Search

Purpose: For observational cohort studies that employ matching by propensity scores (PS), preliminary stratification by consequential predictors of outcome better emulates stratified randomization and potentially reduces variance and bias through relaxed dependence on modeling assumptions. We assessed the impact of pre‐stratification in two real‐life examples. For both, prior evidence from placebo‐controlled randomized clinical trials (RCTs) suggested small or no risk reduction, but observational analysis suggested protection, presumably the result of confounding bias. Study Design and Setting: The study populations consisted of Medicare beneficiaries (2014‐18) with type 2 diabetes initiating either (i) empagliflozin versus dipeptidyl peptidase‐4 inhibitors (DPP‐4i) or (ii) empagliflozin versus glucagon‐like peptide‐1 receptor agonists (GLP‐1RA). The outcome was myocardial infarction or stroke. We estimated hazard ratios (HR) and rate differences (RD) after controlling for 143 pre‐exposure covariates via 1:1 PS matching after (1) PS estimation in the total cohort (total‐cohort PS‐matching) and (2) PS estimation separately by baseline cardiovascular disease (stratified PS matching). Results: Stratified PS matching resulted in HRs that exceeded those from total‐cohort PS‐matching by 13% and 9%, respectively, for the comparisons of empagliflozin to DPP‐4i and GLP‐1RA. Against both comparators, HRs and RDs after stratified PS matching were closer to the null, with slightly higher variances (2%–3%) than those after total‐cohort PS matching. Conclusion: Stratified PS matching produced effect estimates closer to the expected trial findings than total‐cohort PS matching. The price paid in increased variance was minimal. [ABSTRACT FROM AUTHOR]

Purpose: Azithromycin is a common first‐line antibiotic for respiratory infection; however, there is conflicting evidence regarding risk of cardiovascular death. We assessed cardiovascular and noncardiovascular mortality associated with azithromycin versus amoxicillin–clavulanate among US Veterans treated for nonear–nose–throat respiratory infection ("respiratory") or ear–nose–throat infection indication. Methods: Electronic health record data from the US Veterans Health Administration database were used to identify Veterans (30–74 years) with outpatient dispensings of oral azithromycin versus amoxicillin–clavulanate for respiratory or ear–nose–throat infection (January 01, 2000–December 31, 2014). Outcomes assessed were risk of cardiovascular death and noncardiovascular death within 1–5 and 6–10 days postdispensing. Inverse probability of treatment‐weighted proportional hazards models and binomial regression models were used to estimate hazard ratios (HRs) and compute risk differences (RD) per million courses of therapy. Cardiac death (subset of cardiovascular death) was assessed in sensitivity analyses. Results: There were 629 345 azithromycin and 168 429 amoxicillin–clavulanate dispensings for respiratory indications, 143 783 azithromycin, and 203 142 amoxicillin–clavulanate dispensings for ear–nose–throat indications. For respiratory indications, azithromycin was not associated with a significantly different risk of cardiovascular death versus amoxicillin–clavulanate within 1–5 days postdispensing (HR [95% confidence interval (CI)]: 1.12 [0.63, 2.00]; RD [95% CI]: 11 [−43, 64] deaths/million courses of therapy). No elevated risk for azithromycin was found for ear–nose–throat indications. Pooled results for both indications via meta‐analysis showed no association between antibiotics and cardiovascular mortality. There was no significant difference in risk of noncardiovascular or cardiac death between antibiotics postdispensing. Conclusion: Azithromycin was not associated with elevated risk of cardiovascular or noncardiovascular death versus amoxicillin–clavulanate among US Veterans. [ABSTRACT FROM AUTHOR]

Analyses of suspended sediment transport provide valuable insight into the role that sediment supply plays in causing geomorphic change. The sediment supply within a river system evolves depending on the discharge, flood frequency and duration, changes in sediment input, and ecohydraulic conditions that modify sediment transport processes. Changes in supply can be evaluated through analyses of coupled changes in suspended sediment concentration and grain size. The concentration of sand in transport in the Green and Colorado Rivers is most strongly controlled by discharge and the bed sand grain size distribution. Since the 1950s, sand loads have decreased in response to declines in peak discharge in the Green River and coarsening of the bed sand in the Colorado River. However, changes in the bed sand grain size distribution are associated with large changes in suspended sand concentration in both rivers; concentration varies by a factor of ~3 in the Green River and a factor of ~8 in the Colorado River, depending on the bed sand grain size distribution. Analyses of hysteresis in suspended sediment measurements show that sediment depletion during annual floods is most strongly controlled by flood duration, with peak discharge being nearly equally important in the Green River. Despite channel narrowing in both rivers, periods of bed sand coarsening and sediment depletion during annual floods indicate that these rivers are not necessarily in sediment surplus. Channel narrowing appears to be strongly controlled by short‐term declines in flood magnitude and the ecohydraulic effects of vegetation and may not be indicative of the long‐term sediment budget. Plain Language Summary: River channels change size and shape in response to changes in the amount of sediment transported downstream. Changes in streamflow and/or the upstream sediment supply are the cause(s) of such changes in sediment transport. The channels of the Green and Colorado Rivers near Canyonlands National Park, Utah, have both narrowed over the last century. We use measurements of suspended sediment transport to investigate how changes in the sediment supply influence sediment transport and channel change. In most cases, the transport of suspended sand is primarily controlled by the discharge of water and secondarily controlled by the bed sediment grain size distribution. Depletion of the upstream sand supply leads to bed sand coarsening and erosion, whereas enrichment of the upstream sand supply leads to bed sand fining and deposition. The sand supply is progressively depleted during annual snowmelt floods on the Green and Colorado Rivers, with greater depletion occurring during longer floods. Larger floods also cause greater depletion of the upstream sand supply in the Green River but are of less importance in the Colorado River. The size and shape of the present‐day river channels may therefore be maintained, and channel narrowing may be limited, if longer‐duration floods occur in the future. Key Points: Sand transport has declined because of reductions in discharge in the Green River and reductions in sand supply in the Colorado RiverSediment depletion occurs during annual floods and is determined by flood duration and magnitude, depending on the riverDespite channel narrowing, periods of bed sand coarsening and sediment depletion do not indicate conditions of sediment surplus [ABSTRACT FROM AUTHOR]

Background We reviewed the results of the Observational Medical Outcomes Research Partnership (OMOP) 2010 Experiment in hopes of finding examples where apparently well-designed drug studies repeatedly produce anomalous findings. OMOP had applied thousands of designs and design parameters to 53 drug-outcome pairs across 10 electronic data resources. Our intent was to use this repository to elucidate some sources of error in observational studies. Method From the 2010 OMOP Experiment, we sought drug-outcome-method combinations (DOMCs) that met consensus design criteria, yet repeatedly produced results contrary to expectation. We set aside DOMCs for which we could not agree on the suitability of the designs, then selected for an in-depth scrutiny one drug-outcome pair analyzed by a seemingly plausible methodological approach, whose results consistently disagreed with the a priori expectation. Results The OMOP 'all-by-all' assessment of possible DOMCs yielded many combinations that would not be chosen by researchers as actual study options. Among those that passed a first level of scrutiny, two of seven drug-outcome pairs for which there were plausible research designs had anomalous results. The use of benzodiazepines was unexpectedly associated with acute renal failure and upper gastrointestinal bleeding. We chose the latter as an example for in-depth study. The factitious appearance of a bleeding risk may have been partly driven by an excess of procedures on the first day of treatment. A risk window definition that excluded the first day largely removed the spurious association. Conclusion One cause of reproducible 'error' may be repeated failure to tie design choices closely enough to the research question at hand. Copyright © 2016 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Purpose The purpose of this review is to assist researchers in developing, using, and interpreting case-identifying algorithms in electronic healthcare databases. Methods We review clinical characteristics of health outcomes, data settings and informatics, and epidemiologic and statistical methods aspects as they pertain to the development and use of case-identifying algorithms. Results We offer a framework for thinking critically about the use of electronic health insurance data and electronic health records to identify the occurrence of health outcomes. Accuracy of case ascertainment in database research depends on many factors, including clinical and behavioral aspects of the health outcome, and details of database construction as it pertains to completeness and reliability of database content. Existing methods for diagnostic and screening tests, misclassification, validation studies, and predictive modelling can be usefully applied to improve case ascertainment in database research. Conclusions Good case-identifying algorithms are based on a sound understanding of care-seeking behavior and patterns of clinical diagnosis and treatment in the study population and details about the construction and characteristics of the database. Researchers should use quantitative bias analyses to take into account the performance characteristics of case-identifying algorithms and their impact on study results. Copyright © 2015 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

ABSTRACT Purpose Clinical trials of olmesartan for prevention of progression of renal disease in patients with diabetes showed renal protection but an unexpected imbalance in cardiac deaths. The US Food and Drug Administration requested from the manufacturer a cohort study of olmesartan, other angiotensin-receptor blockers (ARBs), and angiotensin-converting enzyme (ACE) inhibitors in a large population. Methods A retrospective cohort study was conducted with the cooperation of a US health insurer. Subject entry and follow-up ran from 2002 through 2009. In propensity-matched cohorts, the primary analysis considered continuous current users. Endpoints were sudden cardiac death (SCD) and all-cause mortality, identified through the US National Death Index, supplemented by insurance and hospital discharge data. Statistical estimation was based on proportional hazards analyses with 95% confidence intervals. Power calculations had shown that 25 000 olmesartan initiators would be required to detect relative risks (RRs) of SCD of twofold or greater. Results A total of 57 123 initiators of olmesartan were matched 1:2 to initiators of other ARBs and 41 801 to initiators of ACE inhibitors. Average follow-up time ranged from 8 to 9 months. Olmesartan initiators and comparators experienced similar patterns of both outcomes, with RRs ≤1.0 and upper confidence bounds ≤1.6. Among persons with prior use of hypoglycemic agents, in comparison with other ARBs, the RR of SCD for olmesartan users was 0.8, with an upper confidence bound of 2.2. Conclusion The results of this well-powered study do not raise concerns for the risk of SCD or death from all causes among olmesartan users in comparison with users of other ARBs or ACE inhibitors. Copyright © 2013 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

ABSTRACT Purpose A new meningococcal conjugate vaccine (MCV4) was introduced in 2005. Shortly after, case reports of Guillain-Barré syndrome (GBS), a serious demyelinating disease, began to be reported to the Vaccine Adverse Event Reporting System. In 2006, the Centers for Disease Control and Prevention and the Food and Drug Administration requested the evaluation of GBS risk after MCV4 vaccination. We conducted a study to assess the risk of GBS after MCV4 vaccination using health plan administrative and claims data together with the review of primary medical records of potential cases. Methods Retrospective cohort study among 12.6 million 11- to 21-year-old members of five US health plans with a total membership of 50 million. Automated enrollment and medical claims data from March 2005 through August 2008 were used to identify the population, the vaccinations administered, and the medical services associated with possible GBS. Medical records were reviewed and adjudicated by a neurologist panel to confirm cases of GBS. The study used distributed data analysis methods that minimized sharing of protected health information. Results We confirmed 99 GBS cases during 18,322,800 person-years (5.4/1,000,000 person-years). More than 1.4 million MCV4 vaccinations were observed. No confirmed cases of GBS occurred within 6 weeks after vaccination. The upper 95% CI for the attributable risk of GBS associated with MCV4 is estimated as 1.5 cases per 1,000,000 doses. Conclusions Among members of five US health plans, MCV4 vaccination was not associated with increased GBS risk. Copyright © 2012 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

ABSTRACT Background Prospective medical product monitoring is intended to alert stakeholders about whether and when safety problems are identifiable in longitudinal electronic healthcare data. Little attention has been given to how to compare methods in this setting. Purpose To explore aspects of prospective monitoring that should be considered when comparing method performance and to develop a metric that explicitly accounts for these considerations. Methods We reviewed existing metrics and propose an event-based approach that classifies exposed outcomes according to whether a prior alert was generated. Results In comparing performance of methods for prospective monitoring, three factors must be considered: (1) accuracy in alerting; (2) timeliness of alerting; and (3) the trade-offs between the costs of false negative and false positive alerting. Traditional scenario-based measures of accuracy, such as sensitivity and specificity, which classify only at the end of monitoring, fail to appreciate timeliness of alerting and impose fixed tradeoffs between false positive versus false negative consequences. We provide an expression that summarizes event-based sensitivity (the proportion of exposed events that occur after alerting among all exposed events in scenarios with true safety issues) and event-based specificity (the proportion of exposed events that occur in the absence of alerting among all exposed events in scenarios with no true safety issues) by taking an average weighted by relative costs of false positive and false negative alerting. Conclusions The proposed approach explicitly accounts for accuracy in alerting, timeliness in alerting, and the trade-offs between the costs of false negative and false positive alerting. Copyright © 2012 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

ABSTRACT Background Studies based on the first appearance of a code for esophageal cancer in the General Practice Research Database (GPRD) have yielded conflicting results concerning the relationship between disease onset and prior bisphosphonate use. The literature on the timing of cancer code appearance in general practice records is scanty but suggests that there may be substantial error by comparison with actual dates of clinical onset. Objectives To assess the accuracy of codes for esophageal cancer in the GPRD and to determine whether it was possible to infer clinical date of onset. Methods A reviewer adjudicated the records of women with codes for esophageal cancer from 1996 to 2008 by using chronological GPRD listings of codes for diagnoses, services, and tests. GPRD staff sought reviews by the general practices for 75 women whom the reviewer classified as having esophageal cancer and 25 classified as not having the condition. Results Essentially all cases bearing a code for the condition had esophageal cancer. Where data were available to permit dating of clinical onset (75% of cases), the result agreed with the date recorded in the clinical record to within 60 days in 89% of cases. The remaining 25% of cases had little or no clinical information and could not be dated. Clinical onset preceded the first appearance of a cancer code by more than 6 months in 10% of cases. Conclusions The accurate timing of clinical onset and diagnosis in the GPRD, at least of esophageal cancer in women, may require the review of clinical records or future linkage to cancer registries. Copyright © 2011 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

To evaluate health care resource utilization (HRU) in active epilepsy. Thomson-Reuters insurance databases included 14 million persons in 2005–2007. We extracted information for individuals with insurance claims suggestive of epilepsy. Using iterative expert classification, we sorted patients by type of epilepsy. For each type we calculated prevalence and HRU. A distance analysis identified closely similar types, and a principal components analysis revealed dimensions of variation in HRU. The prevalence of active epilepsy was 3.4 per 1,000. Most common diagnoses among 46,847 patients were generalized convulsive epilepsy (33.3%) and complex partial seizures (24.8%). Patients averaged 10 physician visits per year, 24 diagnostic tests/procedures per year, >30 drug dispensings per year, and <1 emergency room (ER) visit per year, the minority of each of these being related to epilepsy. Female patients generally had more HRU, and HRU increased with age. Patients were hospitalized most frequently for disorders other than epilepsy. HRU was similar for most epilepsy types, excepting grand mal status, epilepsia partialis continua, and infantile spasms. The first principal components of HRU variation was nonepilepsy HRU, followed by components of epilepsy-related medications, other epilepsy/emergency care, and epilepsy visits/diagnostic procedures. The prevalence of active epilepsy in the United States is substantially less than the prevalence of any history of recurrent seizure. Nonepilepsy-related HRU dominated HRU in epilepsy patients and was the principal source of variation. There is a core set of epilepsy diagnoses, the HRU patterns of which are indistinguishable, whereas patients with grand mal status, epilepsia partialis continua, and infantile spasms all have distinct patterns. To provide more specific insights into the economic impact of the condition, studies of HRU in epilepsy should make a distinction about epilepsy-related and unrelated care. [ABSTRACT FROM AUTHOR]

When a biological hypothesis of causal effect can be inferred, the hypothesis can sometimes be tested in the selfsame database that gave rise to the study data from which the hypothesis grew. Valid testing happens when the inferred biological hypothesis has scientific implications that predict new relations between observations already recorded. Testing for the existence of the new relations is a valid assessment of the biological hypothesis, so long as the newly predicted relations are not a logical correlate of the observations that stimulated the hypothesis in the first place. These predictions that lead to valid tests might be called 'orthogonal' predictions in the data, and stand in marked contrast to 'scrawny' hypotheses with no biological content, which predict simply that the same data relations will be seen in a new database. The Universal Data Warehouse will shortly render moot searches for new databases in which to test. Copyright © 2010 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

An information-theoretic definition of similarity proposed by Dekang Lin in 1998 appears to have attractive properties for identifying medically similar cases in a large database. Examination of similar cases that appear following vaccination may provide a way of detecting previously unexpected associations. A key element here is the inclusion of shared time-since-vaccination as a dimension of similarity while considering concordance on many other features. The system is potentially much more sensitive to patterns with multiple features than are very large, high-dimension contingency tables. Copyright © 2009 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Task performance while sleep deprived may be moderated by the controlled attention required by the task (Pilcher, Band, Odle-Dusseau, & Muth, 2007). This study examined the effects of 28 h of sleep deprivation on respiratory sinus arrhythmia (RSA) during tasks with low and high controlled attention demands. The results showed that RSA increased throughout the night for both task types, but was consistently reduced during the low compared to high controlled attention tasks. The increase in RSA was linear for the high controlled attention tasks but curvilinear for the low ones. Hence, RSA followed a circadian pattern during the low controlled attention tasks but not the high ones. These results suggest that the effects of sleep deprivation on task performance may be moderated by parasympathetic activity and task type, and this has implications for task assignment during sustained operations that cause sleep deprivation. [ABSTRACT FROM AUTHOR]

Purpose We describe a multi-center post-marketing safety study that uses distributed data methods to minimize the need for covered entities to share protected health information (PHI). Implementation has addressed several issues relevant to creation of a large scale post-marketing drug safety surveillance system envisioned by the FDA's Sentinel Initiative. Methods This retrospective cohort study of Guillain-Barré syndrome (GBS) following meningococcal conjugate vaccination incorporates the data and analytic expertise of five research organizations closely affiliated with US health insurers. The study uses administrative claims data, plus review of full text medical records to adjudicate the status of individuals with a diagnosis code for GBS (ICD9 357.0). A distributed network approach is used to create the analysis files and to perform most aspects of the analysis, allowing nearly all of the data to remain behind institutional firewalls. Pooled analysis files transferred to a central site will contain one record per person for approximately 0.2% of the study population, and contain PHI limited to the month and year of GBS onset for cases or the index date for matched controls. Results The first planned data extraction identified over 9 million eligible adolescents in the target age range of 11-21 years. They contributed an average of 14 months of eligible time on study over 27 months of calendar time. MCV4 vaccination coverage levels exceeded 20% among 17-18-year olds and 16% among 11-13 and 14-16-year-old age groups by the second quarter of 2007. Conclusion This study demonstrates the feasibility of using a distributed data network approach to perform large scale post-marketing safety analyses and is scalable to include additional organizations and data sources. We believe these results can inform the development of a large national surveillance system. Copyright © 2008 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Purpose To assess the relative incidence of peptic ulcer bleeding and perforation (PUBP) in users of selective cyclooxygenase-2 (COX-2) inhibitors and nonselective nonsteroidal anti-inflammatory drugs (NSAIDs). Methods This was a retrospective case-control study based on insurance claims, with chart review to confirm cases. The base population consisted of 2.2 million adult users of celecoxib, diclofenac, ibuprofen, naproxen, rofecoxib, or valdecoxib who were covered by commercial health insurance in the United States from 1999-2003. For 790 persons hospitalized for PUBP and 7887 controls, most recent NSAID/COX-2 use and a variety of medical and pharmacy risk factors were ascertained from the insurance files and analyzed by conditional logistic regression. The main outcome measure was the odds ratio (OR) associated with use of each drug by comparison to naproxen, and in successive intervals after last dispensing. Results Hospitalization for PUBP was highest from the date of NSAID/COX-2 dispensing through the end of an exposure period that corresponded to the days supply; the rate dropped steadily thereafter. The covariate-adjusted ORs by comparison to naproxen were: ibuprofen 0.86 (95% confidence interval (CI) 0.68, 1.09), rofecoxib 0.79 (0.62, 1.02), diclofenac 0.66 (0.47, 0.94), valdecoxib 0.50 (0.26, 0.97), and celecoxib 0.45 (0.35, 0.58). The nonselective NSAIDs had an OR for PUBP of 1.51 (1.26, 1.98) compared to the selective COX-2 inhibitors. Conclusion Nonselective NSAIDs convey half again as much risk of hospitalization for PUBP as do COX-2 inhibitors. Celecoxib users had about half the hospitalization rate of users of naproxen. Copyright © 2008 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Background The risk of coronary heart disease (CHD) in users of antidiabetic agents must be quantified to permit reasoned therapeutic choices. Objectives To assess the risk of myocardial infarction (MI) and coronary revascularization (CR), in diabetic patients who began rosiglitazone, pioglitazone, metformin, or sulfonylureas. Methods We conducted a retrospective cohort study of MI and CR in the PharMetrics database. We performed head-to-head comparisons using propensity-score-stratified Cox proportional hazards models, examining risks both on-treatment and during total follow-up before regimen switches. Results For the combined outcome (MI and CR), the crude rates per 1000 person years were 9 on monotherapy, 13 on dual therapy, and 21 on therapies combined with insulin. In the absence of insulin, regimens containing thiazolidinediones (TZDs) tended toward lower risk than comparable regimens containing sulfonylureas and higher risk than those containing metformin. The summary hazard ratio for rosiglitazone versus pioglitazone was 1.04 (95%CI: 0.94-1.14) for total follow-up and 1.05 (0.92-1.19) for on-treatment time. For MI, the hazard ratios were 1.07 (0.89-1.27) for total follow-up and 1.21 (0.95-1.54) for on-treatment time. Conclusions The present data indicate that the risk of CHD in patients using TZDs appears to lie between the risks associated with sulfonylureas and metformin. Neither the risk of MI and CR together nor the risk of MI alone was significantly different between rosiglitazone and pioglitazone. A nonsignificant observed excess risk of 21% for MI during on-treatment time will require combination with the results of other studies to provide a reliable assessment. Copyright © 2008 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Purpose Residual confounding is a potential limitation of pharmacoepidemiologic studies, and in particular, studies based on administrative claims data that do not capture lifestyle and clinical confounders. We describe an application of the case-cohort design to assess residual confounding by thromboembolic risk factors (e.g., smoking and obesity) not captured in claims data in a claims-based cohort study of thromboembolism among matched oral contraceptive (OC) initiators. Methods This study was conducted using the Ingenix Research Data Mart, a database containing medical claims for approximately 12 million members of a large health plan of the United States. We randomly sampled 701 OC initiators from cohorts of ethinyl estradiol/drospirenone ( n = 22 429) and other OC initiators ( n = 44 858) identified in the years 2001-2004 and matched by propensity score in a claims-based cohort study. Supplementary data on risk factors not measured in the cohort study were collected from medical records for the sample. We estimated the risk ratio of thromboembolism adjusted for the supplementary variables using Cox regression modified for a case-cohort design, and compared it to the rate ratio from the cohort study. Results The risk ratio adjusted for the supplementary variables was 0.90 (95 per cent (%) confidence interval (CI): 0.49, 1.68) which was similar to the rate ratio (0.92; 95%CI: 0.50, 1.63), indicating negligible confounding by the supplementary variables in the cohort study. Conclusions Case-cohort methods were used to assess residual confounding in a claims-based cohort study. This approach adds to a growing number of methods to evaluate residual confounding in cohort studies. Copyright © 2008 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Purpose To refine a preliminary analysis identifying a possibly increased prevalence of malformations among infants born to women exposed to paroxetine in the first trimester. Methods This study used data from UnitedHealthcare, a large U.S. insurer, using datasets originally for a study of bupropion in pregnancy. We identified women with a live-born delivery between January 1995 and September 2004. We classified women according to their first trimester mono- or mono/polytherapy exposure to paroxetine and other antidepressants. We confirmed malformation cases by medical record abstraction. We calculated the adjusted odds ratios (AORs) through logistic regression. Results For paroxetine, there were 815 infants among 791 women exposed as monotherapy, and 1020 infants among 989 women exposed as mono- or polytherapy. For other antidepressants, there were 4198 infants among 4072 women exposed as monotherapy, and 4936 infants among 4767 women exposed as mono- or polytherapy. AORs for all congenital malformations associated with paroxetine were 1.89 (95%CI 1.20-2.98) for monotherapy, and 1.76 (95%CI 1.18-2.64) for mono- or polytherapy. AORs for cardiovascular malformations associated with paroxetine were 1.46 (95%CI 0.74-2.88) for monotherapy, and 1.68 (95%CI 0.95-2.97) for mono- or polytherapy. Conclusions These more detailed paroxetine findings confirm previous findings of analyses of these data among women exposed to all types of antidepressants. The present findings are consistent with other recent results suggesting the possibility of a modestly increased occurrence of congenital malformations following first trimester exposure to paroxetine compared to other antidepressants. Copyright © 2007 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Background There is conflicting evidence on the reduction of cardiovascular risk in diabetic patients treated with oral antidiabetic agents. Objectives To compare the risk of myocardial infarction (MI) and coronary revascularization (CR) in type 2 diabetic patients treated with rosiglitazone, metformin, or sulfonylurea. Methods Using data from a large US insurer, we created propensity-matched cohorts. We identified hospitalizations for MI or CR. We calculated incidence rates and 95% confidence intervals for the outcomes and estimated risks from Cox proportional hazards models. Results We identified 26,931 initiators of monotherapy, 4,086 initiators of dual-therapy, and 2,346 initiators of combination with insulin therapy. There was no difference between the risk of the composite outcome with rosiglitazone monotherapy compared to metformin monotherapy (HR 1.07, 95% CI: 0.85, 1.34), and similarly with rosiglitazone monotherapy compared to sulfonylurea monotherapy (HR 0.82, 95% CI: 0.67, 1.02). There was no difference in the risk of outcome with rosiglitazone in combination with insulin therapy compared to other oral antidiabetic agents in combination with insulin (HR 0.88, 95% CI: 0.59, 1.32). Overall, there was little difference in the risk of the composite outcome or of the individual outcomes of MI and CR comparing rosiglitazone therapies to non-rosiglitazone therapies (HR for composite outcome 0.93, 95% CI: 0.80, 1.10). Conclusions The results from the monotherapy and the dual-therapy comparisons, though not individually significant, are consistent in suggesting that the risk of cardiovascular outcome events in patients using rosiglitazone may lie between the risks associated with sulfonylureas (higher incidence) and metformin (lower incidence). Copyright © 2007 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Purpose: To estimate the rate of new-onset afebrile provoked and unprovoked seizure in a general pediatric population and subgroups of patients with and without psychiatric diagnoses other than attention deficit hyperactivity disorder (ADHD). Methods: A retrospective cohort study of 133,440 pediatric patients, between the ages of 6 and 17 years, and without history of seizure or prior use of anticonvulsant medications, with follow-up during 2003. The data source for this study was Ingenix's research database containing pharmacy and medical claims for members of a large US-based managed care organization. The main outcome measure was new-onset nonfebrile seizure. Incidence rates of seizure and 95% confidence intervals (CI) were calculated and expressed as rates per 100,000 person-years. Results: There were 132 new-onset provoked and unprovoked seizures in 78,423 person-years of follow-up among the general pediatric population sample. The incidence rate of seizure among the general pediatric population was 168 per 100,000 p-y (95% CI 141–200). The incidence rate of seizure among patients without psychiatric diagnoses was 149 per 100,000 p-y (95% CI 122–180). The incidence rate of seizure among patients with psychiatric diagnoses other than ADHD was 513 per 100,000 p-y (95% CI 273–878). There were increases in the incidence rates of seizure in all of the seizure risk factor groups, but this was more pronounced among males ages 6–12 with psychiatric diagnoses. Conclusions: The results of this study are consistent with previous reports showing that pediatric patients with psychiatric disorders have a higher incidence rate of seizure than the general pediatric population. [ABSTRACT FROM AUTHOR]

Purpose Case reports suggest that leukotriene modifier use may be associated with the onset of Churg-Strauss syndrome (CSS). Using pooled data from two nested case-control studies, we examined the association between asthma drug use and the development of CSS. Methods The study was performed in three US managed care organizations and a US national health plan with chart access and complete electronic pharmacy data, with a covered population of 13.9 million. There were 47 cases of possible or definite CSS and 4700 asthma drug user controls identified between January 1, 1995 and December 31, 2002. We examined exposure to asthma drugs in cases and controls, including leukotriene modifiers (6 cases and 202 controls), in the two to 6 months prior to the onset of adjudicated CSS. Results While the crude association between use of leukotriene modifiers and CSS was strong (odds ratio (OR) 4.00, 95% confidence interval (CI): 1.49-10.60), in a multivariable analysis controlling for use of oral corticosteroids, inhaled corticosteroids, and number of categories of asthma drugs dispensed, there was no significant association (OR 1.32, 95% CI: 0.44-3.96). Use of inhaled and oral corticosteroids, evaluated as markers of asthma severity, were associated with CSS (OR 3.07, 95% CI: 1.34-7.03 and OR 5.36, 95% CI: 2.51-11.45, respectively). Conclusions No association was found between CSS and leukotriene modifiers after controlling for asthma drug use However, it is not possible to rule out modest associations with asthma treatments given CSS is so rare and so highly correlated with asthma severity, suggesting further investigation is warranted. Copyright © 2006 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Purpose Reports from the GlaxoSmithKline Bupropion Pregnancy Registry suggested an increase in cardiovascular defects following exposure to bupropion during pregnancy. We conducted a study of congenital malformations among infants born to women exposed to bupropion during their first trimester. Methods The study used data from UnitedHealthcare between January 1995 and September 2004. We calculated the prevalence of all congenital malformations and cardiovascular malformations associated with bupropion exposure in the estimated first trimester (1213 infants), compared with (1) other antidepressant exposure in the first trimester (4743 infants) and (2) bupropion exposure outside the first trimester (1049 infants). Malformation cases were confirmed through medical record abstraction. We calculated adjusted odds ratios (AORs) using the GEE form of logistic regression. Results For all congenital malformations, the prevalence associated with bupropion first trimester was 23.1 per 1000 infants. The AORs were 0.95 (95%CI 0.62-1.45) and 1.00 (95%CI 0.57-1.73) in comparison to other antidepressants (prevalence 23.2 per 1000) and bupropion outside the first trimester (prevalence 21.9 per 1000), respectively. For cardiovascular malformations, the prevalence associated with bupropion first trimester was 10.7 per 1000 infants. The AORs were 0.97 (95%CI 0.52-1.80) and 1.07 (95%CI 0.48-2.40) in comparison to other antidepressants (prevalence 10.8 per 1000) and bupropion outside the first trimester (prevalence 9.5 per 1000), respectively. Conclusions Results do not support a hypothesis of a teratogenic effect of first trimester bupropion exposure. The prevalence of malformations associated with bupropion exposure in the first trimester was not increased relative to the comparison groups. Copyright © 2006 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Purpose Abacavir is associated with an infrequent but potentially serious hypersensitivity reaction (HSR) that can include a wide range of signs and symptoms. Identification of this reaction through medical insurance claims could provide a simple and efficient means of monitoring the incidence of abacavir hypersensitivity in large populations of patients. Methods Using data from a safety study of 948 abacavir users with 22 hypersensitivity events identified from claims and validated through medical record review, we used a recursive partitioning analysis to construct an algorithm to differentiate between patients with and without validated adverse events. Bootstrap resampling techniques provided validation for the analysis. Results The analysis produced a classification tree with three decision nodes that comprised the best indicators of HSRs. The predictors included any one of several specific symptoms commonly found with this reaction, a claims diagnosis of adverse effect of drug, anaphylactic shock or unspecified allergy, and a discontinuation in abacavir prior to completing a 90-day course of therapy. The algorithm demonstrated 95% sensitivity and 90% specificity when tested using a bootstrap resampling approach with the current data. Conclusions A sensitive and specific algorithm for identifying abacavir hypersensitivity from claims was created. This algorithm would permit efficient identification of charts for medical review. Further testing of the algorithm with additional medical claims data for abacavir users will be required to ascertain its validity across databases. Copyright © 2007 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Background Case reports and observational studies have implicated fluoroquinolone antibiotic exposure as a risk factor for Achilles tendon rupture (ATR), an uncommon condition for which there are few formal studies. We sought to quantify the strength of association between exposure to fluoroquinolone antibiotics and the occurrence of ATR, accounting for other risk factors. Methods This was a case-control study nested within a health insurer cohort. Cases of ATR were identified and confirmed using patterns of health insurance claims that were validated through sampled medical record review. Information on risk factors, including fluoroquinolone exposure, came from health insurance claims. Results There were 947 cases of ATR and 18 940 controls. A dispensing of a fluoroquinolone antibiotic in the past 6 months was more common among ATR cases than controls, although not significantly so (odds ratio (OR) = 1.2; 95% confidence interval (CI) = 0.9-1.7), and exposure to a higher cumulative fluoroquinolone dose was more strongly associated (OR = 1.5, 95%CI = 1.0-2.3). Other risk factors for ATR were trauma (OR = 17.2, 95%CI = 14.0-20.2), male sex (OR = 3.0, 95%CI = 2.6-3.5), injected corticosteroid administration (OR = 2.2, 95%CI = 1.6-2.9), obesity (OR = 2.0, 95%CI = 1.2-3.1), rheumatoid arthritis (OR = 1.9, 95%CI = 1.0-3.7), skin or soft tissue infections (OR = 1.5, 95%CI = 0.9-2.3), oral corticosteroids (OR = 1.4, 95%CI = 1.0-1.8), and non-fluoroquinolone antibiotics (OR = 1.2, 95%CI = 1.1-1.5). Conclusions The elevation in ATR risk associated with fluoroquinolones was similar in magnitude to that associated with oral corticosteroids or non-fluoroquinolone antibiotics. Trauma and male sex were more strongly associated with ATR, as were obesity and injected corticosteroids. Copyright © 2006 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Purpose The purpose of this study was to estimate rates of acute coronary syndrome (ACS), and sudden cardiac death in relation to use of rofecoxib, celecoxib, naproxen, diclofenac, and ibuprofen. Methods We conducted a retrospective cohort study among 424 584 health plan enrollees ages 40-64 who used non-aspirin non-steroidal anti-inflammatory drugs (NANSAIDs) by prescription from 1999 to 2001. We used automated medical and pharmacy claims data to compute person-time exposed to study drugs, and to identify hospitalizations for ACS. The primary endpoint was ACS, inclusive of myocardial infarction (MI), unstable angina, and sudden cardiac death as identified through claims data and confirmed through hospital medical records, or through the National Death Index. Results Compared with ibuprofen or diclofenac use, the relative risk (RR) of ACS during periods of current rofecoxib use was 1.35 (95%CI 1.09-1.68). For current use of celecoxib, the RR was 1.03 (95%CI 0.83-1.27). Risks in the first 30 days of rofecoxib and celecoxib use were modestly elevated. Rofecoxib use at the 25 mg/day (modal) dose was associated with an elevated risk of ACS (RR 1.54, 95%CI 1.15-2.04), while use at 26-50 mg/day (>modal) was not (RR 0.81, 95%CI 0.41-1.60). There were no increased risks with modal or greater than modal doses of celecoxib or naproxen compared with all doses of ibuprofen or diclofenac combined. Conclusions The incidence of ACS was 1.35 times greater during rofecoxib use than use of ibuprofen or diclofenac. No statistically significant elevation in risk was observed with celecoxib use. Naproxen use was not associated with risk of ACS. Copyright © 2005 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Study Objective. To measure the association among prescription copayment, drug adherence, and subsequent health outcomes among patients with chronic heart failure (CHF). Design. Retrospective cohort study. Data Source. Database of a large, national health insurance plan. Patients. Patients with CHF receiving commercial and Medicare supplemental benefits. Measurements and Main Results. We estimated adherence to therapy with β-blockers or angiotensin-converting enzyme (ACE) inhibitors in 2002 by using the medication possession ratio, an estimate of the proportion of days a patient was exposed to a drug while taking a drug regimen. For 2003, we measured the annualized total cost of health care and identified hospitalizations with a diagnostic code for CHF. We used a two-stage regression approach to model the association among copayment, adherence, and patient outcomes. For patients taking ACE inhibitors, a $10 increase in copayment was associated with a 2.6% decrease in the medication possession ratio (95% confidence interval [CI] 2.0-3.1%). This change in adherence was associated with a predicted 0.8% decrease in medical costs (95% CI -4.2-2.5%) but a predicted 6.1% increase in the risk of hospitalization for CHF (95% C1 0.5-12.0%). Among patients taking β-blockers, a $10 increase in copayment was associated with a 1.8% decrease in the medication possession ratio (95% C1 1.4-2.2%). This change in adherence was associated with a predicted 2.8% decrease in medical costs (95% Cl -5.9-0.1%) and a predicted 8.7% increase in the risk of hospitalization for CHF (95% CI 3.8-13.8%). Conclusion. Among patients with CHF, higher drug copayments were associated with poorer adherence. The change was relatively small and did not affect predicted total health care costs, but it was sufficient to increase the predicted risk of hospitalization for CHF. [ABSTRACT FROM AUTHOR]

Purpose The purpose of this study was to compare incidence rates of hospitalization associated with rhabdomyolysis, myopathy, renal, or hepatic dysfunction, and of in-hospital death, between initiators of rosuvastatin and other statins. Methods This was a matched cohort study of statin initiators from the administrative database of a large health insurer in the US, during the first 6 months of rosuvastatin availability with up to 18 months of follow-up. All outcome events were verified by medical record review. Incidence rates, risk ratios, and associated 95% confidence intervals were estimated. Results From an initial pool of 12 217, 11 249 eligible rosuvastatin initiators were matched to 37 282 initiators of other statins. The incidence rate (IR) per 1000 person-years for rhabdomyolysis was 0.10 [0.00, 0.55] for rosuvastatin initiators (n = 1) and 0.06 [0.01, 0.22] for other statin initiators (n = 2), for a hazard ratio (HR) of 1.98 [0.18, 21.90]. The IR for myopathy was 0.20 [0.02, 0.71] for rosuvastatin initiators (n = 2) and 0.00 [0.00, 0.09] for other statin initiators (n = 0). The IR for renal dysfunction was 1.18 [0.61, 2.06] for rosuvastatin initiators (n = 12) and 1.26 [0.91, 1.71] for other statin initiators (n = 42), for a HR of 0.90 [0.47, 1.73]. The IR for hepatic dysfunction was 0.20 (0.02, 0.71) for rosuvastatin initiators (n = 2) and 0.24 (0.10, 0.47) for other statin initiators (n = 8), for a HR of 0.87 (0.18, 4.14). Conclusions This study found no difference between rosuvastatin and the other statins in the incidence of hospitalizations associated with renal or hepatic events, or death. The absolute incidence rates of rhabdomyolysis and myopathy were reassuringly low among all statin initiators but remain too small for firm conclusions to be drawn on any difference between the statins. Copyright © 2006 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Results from clinical trials and clinical practice have shown statins to be generally well tolerated with a low frequency of clinically relevant side effects. Nevertheless, there are rare occasions when adverse events (AEs), sometimes serious, may occur. Rosuvastatin is the newest statin to be approved in the USA and many other countries. As part of the continued assessment of the benefit-risk profile of rosuvastatin, AstraZeneca has developed a progressive, comprehensive pharmacoepidemiology programme to complement safety data obtained from randomised clinical trials and spontaneous reporting systems, which have demonstrated that rosuvastatin has a safety profile in line with comparator statins. This programme comprises nine studies conducted in recognised centres of excellence assessing over 50 000 patients treated with rosuvastatin. It consists of three components: patient characteristics studies (four studies), safety evaluation studies (four studies); and review of data generated from the Prescription-Event Monitoring (PEM) study, designed and run by an independent third party. Patient characteristics studies are designed to describe the characteristics and drug utilisation patterns of new users of rosuvastatin compared with new users of other statins in automated databases. Safety evaluation studies will examine the rates of specific AEs in different cohorts of statin users and determine risk factors for these events using data recorded prospectively in automated databases with case adjudication via medical record review. The independent PEM study will monitor any significant events recorded by general practitioners since starting rosuvastatin treatment. This article is an overview of the rationale and methodology of the rosuvastatin pharmacoepidemiology programme. Copyright © 2006 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Purpose A challenge in the use of insurance claims databases for epidemiologic research is accurate identification and verification of medical conditions. This report describes the development and validation of claims-based algorithms to identify colonic ischemia, hospitalized complications of constipation, and irritable bowel syndrome (IBS). Methods From the research claims databases of a large healthcare company, we selected at random 120 potential cases of IBS and 59 potential cases each of colonic ischemia and hospitalized complications of constipation. We sought the written medical records and were able to abstract 107, 57, and 51 records, respectively. We established a 'true' case status for each subject by applying standard clinical criteria to the available chart data. Comparing the insurance claims histories to the assigned case status, we iteratively developed, tested, and refined claims-based algorithms that would capture the diagnoses obtained from the medical records. We set goals of high specificity for colonic ischemia and hospitalized complications of constipation, and high sensitivity for IBS. Results The resulting algorithms substantially improved on the accuracy achievable from a naïve acceptance of the diagnostic codes attached to insurance claims. The specificities for colonic ischemia and serious complications of constipation were 87.2 and 92.7%, respectively, and the sensitivity for IBS was 98.9%. Conclusions U.S. commercial insurance claims data appear to be usable for the study of colonic ischemia, IBS, and serious complications of constipation. Copyright © 2005 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Background Confounding by indication is a common problem in pharmacoepidemiology, where predictors of treatment also have prognostic value for the outcome of interest. The tools available to the epidemiologist that can be used to mitigate the effects of confounding by indication often have limits with respect to the number of variables that can be simultaneously incorporated as components of the confounding. This constraint becomes particularly apparent in the context of a rich data source (such as administrative claims data), applied to the study of an outcome that occurs infrequently. In such settings, there will typically be many more variables available for control as potential confounders than traditional epidemiologic techniques will allow. Methods One tool that can indirectly permit control of a large number of variables is the propensity score approach. This paper illustrates the application of the propensity score to a study conducted in an administrative database, and raises critical issues to be addressed in such an analysis. In this example, the effect of statin therapy on the occurrence of myocardial infarction was examined, and numerous potential confounders of this association were adjusted simultaneously using a propensity score to form matched cohorts of statin initiators and non-initiators. Results The incidence of myocardial infarction observed in the statin treated cohort was lower than the incidence in the untreated cohort, and the magnitude of this effect was consistent with results from randomized placebo controlled clinical trials of statin therapy. Conclusions This example illustrates how confounding by indication can be mitigated by the propensity score matching technique. Concerns remain over the generalizability of estimates obtained from such a study, and how to know when propensity scores are removing bias, since apparent balance between compared groups on measured variables could leave variables not included in the propensity score unbalanced and lead to confounded effect estimates. Copyright © 2005 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

Objective.—To estimate rates of vascular events in relation to dispensing of triptans and ergot alkaloids among migraineurs, and to compare these rates with those of nonmigraineurs. Context. —It has been speculated that the use of triptans or ergot alkaloid drugs might increase risk of ischemic events through vasoconstriction. Design.—A retrospective cohort study of 130,411 migraineurs and 130,411 age-, sex-, and health plan-matched nonmigraineurs who were members of UnitedHealthcare during 1995 through 1999. The data source for this study was Ingenix's research database containing pharmacy and medical claims for UnitedHealthcare members, and the National Death Index. Main Outcome Measures .—Incidence of cardiovascular and cerebrovascular events and mortality. Results.—Migraineurs and nonmigraineurs had identical rates of myocardial infarction: 1.4 per 1000 person-years. Migraineurs were 67% more likely to suffer a stroke than nonmigraineurs (adjusted relative risk [RR] 1.67, 95% confidence interval [CI] 1.31–2.13), and had higher rates of unstable angina and transient ischemic attacks. There was no increase in risk of myocardial infarction with current (adjusted RR 0.80, 95% CI 0.58–1.11) or recent (adjusted RR 1.15, 95% CI 0.71–1.87) triptan use. Neither current (adjusted RR 0.90, 95% CI 0.64–1.26) nor recent (adjusted RR 0.84, 95% CI 0.46–1.55) triptan use was associated with risk of stroke. Current users of ergot alkaloids were somewhat more likely to have a stroke than other migraineurs (adjusted RR 1.49, 95% CI 0.93–2.41), but there was no dose-response relationship. Conclusions.—Use of triptans is not associated with increased risk of any ischemic events, including myocardial infarction and stroke, or mortality. Consistent with previous studies, migraineurs in general have an elevated risk of stroke, but not myocardial infarction, compared with nonmigraineurs. [ABSTRACT FROM AUTHOR]

Introduction: In view of current international interest and research on the elderly, numerous questions present themselves. Examples are: how rapidly are populations ageing? Is physical and functional decline in old age inevitable? Can morbidity be compressed into the final years of life? Are social, medical and other resources available for coping with the increasing numbers of elderly? Not least, how do old people themselves feel about ageing? [ABSTRACT FROM AUTHOR]

DMFT scores, total sugar intakes and snack habits were determined in 1918 South African Black pupils (923 rural, 995 urban) and 724 White pupils (English and Afrikaans speaking) aged 16-18 years inclusive. Mean DMFT scores of school groups of Black pupils (both sexes) ranged from 0.9 and 2.0 in rural areas, to 4.2-6.7 in urban areas (where data were far higher than such obtained 7 years ago), and were 9.2 and 10.2 for White pupils. Corresponding mean daily sugar intakes were - rural Blacks, 69 g and 97 g, urban Blacks 118-141 g, respectively (all slightly higher than previously); and Whites 102 g and 123 g. Mean DMFT scores of girls were higher than those of boys; yet while mean sugar intakes of Black girls and boys were somewhat similar, White girls' intakes were much lower than those of White boys. In the ethnic-sex groups studied, mean DMFT scores for pupils in upper, compared with lower, third of sugar intake, were higher in nine of the 14 sub-groups. Mean DMFT scores in upper, compared with lower, third of exposure to snack practices were higher in 11 of the 14 groups. However, most of the differences were slight. In further research, more intensive enquiries should be pursued on roles of ethnic and familial factors, as well as on roles of inter-acting dietary components additional to sugar and sugar-containing foods. [ABSTRACT FROM AUTHOR]