the best-performing metrics, the C-statistic and general weighted difference also have the advantage that they automatically evaluate balance on all covariates simultaneously and can easily incorporate balance on interactions among covariates. Therefore, when combined with the usual practice of comparing individual covariate means and standard deviations across exposure groups, these metrics may provide useful summaries of the observed covariate imbalance.

A modular, prospective, semi-automated drug safety monitoring system for use in a distributed data environment.

PURPOSE: The aim of this study was to develop and test a semi-automated process for conducting routine active safety monitoring for new drugs in a network of electronic healthcare databases. METHODS: We built a modular program that semi-automatically performs cohort identification, confounding adjustment, diagnostic checks, aggregation and effect estimation across multiple databases, and application of a sequential alerting algorithm. During beta-testing, we applied the system to five databases to evaluate nine examples emulating prospective monitoring with retrospective data (five pairs for which we expected signals, two negative controls, and two examples for which it was uncertain whether a signal would be expected): cerivastatin versus atorvastatin and rhabdomyolysis; paroxetine versus tricyclic antidepressants and gastrointestinal bleed; lisinopril versus angiotensin receptor blockers and angioedema; ciprofloxacin versus macrolide antibiotics and Achilles tendon rupture; rofecoxib versus non-selective non-steroidal anti-inflammatory drugs (ns-NSAIDs) and myocardial infarction; telithromycin versus azithromycin and hepatotoxicity; rosuvastatin versus atorvastatin and diabetes and rhabdomyolysis; and celecoxib versus ns-NSAIDs and myocardial infarction. RESULTS: We describe the program, the necessary inputs, and the assumed data environment. In beta-testing, the system generated four alerts, all among positive control examples (i.e., lisinopril and angioedema; rofecoxib and myocardial infarction; ciprofloxacin and tendon rupture; and cerivastatin and rhabdomyolysis). Sequential effect estimates for each example were consistent in direction and magnitude with existing literature. CONCLUSIONS: Beta-testing across nine drug-outcome examples demonstrated the feasibility of the proposed semi-automated prospective monitoring approach. In retrospective assessments, the system identified an increased risk of myocardial infarction with rofecoxib and an increased risk of rhabdomyolysis with cerivastatin years before these drugs were withdrawn from the market.

Plasmode simulation for the evaluation of pharmacoepidemiologic methods in complex healthcare databases.

Longitudinal healthcare claims databases are frequently used for studying the comparative safety and effectiveness of medications, but results from these studies may be biased due to residual confounding. It is unclear whether methods for confounding adjustment that have been shown to perform well in small, simple nonrandomized studies are applicable to the large, complex pharmacoepidemiologic studies created from secondary healthcare data. Ordinary simulation approaches for evaluating the performance of statistical methods do not capture important features of healthcare claims. A statistical framework for creating replicated simulation datasets from an empirical cohort study in electronic healthcare claims data is developed and validated. The approach relies on resampling from the observed covariate and exposure data without modification in all simulated datasets to preserve the associations among these variables. Repeated outcomes are simulated using a true treatment effect of the investigator's choice and the baseline hazard function estimated from the empirical data. As an example, this framework is applied to a study of high versus low-intensity statin use and cardiovascular outcomes. Simulated data is based on real data drawn from Medicare Parts A and B linked with a prescription drug insurance claims database maintained by Caremark. Properties of the data simulated using this framework are compared with the empirical data on which the simulations were based. In addition, the simulated datasets are used to compare variable selection strategies for confounder adjustmentvia the propensity score, including high-dimensional approaches that could not be evaluated with ordinary simulation methods. The simulated datasets are found to closely resemble the observed complex data structure but have the advantage of an investigator-specified exposure effect.

Heart failure risk among patients with rheumatoid arthritis starting a TNF antagonist.

BACKGROUND: While heart failure (HF) is associated with elevations in tumor necrosis factor (TNF)α, several trials of TNF antagonists showed no benefit and possibly worsening of disease in those with known severe HF. We studied the risk of new or recurrent HF among a group of patients receiving these agents to treat rheumatoid arthritis (RA). METHODS: We used data from four different US healthcare programmes. Subjects with RA receiving methotrexate were eligible to enter the study cohort if they added or switched to a TNF antagonist or another non-biological disease modifying antirheumatic drug (nbDMARD). These groups were compared in Cox regression models stratified by propensity score decile and adjusted for oral glucocorticoid dosage, prior HF hospitalisations, and the use of loop diuretics. RESULTS: We compared 8656 new users of a nbDMARD with 11 587 new users of a TNF antagonist with similar baseline covariates. The HR for the TNF antagonists compared with nbDMARD was 0.85 (95% CI 0.63 to 1.14). The HR was also not elevated in subjects with a history of HF. But, it was elevated prior to 2002 (HR 2.17, 95% CI 0.45 to 10.50, test for interaction p=0.036). Oral glucocorticoids were associated with a dose-related gradient of HF risk: compared with no use, 1≤5 mg HR 1.30 (95% CI 0.91 to 1.85), ≥5 mg HR 1.54 (95% CI 1.09 to 2.19). CONCLUSIONS: TNF antagonists were not associated with a risk of HF hospital admissions compared with nbDMARDs in this RA population.