The Experts below are selected from a list of 360 Experts worldwide ranked by ideXlab platform
Til Sturmer - One of the best experts on this subject based on the ideXlab platform.
-
the active comparator new user study design in Pharmacoepidemiology historical foundations and contemporary application
Current Epidemiology Reports, 2015Co-Authors: Jennifer L Lund, David B Richardson, Til SturmerAbstract:Better understanding of biases related to selective prescribing of, and adherence to, preventive treatments has led to improvements in the design and analysis of pharmacoepidemiologic studies. One influential development has been the “active comparator, new user” study design, which seeks to emulate the design of a head-to-head randomized controlled trial. In this review, we first discuss biases that may affect pharmacoepidemiologic studies and describe their direction and magnitude in a variety of settings. We then present the historical foundations of the active comparator, new user study design and explain how this design conceptually mitigates biases leading to a paradigm shift in Pharmacoepidemiology. We offer practical guidance on the implementation of the study design using administrative databases. Finally, we provide an empirical example in which the active comparator, new user study design addresses biases that have previously impeded pharmacoepidemiologic studies.
-
evidence of sample use among new users of statins implications for Pharmacoepidemiology
Medical Care, 2014Co-Authors: Til Sturmer, Alan M BrookhartAbstract:Background:Epidemiologic studies of prescription medications increasingly rely on large administrative health care databases. These data do not capture patients’ use of medication samples. This could potentially bias studies of short-term effects where date of initiation may be inaccurate.Objectives
-
the implications of propensity score variable selection strategies in Pharmacoepidemiology an empirical illustration
Pharmacoepidemiology and Drug Safety, 2011Co-Authors: Amanda R Patrick, Sebastian Schneeweiss, Kenneth J Rothman, Jerry Avorn, Robert J Glynn, Alan M Brookhart, Til SturmerAbstract:Purpose To examine the effect of variable selection strategies on the performance of propensity score (PS) methods in a study of statin initiation, mortality and hip fracture assuming a true mortality reduction of <15% and no effect on hip fracture.
-
indications for propensity scores and review of their use in Pharmacoepidemiology
Basic & Clinical Pharmacology & Toxicology, 2006Co-Authors: Robert J Glynn, Sebastian Schneeweiss, Til SturmerAbstract:Use of propensity scores to identify and control for confounding in observational studies that relate medications to outcomes has increased substantially in recent years. However, it remains unclear whether, and if so when, use of propensity scores provides estimates of drug effects that are less biased than those obtained from conventional multivariate models. In the great majority of published studies that have used both approaches, estimated effects from propensity score and regression methods have been similar. Simulation studies further suggest comparable performance of the two approaches in many settings. We discuss five reasons that favour use of propensity scores: the value of focus on indications for drug use; optimal matching strategies from alternative designs; improved control of confounding with scarce outcomes; ability to identify interactions between propensity of treatment and drug effects on outcomes; and correction for unobserved confounders via propensity score calibration. We describe alternative approaches to estimate and implement propensity scores and the limitations of the C-statistic for evaluation. Use of propensity scores will not correct biases from unmeasured confounders, but can aid in understanding determinants of drug use and lead to improved estimates of drug effects in some settings.
Samy Suissa - One of the best experts on this subject based on the ideXlab platform.
-
time related biases in Pharmacoepidemiology
Pharmacoepidemiology and Drug Safety, 2020Co-Authors: Samy Suissa, Sophie DellanielloAbstract:PURPOSE: Observational studies using computerized healthcare databases have become popular to investigate the potential effectiveness of old drugs for new indications. Many of these studies reporting remarkable effectiveness were shown to be affected by different time-related biases. We describe these biases and illustrate their effects using a cohort of patients treated for chronic obstructive pulmonary disease (COPD). METHODS: The Quebec healthcare databases were used to form a cohort of 124 030 patients with COPD, 50 years or older, treated between 2000 and 2015. Inhaled corticosteroids (ICS) and long-acting bronchodilators were used as exposures, with diverse outcomes, including lung cancer, acute myocardial infarction and death, to illustrate protopathic, latency time, immortal time, time-window, depletion of susceptibles, and immeasurable time biases. RESULTS: Protopathic bias affected bronchodilator-defined cohort entry with an incident rate of lung cancer of 23.9 per 1000 in the first year, compared with around 12.0 in the subsequent years. When latency and immortal times were misclassified, ICS were associated with decreased incidence of lung cancer (hazard ratio [HR] 0.32; 95% CI: 0.30-0.34), compared with 0.50 (95% CI: 0.48-0.53) after correcting for immortal time bias and 0.96 (95% CI: 0.91-1.02) after also correcting for latency time bias. Time-window, depletion of susceptibles and immeasurable time biases also affected the findings similarly. CONCLUSIONS: Many observational studies of new indications for older drugs reporting unrealistic effectiveness were affected by avoidable time-related biases. The apparent effectiveness often disappears with proper design and analysis. Future studies should consider these time-related issues to avoid severely biased results.
-
how Pharmacoepidemiology networks can manage distributed analyses to improve replicability and transparency and minimize bias
Pharmacoepidemiology and Drug Safety, 2020Co-Authors: Jeffrey S Brown, Olaf H. Klungel, Robert W Platt, Richard Platt, David Henry, Samy SuissaAbstract:Several Pharmacoepidemiology networks have been developed over the past decade that use a distributed approach, implementing the same analysis at multiple data sites, to preserve privacy and minimize data sharing. Distributed networks are efficient, by interrogating data on very large populations. The structure of these networks can also be leveraged to improve replicability, increase transparency, and reduce bias. We describe some features of distributed networks using, as examples, the Canadian Network for Observational Drug Effect Studies, the Sentinel System in the USA, and the European Research Network of Pharmacovigilance and Pharmacoepidemiology. Common protocols, analysis plans, and data models, with policies on amendments and protocol violations, are key features. These tools ensure that studies can be audited and repeated as necessary. Blinding and strict conflict of interest policies reduce the potential for bias in analyses and interpretation. These developments should improve the timeliness and accuracy of information used to support both clinical and regulatory decisions.
-
the quasi cohort approach in Pharmacoepidemiology upgrading the nested case control
Epidemiology, 2015Co-Authors: Samy SuissaAbstract:Observational studies of drug effects conducted using health care mega-databases often involve large cohorts with multiple time-varying exposures and covariates. These present formidable technical challenges in data analysis, necessitating sampling approaches such as nested case–control designs. The
-
the case crossover study design in Pharmacoepidemiology
Statistical Methods in Medical Research, 2009Co-Authors: Joseph A C Delaney, Samy SuissaAbstract:In the study of the association of transient drug exposures with acute outcomes, the case-crossover design is an efficient alternative to the case-control approach. This design based exclusively on the case series uses within-subject comparisons of drug exposures over time to estimate the rate ratio of the outcome associated with the drug under study. This design inherently removes the biasing effects of unmeasured, time-invariant confounding factors from the estimated rate ratio, but is sensitive to several assumptions. We illustrated the case-crossover design and explored its sensitivity using data from 4028 cases of gastrointestinal bleeding from the General Practice Research Database in assessing the effects of the drug warfarin. We compared the use of different time window lengths to assess exposure and considered the use of a case-time-control design to account for exposure time trends. The case-crossover approach found no excess risk of bleeding with warfarin exposure [rate ratio 0.98; 95% confidence interval (CI): 0.74-1.28] using a 1-month time window. When we restricted the analysis to subjects with truly transient drug exposure, defined by 1 to 3 prescriptions in the previous year, the rate ratio was 2.59 (95% CI: 1.42-4.74). To consider the longer 1-year exposure time window, the case-time-control approach was used and resulted in a rate ratio of 1.72 (95% CI: 1.08-2.43). In conclusion, the case-crossover design is potentially a powerful approach to assess the risk of drugs. This design is, however, highly sensitive to assumptions about intermittency of drug use and the length of the exposure time window, as demonstrated with the example of bleeding associated with warfarin use.
-
immortal time bias in Pharmacoepidemiology
American Journal of Epidemiology, 2008Co-Authors: Samy SuissaAbstract:Immortal time is a span of cohort follow-up during which, because of exposure definition, the outcome under study could not occur. Bias from immortal time was first identified in the 1970s in epidemiology in the context of cohort studies of the survival benefit of heart transplantation. It recently resurfaced in pharmaco-epidemiology, with several observational studies reporting that various medications can be extremely effective at reducing morbidity and mortality. These studies, while using different cohort designs, all involved some form of immortal time and the corresponding bias. In this paper, the author describes various cohort study designs leading to this bias, quantifies its magnitude under different survival distributions, and illustrates it by using data from a cohort of lung cancer patients. The author shows that for time-based, event-based, and exposure-based cohort definitions, the bias in the rate ratio resulting from misclassified or excluded immortal time increases proportionately to the duration of immortal time. The bias is more pronounced with a decreasing hazard function for the outcome event, as illustrated with the Weibull distribution compared with a constant hazard from the exponential distribution. In conclusion, observational studies of drug benefit in which computerized databases are used must be designed and analyzed properly to avoid immortal time bias.
Olaf H. Klungel - One of the best experts on this subject based on the ideXlab platform.
-
how Pharmacoepidemiology networks can manage distributed analyses to improve replicability and transparency and minimize bias
Pharmacoepidemiology and Drug Safety, 2020Co-Authors: Jeffrey S Brown, Olaf H. Klungel, Robert W Platt, Richard Platt, David Henry, Samy SuissaAbstract:Several Pharmacoepidemiology networks have been developed over the past decade that use a distributed approach, implementing the same analysis at multiple data sites, to preserve privacy and minimize data sharing. Distributed networks are efficient, by interrogating data on very large populations. The structure of these networks can also be leveraged to improve replicability, increase transparency, and reduce bias. We describe some features of distributed networks using, as examples, the Canadian Network for Observational Drug Effect Studies, the Sentinel System in the USA, and the European Research Network of Pharmacovigilance and Pharmacoepidemiology. Common protocols, analysis plans, and data models, with policies on amendments and protocol violations, are key features. These tools ensure that studies can be audited and repeated as necessary. Blinding and strict conflict of interest policies reduce the potential for bias in analyses and interpretation. These developments should improve the timeliness and accuracy of information used to support both clinical and regulatory decisions.
-
the reporting of studies conducted using observational routinely collected health data statement for Pharmacoepidemiology record pe
BMJ, 2018Co-Authors: Sinead Langan, Olaf H. Klungel, Vera Ehrenstein, Sigrun Aj Schmidt, Kevin Wing, Stuart G Nicholls, Kristian B Filion, Irene PetersenAbstract:In Pharmacoepidemiology, routinely collected data from electronic health records (including primary care databases, registries, and administrative healthcare claims) are a resource for research evaluating the real world effectiveness and safety of medicines. Currently available guidelines for the reporting of research using non-randomised, routinely collected data—specifically the REporting of studies Conducted using Observational Routinely collected health Data (RECORD) and the Strengthening the Reporting of OBservational studies in Epidemiology (STROBE) statements—do not capture the complexity of pharmacoepidemiological research. We have therefore extended the RECORD statement to include reporting guidelines specific to pharmacoepidemiological research (RECORD-PE). This article includes the RECORD-PE checklist (also available on www.record-statement.org) and explains each checklist item with examples of good reporting. We anticipate that increasing use of the RECORD-PE guidelines by researchers and endorsement and adherence by journal editors will improve the standards of reporting of pharmacoepidemiological research undertaken using routinely collected data. This improved transparency will benefit the research community, patient care, and ultimately improve public health.
-
methods to control for unmeasured confounding in Pharmacoepidemiology an overview
International Journal of Clinical Pharmacy, 2016Co-Authors: Jamal Uddin, Rolf H.h. Groenwold, Anthonius De Boer, Kit C.b. Roes, M S Ali, Muhammad Abdul Baker Chowdhury, Olaf H. KlungelAbstract:Background Unmeasured confounding is one of the principal problems in pharmacoepidemiologic studies. Several methods have been proposed to detect or control for unmeasured confounding either at the study design phase or the data analysis phase. Aim of the Review To provide an overview of commonly used methods to detect or control for unmeasured confounding and to provide recommendations for proper application in Pharmacoepidemiology. Methods/Results Methods to control for unmeasured confounding in the design phase of a study are case only designs (e.g., case-crossover, case-time control, self-controlled case series) and the prior event rate ratio adjustment method. Methods that can be applied in the data analysis phase include, negative control method, perturbation variable method, instrumental variable methods, sensitivity analysis, and ecological analysis. A separate group of methods are those in which additional information on confounders is collected from a substudy. The latter group includes external adjustment, propensity score calibration, two-stage sampling, and multiple imputation. Conclusion As the performance and application of the methods to handle unmeasured confounding may differ across studies and across databases, we stress the importance of using both statistical evidence and substantial clinical knowledge for interpretation of the study results.
-
Performance of prior event rate ratio adjustment method in Pharmacoepidemiology: a simulation study†
Pharmacoepidemiology and drug safety, 2014Co-Authors: Jamal Uddin, Tjeerd Van Staa, Rolf H.h. Groenwold, Anthonius De Boer, Svetlana V. Belitser, Arno W. Hoes, Kit C.b. Roes, Olaf H. KlungelAbstract:PURPOSE: Prior event rate ratio (PERR) adjustment method has been proposed to control for unmeasured confounding. We aimed to assess the performance of the PERR method in realistic pharmacoepidemiological settings. METHODS: Simulation studies were performed with varying effects of prior events on the probability of subsequent exposure and post-events, incidence rates, effects of confounders, and rate of mortality/dropout. Exposure effects were estimated using conventional rate ratio (RR) and PERR adjustment method (i.e. ratio of RR post-exposure initiation and RR prior to initiation of exposure). RESULTS: In the presence of unmeasured confounding, both conventional and the PERR method may yield biased estimates, but PERR estimates appear generally less biased estimates than the conventional method. However, when prior events strongly influence the probability of subsequent exposure, the exposure effect from the PERR method was more biased than the conventional method. For instance, when the effect of prior events on the exposure was RR = 1.60, the effect estimate from the PERR method was RR = 1.13 and from the conventional method was RR = 2.48 (true exposure effect, RR = 2). In all settings, the variation of the estimates was larger for the PERR method than for the conventional method. CONCLUSION: The PERR adjustment method can be applied to reduce bias as a result of unmeasured confounding. However, only in particular situations, it can completely remove the bias as a result of unmeasured confounding. When applying this method, theoretical justification using available clinical knowledge for assumptions of the PERR method should be provided. Copyright © 2014 John Wiley & Sons, Ltd.
-
propensity score balance measures in Pharmacoepidemiology a simulation study
Pharmacoepidemiology and Drug Safety, 2014Co-Authors: Sanni M Ali, Rolf H.h. Groenwold, Anthonius De Boer, Svetlana V. Belitser, Arno W. Hoes, Kit C.b. Roes, Wiebe R Pestman, Olaf H. KlungelAbstract:BACKGROUND: Conditional on the propensity score (PS), treated and untreated subjects have similar distribution of observed baseline characteristics when the PS model is appropriately specified. The performance of several PS balance measures in assessing the balance of covariates achieved by a specific PS model and selecting the optimal PS model was evaluated in simulation studies. However, these studies involved only normally distributed covariates. Comparisons in binary or mixed covariate distributions with rare outcomes, typical of pharmacoepidemiologic settings, are scarce. METHODS: Monte Carlo simulations were performed to examine the performance of different balance measures in terms of selecting an optimal PS model, thus reduction in bias. The balance of covariates between treatment groups was assessed using the absolute standardized difference, the Kolmogorov-Smirnov distance, the Levy distance, and the overlapping coefficient. Spearman's correlation coefficient (r) between each of these balance measures and bias were calculated. RESULTS: In large sample sizes (n ≥ 1000), all balance measures were similarly correlated with bias (r ranging between 0.50 and 0.68) irrespective of the treatment effect's strength and frequency of the outcome. In smaller sample sizes with mixed binary and continuous covariate distributions, these correlations were low for all balance measures (r ranging between 0.11 and 0.43), except for the absolute standardized difference (r = 0.51). CONCLUSIONS: The absolute standardized difference, which is an easy-to-calculate balance measure, displayed consistently better performance across different simulation scenarios. Therefore, it should be the balance measure of choice for measuring and reporting the amount of balance reached, and for selecting the final PS model. Copyright © 2014 John Wiley & Sons, Ltd.
Sebastian Schneeweiss - One of the best experts on this subject based on the ideXlab platform.
-
persistent user bias in case crossover studies in Pharmacoepidemiology
American Journal of Epidemiology, 2016Co-Authors: Jesper Hallas, Anton Pottegard, Sebastian Schneeweiss, Shirley V Wang, Joshua J GagneAbstract:Studying the effect of chronic medication exposure by means of a case-crossover design may result in an upward-biased odds ratio. In this study, our aim was to assess the occurrence of this bias and to evaluate whether it is remedied by including a control group (the case-time-control design). Using Danish data resources from 1995-2012, we conducted case-crossover and case-time-control analyses for 3 medications (statins, insulin, and thyroxine) in relation to 3 outcomes (retinal detachment, wrist fracture, and ischemic stroke), all with assumed null associations. Controls were matched on age, sex, and index date, and exposure over the preceding 12 months was ascertained. For retinal detachment, the case-crossover odds ratio was 1.60 (95% confidence interval (CI): 1.42, 1.80) for statins, 1.40 (95% CI: 1.02, 1.92) for thyroxine, and 1.53 (95% CI: 1.04, 2.24) for insulin. Estimates for the retinal detachment controls were similar, leading to near-null case-time-control estimates for all 3 medication classes. For wrist fracture and stroke, the odds ratios were higher for cases than for controls, and case-time-control odds ratios were consistently above unity, thus implying significant residual bias. In case-crossover studies of medications, contamination by persistent users confers a moderate bias upward, which is partly remedied by using a control group. The optimal strategy for dealing with this problem is currently unknown.
-
the implications of propensity score variable selection strategies in Pharmacoepidemiology an empirical illustration
Pharmacoepidemiology and Drug Safety, 2011Co-Authors: Amanda R Patrick, Sebastian Schneeweiss, Kenneth J Rothman, Jerry Avorn, Robert J Glynn, Alan M Brookhart, Til SturmerAbstract:Purpose To examine the effect of variable selection strategies on the performance of propensity score (PS) methods in a study of statin initiation, mortality and hip fracture assuming a true mortality reduction of <15% and no effect on hip fracture.
-
indications for propensity scores and review of their use in Pharmacoepidemiology
Basic & Clinical Pharmacology & Toxicology, 2006Co-Authors: Robert J Glynn, Sebastian Schneeweiss, Til SturmerAbstract:Use of propensity scores to identify and control for confounding in observational studies that relate medications to outcomes has increased substantially in recent years. However, it remains unclear whether, and if so when, use of propensity scores provides estimates of drug effects that are less biased than those obtained from conventional multivariate models. In the great majority of published studies that have used both approaches, estimated effects from propensity score and regression methods have been similar. Simulation studies further suggest comparable performance of the two approaches in many settings. We discuss five reasons that favour use of propensity scores: the value of focus on indications for drug use; optimal matching strategies from alternative designs; improved control of confounding with scarce outcomes; ability to identify interactions between propensity of treatment and drug effects on outcomes; and correction for unobserved confounders via propensity score calibration. We describe alternative approaches to estimate and implement propensity scores and the limitations of the C-statistic for evaluation. Use of propensity scores will not correct biases from unmeasured confounders, but can aid in understanding determinants of drug use and lead to improved estimates of drug effects in some settings.
-
exposure misclassification as a result of free sample drug utilization in automated claims databases and its effect on a Pharmacoepidemiology study of selective cox 2 inhibitors
Pharmacoepidemiology and Drug Safety, 2004Co-Authors: Susanna Jacobus, Sebastian Schneeweiss, Arnold K ChanAbstract:Purpose Free drug samples are widely used in clinical practice. We were concerned about free sample drug utilization as a source of misclassification in Pharmacoepidemiology research using claims data that may result in biased effect estimates. Methods We investigated the magnitude of potential bias with sensitivity analyses based on a published study that examined cardiovascular risk associated with selective cyclooxygenase 2 (COX-2) inhibitors. We derived an estimate of free sample drug utilization with market data for rofexcoxib and calculated sensitivity of the exposure ascertainment method using claims data. We corrected the incidence rate ratio assuming the observed unexposed incidence rate was actually a weighted average rate of the truly unexposed and free sample drug users. The impact of exposure misclassification measured as the percentage change from a corrected to the reported crude incidence rate ratio was examined under a range of free sample drug utilization proportions and unexposed cohort sizes. Results The proportion of free sample drug utilization of all rofecoxib use in our base case scenario was 15.48%, resulting in sensitivity of 84.52% for exposure ascertainment. The magnitude of bias was an underestimation of the unadjusted incidence rate ratio by 0.03%. With a free sample drug utilization proportion of 1.48% and the same unexposed cohort size of 237 975 person-years, the underestimation was 0.003%. If the unexposed cohort were 975 person-years given a 15.48% proportion, the underestimation was 8.82%. Conclusions In the Pharmacoepidemiology study, we examined that uses claims data to ascertain drug exposure, our results suggest that adjustment for free sample drug utilization is probably not warranted. Copyright © 2004 John Wiley & Sons, Ltd.
Laurel A Habel - One of the best experts on this subject based on the ideXlab platform.
-
blood pressure lowering medication initiation and fracture risk a swan Pharmacoepidemiology study
Archives of Osteoporosis, 2019Co-Authors: Daniel H Solomon, Kristine Ruppert, Rasa Kazlauskaite, Joel S Finkelstein, Laurel A HabelAbstract:We examined the fracture risk after initiation of blood pressure-lowering drugs compared with initiation of antidepressants. Multivariable regression models demonstrated an increased risk of fracture among women initiating a blood pressure-lowering medication (HR 1.73, 95% CI 1.02-2.95). This is likely related to an increased risk of falls. Purpose Initiation of blood pressure-lowering drugs has been associated with fractures in several studies, presumably due to an increase in the risk of falls. However, these studies used self-controlled designs without active comparators. We examined the risk of fractures after initiation of blood pressure lowering drugs compared with initiation of antidepressants. Methods Women participants in the Study of Women Across the Nation (SWAN) were potentially eligible if they initiated blood pressure-lowering or antidepressant drugs during follow-up. To reduce the risk of confounding, we estimated a propensity score that included potential confounders including age, menopausal status, osteoporosis, and osteoporosis medication use. The propensity score was used to match subjects in both groups and we then constructed multivariable logistic regression models comparing the risk of any fracture. Sensitivity analyses assessed a limited range of fractures less likely related to trauma. Results Among the 3302 potentially eligible women participating in the SWAN cohort, we were able to propensity-score match 289 women who initiated a blood pressure-lowering medication with 289 who initiated an antidepressant. Multivariable logistic regression models demonstrated an increased risk of fracture among women initiating a blood pressure lowering medication (OR 1.74, 95% CI 1.02-2.95). After excluding fractures of the digits and face, the results were similar (OR 1.57, 95% CI 0.88-2.81). Conclusions There was evidence of an increased risk in fractures among women initiating blood pressure-lowering medications compared to those initiating antidepressants. This is likely related to an increased risk of falling.
-
blood pressure lowering medication initiation and fracture risk a swan Pharmacoepidemiology study
Archives of Osteoporosis, 2019Co-Authors: Daniel H Solomon, Kristine Ruppert, Rasa Kazlauskaite, Joel S Finkelstein, Laurel A HabelAbstract:We examined the fracture risk after initiation of blood pressure–lowering drugs compared with initiation of antidepressants. Multivariable regression models demonstrated an increased risk of fracture among women initiating a blood pressure–lowering medication (HR 1.73, 95% CI 1.02–2.95). This is likely related to an increased risk of falls. Initiation of blood pressure–lowering drugs has been associated with fractures in several studies, presumably due to an increase in the risk of falls. However, these studies used self-controlled designs without active comparators. We examined the risk of fractures after initiation of blood pressure lowering drugs compared with initiation of antidepressants. Women participants in the Study of Women Across the Nation (SWAN) were potentially eligible if they initiated blood pressure–lowering or antidepressant drugs during follow-up. To reduce the risk of confounding, we estimated a propensity score that included potential confounders including age, menopausal status, osteoporosis, and osteoporosis medication use. The propensity score was used to match subjects in both groups and we then constructed multivariable logistic regression models comparing the risk of any fracture. Sensitivity analyses assessed a limited range of fractures less likely related to trauma. Among the 3302 potentially eligible women participating in the SWAN cohort, we were able to propensity-score match 289 women who initiated a blood pressure–lowering medication with 289 who initiated an antidepressant. Multivariable logistic regression models demonstrated an increased risk of fracture among women initiating a blood pressure lowering medication (OR 1.74, 95% CI 1.02–2.95). After excluding fractures of the digits and face, the results were similar (OR 1.57, 95% CI 0.88–2.81). There was evidence of an increased risk in fractures among women initiating blood pressure–lowering medications compared to those initiating antidepressants. This is likely related to an increased risk of falling.