Search

To develop a claims-based model to predict persistent high-dose opioid use among patients undergoing total knee replacement (TKR).Using Medicare claims (2010-2014), we identified patients ages ≥65 years who underwent TKR with no history of high-dose opioid use (mean 25 morphine milligram equivalents [MMEs]/day) in the year prior to TKR. We used group-based trajectory modeling to identify distinct opioid use patterns. The primary outcome was persistent high-dose opioid use in the year after TKR. We split the data into training (2010-2013) and test (2014) sets and used logistic regression with least absolute shrinkage and selection operator regularization, utilizing a total of 83 preoperative patient characteristics as candidate predictors. A reduced model with 10 prespecified variables, which included demographic characteristics, opioid use, and medication history was also considered.The final study cohort included 142,089 patients who underwent TKR. The group-based trajectory model identified 4 distinct trajectories of opioid use (group 1: short-term, low-dose; group 2: moderate-duration, low-dose; group 3: moderate-duration, high-dose; and group 4: persistent high-dose). The model predicting persistent high-dose opioid use achieved high discrimination (receiver operating characteristic area under the curve [AUC] 0.85 [95% confidence interval (95% CI) 0.84-0.86]) in the test set. The reduced model with 10 predictors performed equally well (AUC 0.84 [95% CI 0.84-0.85]).In this cohort of older patients, 10.6% became persistent high-dose (mean 22.4 MME/day) opioid users after TKR. Our model with 10 readily available clinical factors may help identify patients at high risk of future adverse outcomes from persistent opioid use after TKR.

The high-dimensional propensity score (HDPS) is a semi-automated procedure for confounder identification, prioritisation and adjustment in large healthcare databases that requires investigators to specify data dimensions, prioritisation strategy and tuning parameters. In practice, reporting of these decisions is inconsistent and this can undermine the transparency, and reproducibility of results obtained. We illustrate reporting tools, graphical displays and sensitivity analyses to increase transparency and facilitate evaluation of the robustness of analyses involving HDPS.Using a study from the UK Clinical Practice Research Datalink that implemented HDPS we demonstrate the application of the proposed recommendations.We identify seven considerations surrounding the implementation of HDPS, such as the identification of data dimensions, method for code prioritisation and number of variables selected. Graphical diagnostic tools include assessing the balance of key confounders before and after adjusting for empirically selected HDPS covariates and the identification of potentially influential covariates. Sensitivity analyses include varying the number of covariates selected and assessing the impact of covariates behaving empirically as instrumental variables. In our example, results were robust to both the number of covariates selected and the inclusion of potentially influential covariates. Furthermore, our HDPS models achieved good balance in key confounders.The data-adaptive approach of HDPS and the resulting benefits have led to its popularity as a method for confounder adjustment in pharmacoepidemiological studies. Reporting of HDPS analyses in practice may be improved by the considerations and tools proposed here to increase the transparency and reproducibility of study results.

Many real-word evidence (RWE) studies that utilize existing healthcare data to evaluate treatment effects incur substantial but avoidable bias from methodologically flawed study design; however, the extent of preventable methodological pitfalls in current RWE is unknown. To characterize the prevalence of avoidable methodological pitfalls with potential for bias in published claims-based studies of medication safety or effectiveness, we conducted an English-language search of PubMed for articles published from January 1, 2010 to May 20, 2019 and randomly selected 75 studies (10 case-control and 65 cohort studies) that evaluated safety or effectiveness of cardiovascular, diabetes, or osteoporosis medications using US health insurance claims. General and methodological study characteristics were extracted independently by two reviewers, and potential for bias was assessed across nine bias domains. Nearly all studies (95%) had at least one avoidable methodological issue known to incur bias, and 81% had potentially at least one of the four issues considered major due to their potential to undermine study validity: time-related bias (57%), potential for depletion of outcome-susceptible individuals (44%), inappropriate adjustment for postbaseline variables (41%), or potential for reverse causation (39%). The median number of major issues per study was 2 (interquartile range (IQR), 1-3) and was lower in cohort studies with a new-user, active-comparator design (median 1, IQR 0-1) than in cohort studies of prevalent users with a nonuser comparator (median 3, IQR 3-4). Recognizing and avoiding known methodological study design pitfalls could substantially improve the utility of RWE and confidence in its validity.

Actionable real-world evidence (RWE) requires accurate estimation of causal treatment effects. Calibration of RWE against randomized controlled trials (RCTs) is sometimes done to demonstrate that RWE can support the same causal conclusion as RCTs. Disagreements can occur when studies in each pair asked different questions in different populations or due to the presence of residual bias. Distinguishing among reasons for differences will impact the level of confidence in RWE. Several projects, such as the RCT DUPLICATE initiative funded by US FDA, NIH, and others, have launched in the last few years with the aim of assessing whether nonrandomized database studies can in some circumstances produce conclusions on the effectiveness of medications that are similar to those provided by RCTs.1,2 While comparison of randomized and nonrandomized findings is not new, heightened current interest is in part spurred by new initiatives at several regulatory agencies focused on assessing the role real-world evidence (RWE) can play in regulatory decision-making.3,4 FDA defines RWE as evidence on the benefits and risks of medications derived from routinely collected healthcare data, although other data sources such as patient registries may also be used. A key challenge for any project attempting to calibrate RWE findings against RCT findings is that differences between treatment effect estimates from the two study types can be driven by bias due to lack of randomization in RWE and/or by other differences in the design, such as inclusion/exclusion criteria, outcome measurement, or motivations for patients to adhere to study medications. Even if RWE studies are designed to match the corresponding RCT as closely as possible, emulation of all study components is typically impossible. Emulating target trials In clinical epidemiology, it has been recommended for several decades to contemplate how a randomized trial would be designed to answer a specific question before designing the nonrandomized counterpart to study the same question. Hernan and Robins call the former a hypothetical “target trial,” which would then be emulated by a nonrandomized study.5 This process has proven very useful in clarifying design choices for nonrandomized research on medications. It is also a highly flexible process, as specification of the target trial is often iterative. Realities of data collection, patient access, and other practical considerations impose constraints to the nonrandomized emulation; the hypothetical target trial can thus be adjusted so that the design that is feasible in a given healthcare database can accord with the design of the target trial. In calibrating RWE studies against existing RCTs, the process of emulating a target RCT in nonrandomized data is similar, except that in this case the target trial is already underway or completed. Thus, adjusting the design of the target RCT to improve the feasibility of the design in existing data is not possible, making exact emulation more difficult. Instead, investigators must adapt the design elements that they can from the trial, given the constraints of the database, and note the trial specifications that cannot be completely emulated. This process will highlight unavoidable emulation differences between a completed or in-progress RCT and the RWE replication. When assembling a series of RWE replications of RCTs, as in the RCT DUPLICATE initiative, there will be some trials where RWE specifications can closely emulate the RCT and others that cannot be emulated as closely. Examples of the latter are run-in periods during which non-adherent or drug intolerant patients are excluded before randomization, or, to homogenize patients, run-in periods that place all patients on a common medication before randomization. Therefore, enumerating – and quantifying to the extent possible – such emulation differences can provide insight into whether and to what extent differences in treatment effect estimates between RWE and corresponding RCTs are due to bias related to a lack of randomization versus other differences in design between the two study types. Specifically, attempting to correlate measures of emulation difference, such as those suggested in Table 1, with the magnitude of differences in treatment effect estimates may provide understanding of which emulation differences are most important in contributing to the “efficacy-effectiveness gap” between RWE and RCT findings, as long as bias due to confounding is not also correlated with the emulation difference of interest. Once better understood, such metrics could possibly be transformed into a simple three-point emulation scale. Table 1. Challenges in calibrating RWE against RCTs and measures of differences between the study types.

AIM: Recent years have witnessed a proliferation of observational research examining the association between glucose-lowering medications and cancer, paired with an increasing concern on the quality of the produced evidence. Our objective was to review published studies in this area to identify the most common methodological challenges and sources of bias. METHODS: We systematically searched PubMed to identify observational studies on glucose-lowering medications and cancer published between January 2000 and January 2016. We assessed the design and analytical methods used in each study, with a focus on their ability to achieve study validity, and further evaluated the prevalence of major methodological choices over time. RESULTS: Out of 155 studies evaluated, only 26% implemented a new user design, 41% used an active comparator, 33% implemented a lag or latency period, 51% adjusted for diabetes duration. Potential for immortal person-time bias was identified in 63% of the studies; 55% of the studies adjusted for variables measured during the follow-up without appropriate statistical methods. Aside from a decreasing trend in adjusting for variables measured during the follow-up, we observed no trends in methodological choices over time. CONCLUSIONS: The prevalence of well-known design and analysis flaws that may lead to biased results remains high among observational studies on glucose-lowering medications and cancer, limiting the conclusions that can be drawn from these studies. Avoiding known pitfalls could substantially improve the quality and validity of real-world evidence in this field.

PURPOSE Bootstrapping can account for uncertainty in propensity score (PS) estimation and matching processes in 1:1 PS-matched cohort studies. While theory suggests that the classical bootstrap can fail to produce proper coverage, practical impact of this theoretical limitation in settings typical to pharmacoepidemiology is not well studied. METHODS In a plasmode-based simulation study, we compared performance of the standard parametric approach, which ignores uncertainty in PS estimation and matching, with two bootstrapping methods. The first method only accounted for uncertainty introduced during the matching process (the observation resampling approach). The second method accounted for uncertainty introduced during both PS estimation and matching processes (the PS reestimation approach). Variance was estimated based on percentile and empirical standard errors, and treatment effect estimation was based on median and mean of the estimated treatment effects across 1000 bootstrap resamples. Two treatment prevalence scenarios (5% and 29%) across two treatment effect scenarios (hazard ratio of 1.0 and 2.0) were evaluated in 500 simulated cohorts of 10 000 patients each. RESULTS We observed that 95% confidence intervals from the bootstrapping approaches but not the standard approach, resulted in inaccurate coverage rates (98%-100% for the observation resampling approach, 99%-100% for the PS reestimation approach, and 95%-96% for standard approach). Treatment effect estimation based on bootstrapping approaches resulted in lower bias than the standard approach (less than 1.4% vs 4.1%) at 5% treatment prevalence; however, the performance was equivalent at 29% treatment prevalence. CONCLUSION Use of bootstrapping led to variance overestimation and inconsistent coverage, while coverage remained more consistent with parametric estimation.

The emergence and global spread of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has resulted in an urgent need for evidence on medical interventions and outcomes of the resulting disease, coronavirus disease 2019 (COVID-19). Although many randomized controlled trials (RCTs) evaluating treatments and vaccines for COVID-19 are already in progress, the number of clinical questions of interest greatly outpaces the available resources to conduct RCTs. Therefore, there is growing interest in whether nonrandomized real-world evidence (RWE) can be used to supplement RCT evidence and aid in clinical decision making, but concerns about nonrandomized RWE have been highlighted by a proliferation of RWE studies on medications and COVID-19 outcomes with widely varying conclusions. The objective of this paper is to review some clinical questions of interest, potential data types, challenges, and merits of RWE in COVID-19, resulting in recommendations for nonrandomized RWE designs and analyses based on established RWE principles.

Purpose To explore generalized boosted modeling (GBM) as a method for identifying subgroups with greater benefit or harm with dabigatran versus warfarin for treatment of atrial fibrillation. Methods We identified new initiators of warfarin or dabigatran with nonvalvular atrial fibrillation in 2 healthcare claims databases (2009-2013) and used GBM within 1 data source (development cohort) to explore subgroups where their effect on thromboembolism and major bleeding may differ. Identified subgroups were evaluated in the second data source (validation cohort) with stabilized-inverse-probability-of-treatment weights to adjust for confounding. Results Development and validation cohorts included 13 624 (28% dabigatran) and 62 596 (29% dabigatran) initiators, respectively. In development data, the strongest exposure interactions were prior thromboembolism and renal disease. In validation data, reduction in thromboembolism with dabigatran was greater for patients with versus without a history of thromboembolism by 2.8 (95% CI, -0.5 to 5.4) events per 100 patient-years. Major bleeding was reduced by 1.6/100 patient-years for dabigatran compared to warfarin initiators, without evidence of variation by renal disease. Conclusions We explored use of GBM to identify potential subgroups with different treatment effect. Dabigatran's superiority to warfarin at prevention of thromboembolism may be greater in secondary than primary prevention. In practice, secondary prevention patients are more often treated with warfarin.

AIM To evaluate the extent to which balance in unmeasured characteristics of patients with type 2 diabetes (T2DM) was achieved in claims data, by comparing against more detailed information from linked electronic health records (EHR) data. METHODS Within a large US commercial insurance database and using a cohort design, we identified patients with T2DM initiating linagliptin or a comparator agent within class (ie, another dipeptidyl peptidase-4 inhibitor) or outside class (ie, pioglitazone or a sulphonylurea) between May 2011 and December 2012. We focused on comparators used at a similar stage of diabetes to linagliptin. For each comparison, 1:1 propensity score (PS) matching was used to balance >100 baseline claims-based characteristics, including proxies of diabetes severity and duration. Additional clinical data from EHR were available for a subset of patients. We assessed representativeness of the claims-EHR-linked subset, evaluated the balance of claims- and EHR-based covariates before and after PS-matching via standardized differences (SDs), and quantified the potential bias associated with observed imbalances. RESULTS From a claims-based study population of 166 613 patients with T2DM, 7219 (4.3%) patients were linked to their EHR data. Claims-based characteristics in the EHR-linked and EHR-unlinked patients were similar (SD

Regulators consider randomized controlled trials (RCTs) as the gold standard for evaluating the safety and effectiveness of medications, but their costs, duration, and limited generalizability have caused some to look for alternatives. Real world evidence based on data collected outside of RCTs, such as registries and longitudinal healthcare databases, can sometimes substitute for RCTs, but concerns about validity have limited their impact. Greater reliance on such real world data (RWD) in regulatory decision making requires understanding why some studies fail while others succeed in producing results similar to RCTs. Key questions when considering whether RWD analyses can substitute for RCTs for regulatory decision making are WHEN one can study drug effects without randomization and HOW to implement a valid RWD analysis if one has decided to pursue that option. The WHEN is primarily driven by externalities not controlled by investigators, whereas the HOW is focused on avoiding known mistakes in RWD analyses.

In 2008, the US Food and Drug Administration (FDA) issued guidance on the need for cardiovascular outcome trials to assess the safety of new diabetes medications. Using two large commercial databases, we evaluated the effect of the FDA's cardiovascular safety guidance on drug development for type 2 diabetes as well as a comparison group of drugs intended to treat other alimentary and metabolic conditions. The FDA's guidance was associated with a 31% differential decrease in the rate of diabetes drugs entering phase II trials, but the remaining drugs were significantly more likely to target novel biological pathways (72% of drugs had novel mechanisms after the guidance vs. 49% before the guidance). No differential changes were observed for phase I and phase III trials. There was no measurable improvement during the study period in glycemic efficacy among investigational products entering phase III trials. This research highlights how regulatory actions can impact pharmaceutical innovation.

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.

Purpose Skepticism about the safety and effectiveness of certain generic drugs remains, particularly related to generic drugs that are approved by the Food and Drug Administration (FDA) using product-specific bioequivalence studies that differ from the standard testing pathway. The current study was designed to assess patient knowledge and perceptions of the generic drug approval process. Methods We conducted a survey of patients with 10 different chronic diseases. We recruited survey participants from the CVS Advisor Panel, a proprietary database of 124 621 CVS customers pre-consented to participate in online research activities. We created a survey to collect data on patients' perceptions of the FDA's generic drug approval process, as well as their experiences with generic drugs approved using modified bioequivalence approaches used to treat their chronic medical conditions. Results Our survey of 753 patients with chronic diseases (65% response rate) showed that most (74%) expressed little familiarity with FDA's approval process for generic drugs, but nearly all (89%) believed that FDA approval ensures the safety and effectiveness of generic drugs. About one-fifth of respondents reported hearing concerns about their generic drugs, most commonly from physicians (35-36%) and the Internet (32-38%), but there were no differences in patients' reports of concerns about generic versions of the six study drugs approved using product-specific pathways versus comparator drugs. Conclusions Patients have little knowledge about the generic drug approval system, but positive belief in the safety and effectiveness of generic drugs. Patients do not appear to have greater concern about generic drugs approved via product-specific pathways. Copyright © 2017 John Wiley & Sons, Ltd.

Objective While infection burden is high among patients with systemic lupus erythematosus (SLE), there is uncertainty about whether infection rates differ by immunosuppressive drug regimens. We undertook this study to compare infection rates among SLE patients newly initiating immunosuppressive therapy with mycophenolate mofetil (MMF), azathioprine (AZA), or cyclophosphamide (CYC). Methods Within the Medicaid Analytic eXtract database (2000–2010; 29 most populated US states), we identified adults with SLE starting MMF, AZA, or CYC treatment. We estimated propensity scores for receipt of MMF versus AZA and MMF versus CYC based on sociodemographic, comorbidity, and medication use information. After 1:1 propensity score matching, we estimated incidence rates of serious infections up to 6 and 12 months after drug initiation and used Cox regression to estimate hazard ratios (HRs) of first infection and death, with 95% confidence intervals (95% CIs). We performed primary intent-to-treat (ITT) and secondary as-treated analyses. Results We studied 1,350 propensity score–matched pairs of MMF and AZA initiators and 674 propensity score–matched pairs of MMF and CYC initiators. In 6-month ITT analyses, the incidence rate per 100 person-years for first serious hospitalized infection was 14.6 in MMF users and 15.2 in AZA users (HR of MMF versus AZA 0.99 [95% CI 0.74–1.32]). Comparing MMF to CYC, the incidence rate per 100 person-years for first serious infection was 24.1 in MMF users and 24.6 in CYC users (HR 0.95 [95% CI 0.69–1.32]). There were no differences in mortality in either comparison. As-treated analyses yielded similar results. Conclusion In a nationwide longitudinal study of Medicaid SLE patients at high risk of infection, rates of serious infection and mortality did not differ among new users of MMF, AZA, or CYC.

The U.S. Food and Drug Administration (FDA) issued several announcements related to potential risk of bisphosphonates including osteonecrosis of the jaw (2005), atrial fibrillation (2007), and atypical femur fracture (2010). We aimed to evaluate the impact of three FDA drug safety announcements on the use of bisphosphonates in patients with hip fracture using claims data from a U.S. commercial health plan (2004-2013). We calculated the proportion of patients in each quarter who received a bisphosphonate or other osteoporosis medication in the 6 months following hospitalization for hip fracture. Segmented logistic regression models examined the time trends. Among 22,598 patients with hip fracture, use of bisphosphonate decreased from 15% in 2004 to 3% in the last quarter of 2013. Prior to the 2007 announcement, there was a 4% increase in the odds of bisphosphonate use every quarter (OR 1.04; 95% CI, 1.02 to 1.07). After the 2007 announcement, there was a 4% decrease in the odds of bisphosphonate use (OR 0.96; 95% CI, 0.93 to 0.99) every quarter. The announcement in 2007 was associated with a significant decline in the rate of change of bisphosphonate uses over time (p

Purpose Record linkage can enhance data quality of observational database studies. Probabilistic linkage, a method that allows partial match of linkage variables, overcomes disagreements arising from errors and omissions in data entry but also results in false-positive links. The study aimed to assess the validity of probabilistic linkage in the absence of unique personal identifiers (UPI) and the methods of cutoff weight selection. Methods We linked an implantable cardioverter defibrillator placement registry to Medicare inpatient files of 1 year with anonymous nonunique variables and assessed the validity of three methods of cutoff selection against an internally derived gold standard with UPI. Results Of the 64 890 registry records with an expected linkage rate of 55–65%, 55% were linked at cutoffs associated with positive predictive value (PPV) of ≥90%. Histogram inspection suggested an approximate range of optimal cutoffs. The duplicate method made accurate estimates of cutoff and PPV if the method's assumption was met. With adjusted estimates of the sizes of true matches and searched files, the odds formula method made relatively accurate estimates of cutoff and PPV. Conclusions Probabilistic linkage without UPI generated valid linkages when an optimal cutoff was chosen. Cutoff selection remains challenging; however, histogram inspection, the duplicate method, and the odds formula method can be used in conjunction when a gold standard is not available. Copyright © 2015 John Wiley & Sons, Ltd.

Objective While serious infections are significant causes of morbidity and mortality in systemic lupus erythematosus (SLE), the epidemiology in a nationwide cohort of SLE and lupus nephritis (LN) patients has not been examined.

Purpose Trajectory models have been shown to (1) identify groups of patients with similar patterns of medication filling behavior and (2) summarize the trajectory, the average adherence in each group over time. However, the association between adherence trajectories and clinical outcomes remains unclear. This study investigated the association between 12-month statin trajectories and subsequent cardiovascular events. Methods We identified patients with insurance coverage from a large national insurer who initiated a statin during January 1, 2007 to December 31, 2010. We assessed medication adherence during the 360days following initiation and grouped patients based on the proportion of days covered (PDC) and trajectory models. We then measured cardiovascular events during the year after adherence assessment. Cox proportional hazards models were used to evaluate the association between adherence measures and cardiovascular outcomes; strength of association was quantified by the hazard ratio, the increase in model C-statistic, and the net reclassification index (NRI). Results Among 519842 statin initiators, 8777 (1.7%) had a cardiovascular event during follow-up. More consistent medication use was associated with a lower likelihood of clinical events, whether adherence was measured through trajectory groups or PDC. When evaluating the prediction of future cardiovascular events by including a measure of adherence in the model, the best model reclassification was observed when adherence was measured using three or four trajectory groups (NRI=0.189; 95% confidence interval: [0.171, 0.210]). Conclusions Statin adherence trajectory predicted future cardiovascular events better than measures categorizing PDC. Thus, adherence trajectories may be useful for targeting adherence interventions. Copyright © 2015 John Wiley & Sons, Ltd.

Objectives To determine whether adjustment for prognostic indices specifically developed for nursing home (NH) populations affect the magnitude of previously observed associations between mortality and conventional and atypical antipsychotics. Design Cohort study. Setting A merged data set of Medicaid, Medicare, Minimum Data Set (MDS), Online Survey Certification and Reporting system, and National Death Index for 2001 to 2005. Participants Dual-eligible individuals aged 65 and older who initiated antipsychotic treatment in a NH (N = 75,445). Measurements Three mortality risk scores (Mortality Risk Index Score, Revised MDS Mortality Risk Index, Advanced Dementia Prognostic Tool) were derived for each participant using baseline MDS data, and their performance was assessed using c-statistics and goodness-of-fit tests. The effect of adjusting for these indices in addition to propensity scores (PSs) on the association between antipsychotic medication and mortality was evaluated using Cox models with and without adjustment for risk scores. Results Each risk score showed moderate discrimination for 6-month mortality, with c-statistics ranging from 0.61 to 0.63. There was no evidence of lack of fit. Imbalances in risk scores between conventional and atypical antipsychotic users, suggesting potential confounding, were much lower within PS deciles than the imbalances in the full cohort. Accounting for each score in the Cox model did not change the relative risk estimates: 2.24 with PS-only adjustment versus 2.20, 2.20, and 2.22 after further adjustment for the three risk scores. Conclusion Although causality cannot be proven based on nonrandomized studies, this study adds to the body of evidence rejecting explanations other than causality for the greater mortality risk associated with conventional antipsychotics than with atypical antipsychotics.

Gruber and Tchetgen Tchetgen bring up many important limitations of p-value calibration as applied to the assessment of the safety and effectiveness of medications in secondary administrative data. In this discussion, I further examine the role of study design and selection of comparators in interpreting p-value calibration results, and we consider the advantages and disadvantages of automation in the epidemiology of medications, including p-value calibration. Copyright © 2016 John Wiley & Sons, Ltd.

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 Levy 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.

Objective Gout is a common form of inflammatory arthritis with an increasing prevalence in developed countries. It is well known that many patients with gout have significant comorbidities and high health care utilization. We aimed to describe the clinical characteristics and health care utilization patterns in patients with gout who were newly prescribed allopurinol, febuxostat, or colchicine. Methods We used US insurance claims data (2009–2011) to conduct a population-based cohort study. Results There were 25,051 allopurinol, 4,288 febuxostat, and 6,238 colchicine initiators. The mean age was 53 years and 83–87% were men. More than one-half of the patients had hypertension and hyperlipidemia, 20% had diabetes mellitus, and 10% had cardiovascular disease. The mean uric acid level was similar across the groups at baseline, ranging from 8.1–8.5 mg/dl. Compared with allopurinol or colchicine initiators, febuxostat initiators had more comorbidities and greater health care utilization, including outpatient, inpatient, or emergency room visits, both at baseline and during followup. Use of gout-related drugs such as opioids, steroids, and nonsteroidal antiinflammatory drugs was most common in febuxostat initiators and least common in colchicine initiators. The median daily doses at both the start and end of treatment were 300 mg for allopurinol, 40 mg for febuxostat, and 1.2 mg for colchicine. The doses of allopurinol and febuxostat were rarely increased during followup. Conclusion Patients who started allopurinol, febuxostat, or colchicine for gout generally had hyperuricemia and multiple comorbidities. Febuxostat initiators had more comorbidities and greater use of health care resources and gout-related drugs than the other groups. Overall, the doses of allopurinol or febuxostat remained unchanged over time.

Purpose Nursing home residents are of particular interest for comparative effectiveness research given their susceptibility to adverse treatment effects and systematic exclusion from trials. However, the risk of residual confounding because of unmeasured markers of declining health using conventional analytic methods is high. We evaluated the validity of instrumental variable (IV) methods based on nursing home prescribing preference to mitigate such confounding, using psychotropic medications to manage behavioral problems in dementia as a case study. Methods A cohort using linked data from Medicaid, Medicare, Minimum Data Set, and Online Survey, Certification and Reporting for 2001–2004 was established. Dual-eligible patients ≥65 years who initiated psychotropic medication use after admission were selected. Nursing home prescribing preference was characterized using mixed-effects logistic regression models. The plausibility of IV assumptions was explored, and the association between psychotropic medication class and 180-day mortality was estimated. Results High-prescribing and low-prescribing nursing homes differed by a factor of 2. Each preference-based IV measure described a substantial proportion of variation in psychotropic medication choice (β(IV → treatment): 0.22-0.36). Measured patient characteristics were well balanced across patient groups based on instrument status (52% average reduction in Mahalanobis distance). There was no evidence that instrument status was associated with markers of nursing home quality of care. Conclusion Findings indicate that IV analyses using nursing home prescribing preference may be a useful approach in comparative effectiveness studies, and should extend naturally to analyses including untreated comparison groups, which are of great scientific interest but subject to even stronger confounding. Copyright © 2014 John Wiley & Sons, Ltd.