Semiparametric Allocation of Subjects to Cohort Strata.

BACKGROUND: We illustrate a method for stratum assignment in small cohort studies that avoids modeling assumptions. METHODS: Off-the-shelf software ( rgenoud ) made stratum assignments to minimize a loss function built on within-stratum and population-adjusted Euclidean distances. RESULTS: In 100 trials using simulated data of 300 records with a binary treatment and four dissimilar covariate treatment predictors, minimizing a loss based on Euclidean distance reduced covariate imbalance by a median of 99%. Stratification by propensity score and weighting records by the inverse of their probability of treatment reduced imbalance by 76%-89% and 83%-94%, respectively. Loss minimization applied to a cohort of 361 children undergoing immunotherapy achieved nearly complete elimination of covariate differences for important treatment predictors. CONCLUSION: With the availability of semiparametric stratum-assignment algorithms, analysts can tailor loss functions to meet design goals. Here, a loss function that emphasized covariate balance performed well under limited testing.

Stratified analysis in comparative effectiveness studies that emulate randomized trials.

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.

The acute effects of azithromycin use on cardiovascular mortality as compared with amoxicillin-clavulanate in US Veterans.

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.

Long-term risk of hepatocellular carcinoma mortality in 23220 hospitalized patients treated with micafungin or other parenteral antifungals.

BACKGROUND: Liver tumours observed in rats exposed to micafungin led to a black box warning upon approval in Europe in 2008. Micafungin's risk for liver carcinogenicity in humans has not been investigated. We sought to describe the risk of fatal hepatocellular carcinoma (HCC) among persons who received micafungin and other parenteral antifungals (PAFs) with up to 12 years of follow-up. METHODS: We assembled a US multicentre cohort of hospitalized patients who received micafungin or other PAFs between 2005 and 2012. We used propensity score (PS) matching on patient characteristics from electronic medical records to compare rates of HCC mortality identified through the National Death Index though to the end of December 2016. We computed HRs and 95% CIs. RESULTS: A total of 40110 patients who received a PAF were identified; 6903 micafungin recipients (87% of those identified) were successfully matched to 16317 comparator PAF users. Ten incident HCC deaths, one in the micafungin-exposed group and nine among comparator PAF users, occurred in 71285 person-years of follow-up. The HCC mortality rate was 0.05 per 1000 person-years in micafungin patients and 0.17 per 1000 person-years in comparator PAF patients. The PS-matched HR for micafungin versus comparator PAF was 0.29 (95% CI 0.04-2.24). CONCLUSIONS: Both micafungin and comparator PAFs were associated with HCC mortality rates of <0.2 per 1000 person-years. Given the very low event rates, any potential risk for HCC should not play a role in clinical decisions regarding treatment with micafungin or other PAFs investigated in this study.

Structural and Functional Characterization of Dynamic Oligomerization in Burkholderia cenocepacia HMG-CoA Reductase.

The enzyme 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase (HMGR), in most organisms, catalyzes the four-electron reduction of the thioester (S)-HMG-CoA to the primary alcohol (R)-mevalonate, utilizing NADPH as the hydride donor. In some organisms, including the opportunistic lung pathogen Burkholderia cenocepacia, it catalyzes the reverse reaction, utilizing NAD+ as a hydride acceptor in the oxidation of mevalonate. B. cenocepacia HMGR has been previously shown to exist as an ensemble of multiple non-additive oligomeric states, each with different levels of enzymatic activity, suggesting that the enzyme exhibits characteristics of the morpheein model of allostery. We have characterized a number of factors, including pH, substrate concentration, and enzyme concentration, that modulate the structural transitions that influence the interconversion among the multiple oligomers. We have also determined the crystal structure of B. cenocepacia HMGR in the hexameric state bound to coenzyme A and ADP. This hexameric assembly provides important clues about how the transition among oligomers might occur, and why B. cenocepacia HMGR, unique among characterized HMGRs, exhibits morpheein-like behavior.

Structural and Biochemical Characterization of Cinnamoyl-CoA Reductases.

Cinnamoyl-coenzyme A reductase (CCR) catalyzes the reduction of hydroxycinnamoyl-coenzyme A (CoA) esters using NADPH to produce hydroxycinnamyl aldehyde precursors in lignin synthesis. The catalytic mechanism and substrate specificity of cinnamoyl-CoA reductases from sorghum (Sorghum bicolor), a strategic plant for bioenergy production, were deduced from crystal structures, site-directed mutagenesis, and kinetic and thermodynamic analyses. Although SbCCR1 displayed higher affinity for caffeoyl-CoA or p-coumaroyl-CoA than for feruloyl-CoA, the enzyme showed significantly higher activity for the latter substrate. Through molecular docking and comparisons between the crystal structures of the Vitis vinifera dihydroflavonol reductase and SbCCR1, residues threonine-154 and tyrosine-310 were pinpointed as being involved in binding CoA-conjugated phenylpropanoids. Threonine-154 of SbCCR1 and other CCRs likely confers strong substrate specificity for feruloyl-CoA over other cinnamoyl-CoA thioesters, and the T154Y mutation in SbCCR1 led to broader substrate specificity and faster turnover. Through data mining using our structural and biochemical information, four additional putative CCR genes were discovered from sorghum genomic data. One of these, SbCCR2, displayed greater activity toward p-coumaroyl-CoA than did SbCCR1, which could imply a role in the synthesis of defense-related lignin. Taken together, these findings provide knowledge about critical residues and substrate preference among CCRs and provide, to our knowledge, the first three-dimensional structure information for a CCR from a monocot species.

The Enzyme Activity and Substrate Specificity of Two Major Cinnamyl Alcohol Dehydrogenases in Sorghum (Sorghum bicolor), SbCAD2 and SbCAD4.

Cinnamyl alcohol dehydrogenase (CAD) catalyzes the final step in monolignol biosynthesis, reducing sinapaldehyde, coniferaldehyde, and p-coumaraldehyde to their corresponding alcohols in an NADPH-dependent manner. Because of its terminal location in monolignol biosynthesis, the variation in substrate specificity and activity of CAD can result in significant changes in overall composition and amount of lignin. Our in-depth characterization of two major CAD isoforms, SbCAD2 (Brown midrib 6 [bmr6]) and SbCAD4, in lignifying tissues of sorghum (Sorghum bicolor), a strategic plant for generating renewable chemicals and fuels, indicates their similarity in both structure and activity to Arabidopsis (Arabidopsis thaliana) CAD5 and Populus tremuloides sinapyl alcohol dehydrogenase, respectively. This first crystal structure of a monocot CAD combined with enzyme kinetic data and a catalytic model supported by site-directed mutagenesis allows full comparison with dicot CADs and elucidates the potential signature sequence for their substrate specificity and activity. The L119W/G301F-SbCAD4 double mutant displayed its substrate preference in the order coniferaldehyde > p-coumaraldehyde > sinapaldehyde, with higher catalytic efficiency than that of both wild-type SbCAD4 and SbCAD2. As SbCAD4 is the only major CAD isoform in bmr6 mutants, replacing SbCAD4 with L119W/G301F-SbCAD4 in bmr6 plants could produce a phenotype that is more amenable to biomass processing.

The Structure and Catalytic Mechanism of Sorghum bicolor Caffeoyl-CoA O-Methyltransferase.

Caffeoyl-coenzyme A 3-O-methyltransferase (CCoAOMT) is an S-adenosyl methionine (SAM)-dependent O-methyltransferase responsible for methylation of the meta-hydroxyl group of caffeoyl-coenzyme A (CoA) on the pathway to monolignols, with their ring methoxylation status characteristic of guaiacyl or syringyl units in lignin. In order to better understand the unique class of type 2 O-methyltransferases from monocots, we have characterized CCoAOMT from sorghum (Sorghum bicolor; SbCCoAOMT), including the SAM binary complex crystal structure and steady-state enzyme kinetics. Key amino acid residues were validated with site-directed mutagenesis. Isothermal titration calorimetry data indicated a sequential binding mechanism for SbCCoAOMT, wherein SAM binds prior to caffeoyl-CoA, and the enzyme showed allosteric behavior with respect to it. 5-Hydroxyferuloyl-CoA was not a substrate for SbCCoAOMT. We propose a catalytic mechanism in which lysine-180 acts as a catalytic base and deprotonates the reactive hydroxyl group of caffeoyl-CoA. This deprotonation is facilitated by the coordination of the reactive hydroxyl group by Ca(2+) in the active site, lowering the pKa of the 3'-OH group. Collectively, these data give a new perspective on the catalytic mechanism of CCoAOMTs and provide a basis for the functional diversity exhibited by type 2 plant OMTs that contain a unique insertion loop (residues 208-231) conferring affinity for phenylpropanoid-CoA thioesters. The structural model of SbCCoAOMT can serve as the basis for protein engineering approaches to enhance the nutritional, agronomic, and industrially relevant properties of sorghum.

Short-term risk of liver and renal injury in hospitalized patients using micafungin: a multicentre cohort study.

BACKGROUND: Although echinocandins are generally well tolerated, there is little information on the frequency with which renal and hepatic adverse effects occur during use of micafungin or other parenteral antifungal (PAF) agents in clinical practice. METHODS: MYCOS is a multicentre cohort study of adult and paediatric patients who received micafungin or other PAFs between 2005 and 2012 at seven tertiary care hospitals from six centres in the USA. PAF cohort controls were selected through propensity score (PS) matching to micafungin recipients using clinical characteristics, other treatments, procedures and hospital service where PAF treatment was initiated. Analysis was restricted to patients without chronic liver and kidney conditions at the time of cohort entry. Treatment-emergent hepatic and renal injury was documented by changes in liver enzymes or estimated glomerular filtration rate through 30 days following completion of PAF treatment. Comparisons were quantified using the HR from a proportional hazards analysis. RESULTS: There were 2970 micafungin recipients PS matched to 6726 recipients of comparator PAFs. Balance was achieved in all baseline covariates between treatment groups. There were similar rates of hepatic injury (micafungin, 13 events per 100 patients and other PAF, 12 per 100; HR = 0.99; 95% CI 0.86-1.14) and lower rates of renal injury (micafungin, 63 events per 100 patients and other PAF, 65 per 100; HR = 0.93; 95% CI 0.87-0.99) for micafungin recipients versus PAF comparators. CONCLUSION: For a wide spectrum of underlying conditions, we observed no increase in liver injury by micafungin and possibly a reduced risk of renal dysfunction in comparison with other PAF medications.

Design and analysis choices for safety surveillance evaluations need to be tuned to the specifics of the hypothesized drug-outcome association.

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.

Acute nonarteritic anterior ischemic optic neuropathy and exposure to phosphodiesterase type 5 inhibitors.

INTRODUCTION: Nonarteritic anterior ischemic optic neuropathy (NAION), a rare visual disorder, has been reported in men using phosphodiesterase type 5 inhibitors (PDE5i) for erectile dysfunction. AIM: We examined whether intermittent use of PDE5i is associated with acute NAION onset within approximately five half-lives following drug ingestion. METHODS: One hundred two ophthalmology centers in the United States and Europe identified potential cases of NAION. An expert adjudication committee conducted a blind review of the records of those with recent PDE5i use to classify cases as Definite, Possible, or not NAION. Subjects provided information on PDEi use via telephone interview. Each NAION case's PDE5i exposure immediately prior to onset was compared against his recent patterns of use in an observational case-crossover design. A sample size of 40 cases with intermittent PDE5i exposure in the 30 days prior to NAION onset was needed to detect an odds ratio (OR) of 3.0 with 80% power. MAIN OUTCOME MEASURES: The daily relative risk for acute NAION on days within five half-lives of PDE5i use vs. other days was estimated via an OR obtained from conditional logistic regression. RESULTS: Among 43 Definite NAION cases with PDE5i exposure in the prior 30 days, the OR was 2.15 (95% confidence interval [CI]: 1.06, 4.34). When 21 Possible NAION cases were included (n = 64), the OR was 2.36 (95% CI: 1.33, 4.19). CONCLUSIONS: We found an approximately twofold increased risk of acute NAION within five half-lives of PDE5i use compared with use in a more prior time period. Bias from inaccurate recall of exposure was unlikely to have substantially affected the results. Based on our results, we estimate that weekly use of PDE5i adds three NAION cases per 100,000 men 50 years and older annually.

Identifying health outcomes in healthcare databases.

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.

Elucidation of the structure and reaction mechanism of sorghum hydroxycinnamoyltransferase and its structural relationship to other coenzyme a-dependent transferases and synthases.

Hydroxycinnamoyltransferase (HCT) from sorghum (Sorghum bicolor) participates in an early step of the phenylpropanoid pathway, exchanging coenzyme A (CoA) esterified to p-coumaric acid with shikimic or quinic acid as intermediates in the biosynthesis of the monolignols coniferyl alcohol and sinapyl alcohol. In order to elucidate the mode of action of this enzyme, we have determined the crystal structures of SbHCT in its apo-form and ternary complex with shikimate and p-coumaroyl-CoA, which was converted to its product during crystal soaking. The structure revealed the roles of threonine-36, serine-38, tyrosine-40, histidine-162, arginine-371, and threonine-384 in catalysis and specificity. Based on the exact chemistry of p-coumaroyl-CoA and shikimic acid in the active site and an analysis of kinetic and thermodynamic data of the wild type and mutants, we propose a role for histidine-162 and threonine-36 in the catalytic mechanism of HCT. Considering the calorimetric data, substrate binding of SbHCT should occur sequentially, with p-coumaroyl-CoA binding prior to the acyl acceptor molecule. While some HCTs can use both shikimate and quinate as an acyl acceptor, SbHCT displays low activity toward quinate. Comparison of the structure of sorghum HCT with the HCT involved in chlorogenic acid synthesis in coffee (Coffea canephora) revealed many shared features. Taken together, these observations explain how CoA-dependent transferases with similar structural features can participate in different biochemical pathways across species.

Cardiac mortality in users of olmesartan, other angiotensin-receptor blockers and angiotensin-converting enzyme inhibitors.

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.

Characterizing vaccine-associated risks using cubic smoothing splines.

Estimating risks associated with the use of childhood vaccines is challenging. The authors propose a new approach for studying short-term vaccine-related risks. The method uses a cubic smoothing spline to flexibly estimate the daily risk of an event after vaccination. The predicted incidence rates from the spline regression are then compared with the expected rates under a log-linear trend that excludes the days surrounding vaccination. The 2 models are then used to estimate the excess cumulative incidence attributable to the vaccination during the 42-day period after vaccination. Confidence intervals are obtained using a model-based bootstrap procedure. The method is applied to a study of known effects (positive controls) and expected noneffects (negative controls) of the measles, mumps, and rubella and measles, mumps, rubella, and varicella vaccines among children who are 1 year of age. The splines revealed well-resolved spikes in fever, rash, and adenopathy diagnoses, with the maximum incidence occurring between 9 and 11 days after vaccination. For the negative control outcomes, the spline model yielded a predicted incidence more consistent with the modeled day-specific risks, although there was evidence of increased risk of diagnoses of congenital malformations after vaccination, possibly because of a "provider visit effect." The proposed approach may be useful for vaccine safety surveillance.

Active safety monitoring of new medical products using electronic healthcare data: selecting alerting rules.

BACKGROUND: Active medical-product-safety surveillance systems are being developed to monitor many products and outcomes simultaneously in routinely collected longitudinal electronic healthcare data. These systems will rely on algorithms to generate alerts about potential safety concerns. METHODS: We compared the performance of 5 classes of algorithms in simulated data using a sequential matched-cohort framework, and applied the results to 2 electronic healthcare databases to replicate monitoring of cerivastatin-induced rhabdomyolysis. We generated 600,000 simulated scenarios with varying expected event frequency in the unexposed, alerting threshold, and outcome risk in the exposed, and compared the alerting algorithms in each scenario type using an event-based performance metric. RESULTS: We observed substantial variation in algorithm performance across the groups of scenarios. Relative performance varied by the event frequency and by user-defined preferences for sensitivity versus specificity. Type I error-based statistical testing procedures achieved higher event-based performance than other approaches in scenarios with few events, whereas statistical process control and disproportionality measures performed relatively better with frequent events. In the empirical data, we observed 6 cases of rhabdomyolysis among 4294 person-years of follow-up, with all events occurring among cerivastatin-treated patients. All selected algorithms generated alerts before the drug was withdrawn from the market. CONCLUSIONS: For active medical-product-safety monitoring in a sequential matched cohort framework, no single algorithm performed best in all scenarios. Alerting algorithm selection should be tailored to particular features of a product-outcome pair, including the expected event frequencies and trade-offs between false-positive and false-negative alerting.

The Zelnorm epidemiologic study (ZEST): a cohort study evaluating incidence of abdominal and pelvic surgery related to tegaserod treatment.

BACKGROUND: Pre-marketing clinical studies of tegaserod suggested an increased risk of abdominal surgery, particularly cholecystectomy. We sought to quantify the association between tegaserod use and the occurrence of abdominal or pelvic surgery, including cholecystectomy. METHODS: This cohort study was conducted within an insured population. Tegaserod initiators and similar persons who did not initiate tegaserod were followed for up to six months for the occurrence of abdominal or pelvic surgery. Surgical procedures were identified from health insurance claims validated by review of medical records. The incidence of confirmed outcomes was compared using both as-matched and as-treated analyses. RESULTS: Among 2,762 tegaserod initiators, there were 94 abdominal or pelvic surgeries (36 gallbladder): among 2,762 comparators there were 134 abdominal or pelvic surgeries (37 gallbladder) (hazard ratio HR] = 0.70, 95% confidence interval [C.I.] = 0.54-0.91 overall, HR = 0.98, 95% C.I. = 0.62-1.55 for gallbladder). Current tegaserod exposure compared to nonexposure was associated with a rate ratio [RR] of 0.68 (95% C.I. = 0.48-0.95) overall, while the RR was 0.99 (95% C.I. = 0.56-1.77) for gallbladder surgery. CONCLUSIONS: In this study, tegaserod use was not found to increase the risk of abdominal or pelvic surgery nor the specific subset of gallbladder surgery.

An event-based approach for comparing the performance of methods for prospective medical product monitoring.

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.