To summarize experience with Johnson & Johnson's novel Active Safety Surveillance Using Real-world Evidence (ASSURE) program for producing efficient, transparent, applicable, and impactful RWE to support routine pharmacovigilance processes. ASSURE follows a stepwise framework that encompasses database diagnostics to assess fit-for-purpose real-world data (RWD) sources in the US, France, Germany, Australia, and Japan, phenotype development, query specification, and analytic implementation, objective study validity diagnostics, and standardized reporting. Licensed RWD sources are transformed into the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM), from which we generate evidence for safety signal evaluation questions using a suite of open-source analytic tools made publicly available through the OHDSI community. For each signal evaluation request involving a medication exposure and outcome(s), we conducted population characterization and population-level causal effects estimation. Target medication, comparator medication, and indication cohorts utilize predefined phenotypes, when available, for rapid analytics. We provide an evaluation of the initial implementation of this framework, including results from database diagnostics, phenotype development, and objective study validity diagnostics applied prior to unblinding evidence to prevent exposing biased estimates. We defined efficiency as the time from query to results and integration of results into safety decisions. The initial 19 months of the ASSURE program supported 110 safety signal evaluations across 24 unique products that arose from 47 requests. 74 evaluations (67%) passed study diagnostics and yielded effect estimation results with 62 (54%) providing propensity-score adjusted comparative cohort method results and 40 (36%) providing self-controlled case series results. Many analyses failed one or more diagnostics, preventing the generation and unblinding of potentially biased results: database (7 evaluations), phenotype specification (3 evaluations), or the requisite objective study validity diagnostics (26 evaluations). These 26 evaluations (24%) all provided population characterizations of exposure, indication, and outcome, including incidence rates. The ASSURE framework and implementation process, in collaboration with safety management teams, enables rapid response to medical product safety signals under evaluation and produces actionable RWE that supports pharmacovigilance decision making within regulatory timeframes. What ASSURE is ○ ASSURE (Active Safety Surveillance Using Real‐world Evidence) is a program developed to quickly and reliably check whether medicines are safe using real‐world health data—like insurance claims and electronic health records—from multiple countries (US, France, Germany, Australia, and Japan). How it works ○ We convert different health data sources into a single, common format so they can be analyzed the same way. ○ We use publicly available tools and clear clinical definitions for who is exposed to a medicine, who is not, and what counts as the health outcome of concern. ○ Before looking at the final results, we run a set of quality checks to make sure the data and methods answer the question correctly. ○ If the checks pass, we report whether the medicine appears to affect the risk of the outcome. If checks fail, we stop and share what we can (e.g., how often the medicine or outcome occurs) so decisions are not based on unreliable analyses. Key findings from the first 19 months. ○ ASSURE handled 110 safety analyses for 24 different products from 47 requests. ○ 67% (74 analyses) passed the quality checks and produced effect estimates. ○ 54% (62 analyses) produced adjusted comparative results (comparing users of the medicine to an appropriate comparison group). ○ 36% (40 analyses) produced self‐controlled results (which compare a person while taking or not taking medicine). ○ 24% (26 analyses) failed one or more important checks and therefore did not produce reliable comparison; for those, we still provided basic counts and rates. Why these matter ○ The approach makes safety reviews faster, more transparent, and more reliable by using the same standards and reusable tools across many patient populations in different data sources. ○ Running thorough checks before releasing results helps avoid mistaken or biased conclusions and supports better and timely decisions about medicine safety.
Rapid access to new drugs is of interest to patients and society. To analyze duration for clinical development, regulatory review, and reimbursement review of new drugs and assess whether these durations changed over the past decade in the US and European countries (Germany, France, and Switzerland). In this cross-sectional study, databases of the US Food and Drug Administration, the European Medicines Agency, and the Swiss Agency for Therapeutic Products were used to identify all new drugs approved from 2014 to 2024, investigational new drug application dates, submission dates, and approval pathways. Reimbursement dates were extracted from Medicare (US), Ministry of Health and Prevention (France), Federal Joint Committee (Germany), Federal Office of Public Health (Switzerland). Kaplan-Meier estimates were applied to calculate the median time for clinical development, regulatory review, and reimbursement review. This study included 519 drugs in the US, 412 in France, 412 in Germany, and 395 in Switzerland. Median total duration from clinical trial initiation until reimbursement was 9.3 (95% CI, 9.0-9.8) years in the US, 11.1 (95% CI, 10.2-12.0) years in France, 8.9 (95% CI, 8.3-9.5) years in Germany, and 9.8 (95% CI, 9.6-10.5) years in Switzerland. Median total duration from clinical trial initiation until reimbursement increased in the US from 9.0 (95% CI, 8.43-9.90) years in the first period (January 2014-August 2017) to 9.35 (95% CI, 8.75-11.56) years in the most recent period (May 2021-December 2024). In France, the median total duration increased from 9.82 (95% CI, 8.88-11.41) years to 12.18 (95% CI, 10.12-14.78) years. In Switzerland, median total duration increased from 9.42 (95% CI, 8.2-10.17) years to 10.05 (95% CI, 9.39-12.32) years. In Germany, the median total duration decreased from 9.27 (95% CI, 7.67-10.6) years to 8.7 (95% CI, 7.92-9.57) years. Cancer drugs showed a shorter median total duration compared with all drugs, especially in the US and Germany, and less pronounced in Switzerland and France. The findings of this study suggest that the overall median duration from clinical development until reimbursement of new drugs showed only minimal variation over time. Given recent efforts to accelerate new drug approvals, future research should evaluate potentially modifiable factors that increase clinical development durations. This could help to better target policies to accelerate access to new drugs, ideally adequately balancing safety with speed.
Nursing home (NH) residents are at high risk for major bleeding and are frequently exposed to complex medication regimens that may increase the likelihood of clinically important drug-drug interactions (DDIs). However, DDI evidence from community-dwelling populations may not generalize to NH residents, and real-world signals involving clopidogrel and oral anticoagulants remain incompletely characterized in this setting. We aimed to identify and interpret candidate DDI signals for major bleeding involving clopidogrel and oral anticoagulants among NH residents. Using linked Minimum Data Set assessments and Medicare fee-for-service claims (2013-2020), we conducted high-throughput, self-controlled case series (SCCS) screening analyses for four object drugs (clopidogrel, apixaban, rivaroxaban, and warfarin) paired with candidate precipitant medications. We estimated rate ratios (RRs) and 95% confidence intervals (CIs) comparing periods of concomitant use versus object drug use alone, applying semi-Bayes shrinkage to reduce false-positive signals. To aid interpretation, we incorporated two negative controls: a negative-control object drug (pravastatin) and a negative-control outcome (fall-related injury). A panel of experts evaluated signals for biological plausibility and potential bias. After semi-Bayes adjustment, 5/40 clopidogrel, 1/29 apixaban, 2/15 rivaroxaban, and 6/22 warfarin signals remained significant, and none did for pravastatin. Signals were also observed for fall-related injury. Experts identified only clopidogrel/sertraline and warfarin with cephalexin, sulfamethoxazole-trimethoprim, and furosemide as biologically plausible signals. SCCS screening for DDIs in NH residents identified bleeding signals for clopidogrel and oral anticoagulants, but few were mechanistically coherent and fall-related injury signals suggested residual bias. Nursing home residents often take many medications and are at high risk for serious bleeding, raising concerns about harmful drug–drug interactions. Because most evidence comes from people living in the community, it is unclear whether the same medication combinations pose similar risks in nursing homes. Using linked Medicare claims and nursing home clinical assessment data from 2013 to 2020, we studied residents hospitalized for major bleeding and used a within‐person approach to compare bleeding risk during times when a resident took clopidogrel or an oral anticoagulant (apixaban, rivaroxaban, or warfarin) together with another medication versus times when they took clopidogrel or an anticoagulant alone. We also used two checks to help interpret results: a comparison drug not expected to affect bleeding and an outcome not expected to be caused by these interactions. After statistical adjustment to reduce false signals, we found 14 medication pairs associated with higher bleeding risk, but an expert panel judged only four as biologically plausible: clopidogrel with sertraline, and warfarin with cephalexin, sulfamethoxazole‐trimethoprim, or furosemide. Several signals also appeared for the outcome used to check for bias, suggesting that some findings may reflect factors like illness‐related prescribing rather than true interactions.
Information on the safety profile of disease-modifying therapies (DMTs) in multiple sclerosis (MS) is often lacking in routine clinical practice. Real-world data sources, such as patient registries, support pharmacovigilance to characterise safety signals, as recently recognised by regulatory agencies. In Italy, the "Italian Multiple Sclerosis and Related Disorders Register" (RISM) was officially launched in 2015 to collect demographic and clinical data from patients with MS. To describe the main characteristics of RISM as a research platform supporting the safety evaluation of MS therapies. RISM enables the systematic characterisation of clinical events occurring in treated patients and, over the years, has implemented procedures to ensure high-quality data collection. A study cohort including subjects who initiated a DMT between 2016 and 2025 is analysed using descriptive statistics. Clinical events recorded during the study period are reported as absolute numbers and frequencies. Between 2016 and 2025, a total of 24 259 subjects received at least one DMT, and 4419 of them (18.2%) experienced one or multiple clinical events. For descriptive purposes, the occurrence of infections and malignancies, which accounted for 29.4% and 3.7% of all the clinical events collected in RISM respectively, are reported and categorised. This article outlines the strengths and limitations of RISM in conducting safety studies and highlights the methodological and legal challenges associated with such platforms. Routinely collected healthcare data within RISM provide a comprehensive, patient-centred perspective for the long-term evaluation of DMT safety and can address relevant questions on drug policies and public health. Multiple sclerosis (MS) is a chronic condition that often requires long‐term treatment with therapies that are specific for MS. However, information about the safety of MS therapies in everyday clinical practice is still limited. The Italian Multiple Sclerosis and Related Disorders Register (RISM) was created in 2015 to collect information from clinical practice about people with MS across Italy. By gathering data from specialised MS centres, RISM helps better understand how MS therapies work and what side effects may occur over time. Data quality controls have been implemented with the aim to collect complete and accurate information. Between 2016 and 2025, RISM recorded 24 259 people who received at least one MS treatment. About 18% of them reported at least one clinical event, most often infections (29.4% of all events) or, less frequently, cancers (3.7%). This work describes how RISM can be used to monitor the safety of available MS therapies by analysing data collected in routine clinical practice, as well as the challenges of managing such large data collections. Information from RISM can guide clinicians, researchers, and relevant stakeholders in improving treatment safety and supporting better healthcare decisions for people living with MS.
To determine the trajectories of benzodiazepine (BZD) and Z-drug use among older adults (n = 3590) followed in a network of geriatric outpatient clinics in Brazil. This longitudinal study evaluated BZD and Z-drug use over a 24-month follow-up period. The proportion of use at baseline and at the end of the study period was compared using the McNemar test. Two trajectories were considered: started using and stopped using. Comparisons between trajectories and independent variables were performed using Pearson's chi-square test. Variables with p < 0.20 were eligible for multivariate analysis. Multivariate models were also used to identify factors associated with BZD or Z-drug use at baseline and with trajectories of initiation and discontinuation. Odds ratios (OR) with 95% confidence intervals were estimated. A small but significant reduction in overall BZD or Z-drug use was observed (17.4% to 16.0%; p = 0.007), with a marked decrease in isolated Z-drug use (6.1% to 4.0%; p < 0.001). At baseline, use was associated with female sex (OR = 1.48; 95% CI 1.19-1.84), age 60-74 years (OR = 1.64; 95% CI 1.29-2.10), having ≥ 2 fall-related conditions (OR = 2.53; 95% CI 1.94-3.31), higher frailty scores, and cognitive dysfunction (OR = 1.32; 95% CI 1.00-1.74). Initiation was associated with pharmacist consultations during follow-up (OR = 5.01; 95% CI 2.88-8.71), depression (OR = 2.59; 95% CI 1.80-3.73), insomnia (OR = 1.97; 95% CI 1.19-3.25), and panic syndrome. Pharmacist consultation was also associated with discontinuation (OR = 0.34; 95% CI 0.15-0.76). Clinical vulnerability and psychiatric conditions were associated with BZD or Z-drug use and initiation, while pharmacist consultations were associated with both initiation and discontinuation trajectories. Benzodiazepines and Z‐drugs are commonly used to treat anxiety and sleep problems, but their use in older adults is associated with risks such as falls, cognitive impairment, and dependence. Understanding how the use of these medications changes over time can help to improve the care of older patients. This study followed 3590 older adults treated in a network of geriatric outpatient clinics in Brazil for 24 months. We analyzed two patterns of medication use: individuals who started using benzodiazepines or Z‐drugs and those who stopped using them during the follow‐up. We also examined clinical and care‐related factors associated with these changes. Overall, the use of these medications decreased slightly during the study period, with a more pronounced reduction in Z‐drug use. At the beginning of the study, the use was more common among women, aged 60–74 years, with two or more conditions related to falls, higher frailty scores, and cognitive dysfunction. Starting use during follow‐up was associated with depression, insomnia, panic syndrome, and pharmacist consultations. Pharmacist consultations were also associated with stopping these medications, suggesting that contact with healthcare professionals may influence medication use among older adults.
To develop and describe DrugUtilisation, an open-source R package that facilitates drug utilisation studies using data mapped to the OMOP Common Data Model (CDM). Core functionalities include creating drug user cohorts, identifying and summarising indications, describing the duration and dose of medication/s and assessing treatment adherence. The package works with packages developed within the DARWIN EU initiative to support study-specific workflows. We show the package's workflow by analysing the use of simvastatin in three European real-world databases. This paper outlines the DrugUtilisation package's functions and demonstrates their application with a clinical example of simvastatin use in databases from the United Kingdom, Estonia and the Netherlands. We generated results including cohort counts, indication summaries, measures of dose and duration, as well as publication-ready tables and figures. We implemented comprehensive unit tests and standardised output format, which ensured consistency across databases and minimised coding errors. The development of this software allows for researchers to quickly perform common drug utilisation analyses, while also providing the foundation for additional, bespoke study-specific analyses. Drugs are prescribed in everyday clinical practice and drug utilisation studies help us understand their use. Healthcare databases record prescriptions, refills and related patient information for millions of people, but these data often come in different formats. The OMOP Common Data Model is a shared format that puts different healthcare databases into the same structure so they can be analysed in the same way. We built the DrugUtilisation R package, an open‐source R package to conduct drug utilisation analysis of OMOP CDM data. The package has already been used in 62 studies. With only a few lines of code, researchers can define cohorts of drug users, characterise treatment initiation, quantify dose and duration and track changes in use over time. The package also produces publication‐ready tables and figures. As an illustration, we analysed the use of simvastatin (a cholesterol lowering drug) in three national databases from the United Kingdom, Estonia and the Netherlands. DrugUtilisation makes the monitoring of medication use faster and easier for researchers, clinicians and regulators.
Few studies have investigated the interaction between drug-environment despite the plausibility and concern. We sought to explore the synergistic effects of corticosteroids and PM2.5 on cardiovascular and thromboembolic events in high-risk older adults in Taiwan. To assess the synergistic effect of corticosteroid use and PM2.5 exposure on cardiovascular and thromboembolic events and all cause mortality among elderly individuals in Taiwan. We conducted a retrospective cohort study using the Taiwan National Health Insurance database from 2009 to 2019. We included patients aged 65 years or older and at high risk for cardiovascular and thromboembolic events. Exposures included corticosteroid therapy and seasonal mean PM2.5. Primary outcomes included myocardial infarction or acute coronary syndrome, ischemic stroke or transient ischemic attack, heart failure, venous thromboembolism, atrial fibrillation, and all-cause mortality. We then fitted history-adjusted marginal structural Cox proportional hazard models to investigate both the independent and synergistic effects of PM2.5 and corticosteroid use on the CTE outcomes. We measured whether there was a sufficient cause interaction between PM2.5 and corticosteroids using estimates of relative excess risk due to interaction (RERI). The cohort included 373 402 participants, with average age at index date 73.6 ± 7.2 years; 57.7% of cohort members were female. During the study period, 1 614 331 (43.2%) participants had at least one corticosteroid prescription. The average PM2.5 concentration was 28.3 ± 17.4 μg/m3. Increasing PM2.5 from 15 to 20 μg/m3 resulted in significantly increased excess risk due to interaction (RERI [95% CI]) for all-cause mortality (3.1% [0.27%, 5.9%]). A synergistic effect between seasonal PM2.5 exposure and corticosteroid use was found on all-cause mortality in elderly patients at high risk for cardiovascular and thromboembolic events. Limited research has been conducted on the interaction between drugs and environmental factors. Our study aimed to investigate the combined effects of corticosteroids and PM2.5 on cardiovascular and thromboembolic incidents among high‐risk older adults in Taiwan. We performed a study utilizing the National database from 2009 to 2019, including patients aged 65 and older. The exposures of interest were corticosteroid treatment and seasonal average PM2.5 levels. The primary outcomes measured were cardiovascular and thromboembolic events and overall mortality. The cohort comprised 373 402 participants, with an average age of 73.6 ± 7.2 years, and 57.7% were female. Throughout the study, 161 433 (43.2%) participants received at least one corticosteroid prescription. The average PM2.5 concentration was 28.3 ± 17.4 μg/m3. An increase in PM2.5 from 15 to 20 μg/m3 was associated with a significant rise in risk for overall mortality. Our findings indicate a synergistic effect between seasonal PM2.5 exposure and corticosteroid use on overall mortality in elderly patients at high risk for cardiovascular and thromboembolic events.
Within one week of the Food and Drug Administration granting the COVID-19 vaccine Emergency Use Authorization, the HERO-Together study launched. We describe how the study utilized novel methods for streamlined enrollment and longitudinal data collection to efficiently monitor early and long-term safety of the COVID-19 vaccine. HERO-Together is a fully remote, prospective observational cohort study, conducted through an online portal. Participants were recruited from an existing online community of healthcare workers that expanded to family and community members as vaccine eligibility expanded, as well as through marketing and point-of-vaccination efforts. Individuals received notifications to submit data at scheduled intervals for two years, and automated reminders prompted participants with missed check-ins, with centralized call center outreach for non-responders. Multiple recruitment approaches were adopted, including direct outreach, national and local media campaigns, downloadable enrollment toolkits (e.g., social media), social media influencer partnerships, and pharmacy referrals. A study population of n = 19, 858 was enrolled within nine months of emergency use authorization, most self-reporting as white, female, working in healthcare, and all for whom variable survey completion was observed. Collectively, diverse recruitment strategies, data collection through the online portal, and electronic consent enabled rapid, decentralized enrollment of study participants in order to remotely monitor COVID-19 vaccine safety during the early days of the pandemic. EUPAS38671. The HERO‐Together study started within a week of the COVID‐19 vaccine receiving Emergency Use Authorization in the United States, aiming to quickly and efficiently monitor both short‐ and long‐term safety. The study was fully online, which made it easy to join and participate from home. Participants first included healthcare workers and later expanded to their families and community members as more people became eligible for the COVID‐19 vaccine. Participants were recruited through an existing online community, through marketing efforts, and at vaccination sites. Nearly 20 000 participants joined the study within 9 months of vaccine launch; most were white, female, and worked in healthcare. Participants were asked to complete online surveys regularly for 2 years. Automated reminders were sent to encourage survey completion, and call center staff followed up with those who did not respond.
In 2019, the Innovative Medicines Initiative funded the ConcePTION project to enhance monitoring of medication safety in pregnancy and breastfeeding. This paper describes how the ConcePTION Pregnancy Algorithm (PA) identified pregnancies in 10 diverse European electronic healthcare data sources and estimated their duration. Data sources from six European countries were mapped to the ConcePTION Common Data Model. Any pregnancy-related record was retrieved from various available data banks, including birth register, primary care records, and hospital records, and reconciled into episodes of pregnancy (starting between 01/2015 and 12/2019), each with start date, end date, and type of end. A random forest model was used to estimate missing gestational ages for incomplete records. Parameters were tailored to data sources to address local variations in data availability, collection, and governance. Model performance was evaluated using cross-validated Root Mean Squared Error (RMSE). The PA identified ~2.7 million pregnancies, in over 2.2 million individuals. Most ended in live births (50%-83%), 1%-15% in elective terminations, and 4%-10% in spontaneous abortions, depending on data sources. Pregnancies with unknown type of end were also retrieved (2%-34%). Gestational age was predicted for 6%-89% of records (RMSE: 17-50 days). The median gestational age at first identified pregnancy record ranged from 47 to 280 days. We developed an open-source algorithm to identify and date pregnancies, including early-stage pregnancies with unknown end and/or ongoing at the time of data extraction. This algorithm may facilitate multinational studies, improving generation of timely real-world evidence about use and safety of medicinal products in pregnancy. Pregnant individuals are often excluded from clinical trials, which limits the availability of evidence on the safety of medicines used during pregnancy. To help address this gap, the Innovative Medicines Initiative funded the ConcePTION project, aimed at improving the monitoring and communication of medication safety during pregnancy and breastfeeding. As part of this initiative, we developed the ConcePTION Pregnancy Algorithm (PA), a tool designed to identify and estimate the duration of pregnancies using routinely collected healthcare data. The algorithm combines information from various sources, including hospital records, primary care data, and birth registers. When applied to 10 healthcare databases across six European countries, the PA identified approximately 2.7 million pregnancies between 2015 and 2019. It successfully captured early‐stage pregnancies, as well as those with unknown end or still ongoing at the time of data collection. Those cases are often overlooked by previously published algorithms for the identification of pregnancies. The PA employs machine learning techniques to estimate missing pregnancy start dates. As an open‐source tool, it can be tailored to accommodate diverse healthcare systems and governance restrictions. By enabling more complete and reliable identification of pregnancy episodes, the PA supports high‐quality research and the timely generation of evidence on medicine use and safety during pregnancy. The end of a pregnancy in the early stages, such as spontaneous abortions, is more likely to go unrecorded in electronic healthcare databases compared to an end of pregnancy with longer course. Existing pregnancy‐finding algorithms generally rely on data recorded at the end of pregnancy, potentially introducing selection bias. We developed an algorithm to identify early‐stage, ongoing, and pregnancies with unknown ends, leveraging a Random Forest model to estimate pregnancy start dates when they are not directly recorded. The algorithm operates in a standardized yet adaptable manner across diverse data sources, while maintaining methodological transparency. We described and interpreted the algorithm's results across 10 diverse European data sources, varying in data availability.
Surveys have indicated that patients and clinicians can overestimate the efficacy and safety of drugs approved by the US Food and Drug Administration (FDA). In recent years, only approximately half of new drug approvals have been based on 2 or more adequate and well-controlled trials; furthermore, regulations do not limit how many trials can be conducted or provide clear guidance on how the FDA should consider a drug with conflicting evidence of benefit from multiple trials. To understand how the FDA evaluated a single investigational drug with positive and negative preapproval trials. The case of gepirone extended release (ER), approved for major depressive disorder, was reviewed. The FDA based its efficacy evaluation of gepirone ER on 13 trials: 12 acute treatment trials and 1 maintenance or relapse prevention trial. The FDA judged 2 acute treatment trials as positive. In the others, gepirone ER did not demonstrate superiority to placebo; and for 3, the FDA found evidence of statistical inferiority to an active comparator. Concerned that the positive trials might have occurred by chance and the amount of countervailing evidence, the FDA rejected the New Drug Application from the sponsor 4 times (Organon in 1999, 2002, and 2004, and Fabre-Kramer Pharmaceuticals, Inc in 2007). In 2014, the sponsor filed a dispute resolution request, leading to intervention by senior FDA leaders. In 2015, an FDA Advisory Committee voted that drug efficacy had not been demonstrated. Nevertheless, FDA leaders came to agree with the sponsor's arguments that the 2 positive trials had not occurred by chance and that the drug satisfied the FDA statutory standard for efficacy, and the drug was approved. The case of gepirone shows how the FDA evaluated an investigational drug with conflicting evidence. The FDA sometimes exercises "regulatory flexibility" and focuses on statistical (as opposed to clinical) significance in a few trials, allowing the approval of drugs with scant benefits. Product labeling should transparently report on all adequate and well-controlled trials relating to an FDA-approved indication, not just those with positive outcomes, so that clinicians can make better-informed prescribing decisions.
The global opioid crisis has highlighted substantial differences in prescribing, monitoring, and regulatory practices across health systems. In Costa Rica, comprehensive national data on opioid use have been limited. This study aimed to describe temporal trends in legally prescribed opioid consumption between 2017 and 2024 using nationwide administrative data. A nationwide, retrospective, population-based analysis was conducted using data from the Costa Rican Ministry of Health's narcotics registry. Dispensation records for morphine, methadone, fentanyl, oxycodone, and tapentadol were included. Opioid consumption was standardized to Morphine Milligram Equivalents (MME) using NIH HEAL conversion factors. Temporal trends were assessed using descriptive statistics and simple linear regression. A total of 31.9 million opioid dispensations were recorded during the study period. National opioid consumption peaked in 2018 (395.1 million MME), followed by an approximate 70% decline through 2023 and a modest increase in 2024. Consumption of morphine and methadone declined substantially (-72.2% and -82.4%, respectively), while oxycodone and tapentadol showed moderate increases over time. These findings indicate a marked reduction in overall opioid consumption alongside changes in the distribution of use across opioid agents. Between 2017 and 2024, Costa Rica experienced a sustained decline in total opioid dispensing and a redistribution of opioid use across substances. These trends occurred in the context of strengthened regulatory oversight and expanded digital monitoring of controlled substances. While causal relationships cannot be established, the findings provide relevant population-level evidence to support opioid stewardship and inform regulatory and public health strategies in middle-income health systems. This study examined how patterns of legally prescribed opioid use in Costa Rica changed between 2017 and 2024 using official national data from the Ministry of Health. Opioids are medications used to treat moderate to severe pain, but they require careful monitoring because they are associated with risks such as dependence and overdose. To describe national trends, we analyzed pharmacy dispensation records for several commonly used opioids and converted all doses into morphine milligram equivalents (MME), a standard measure that allows comparisons across different drugs. We found that opioid consumption in Costa Rica reached its highest level in 2018 and then declined by more than 70% through 2023, with a modest increase in 2024. The largest reductions were observed for morphine, methadone, and fentanyl, while oxycodone and tapentadol showed moderate increases over time. Together, these findings describe a redistribution of opioid use across different agents alongside an overall decline in total consumption. These trends occurred during a period of strengthened regulatory oversight, including the implementation of a national Digital Prescription System for controlled substances. Although this study was not designed to determine causal effects, the results provide important population‐level evidence on how opioid dispensing patterns have evolved in Costa Rica. Understanding these changes can help inform future strategies to balance access to effective pain management with patient safety and public health protection.
To extend the Chronic Hypertension and Pregnancy (CHAP) trial findings, this study compared labetalol, nifedipine, and methyldopa initiation at or before 23 weeks' gestation in pregnant patients with chronic hypertension, examining effects on a composite effectiveness outcome and on small for gestational age birth (SGA). We used California Medicaid claims from deliveries between 2016 and 2020, linked to the California Study of Outcomes in Mothers and Infants (SOMI). The exposure was defined as ≥ 1 paid Medicaid fill for labetalol, extended-release nifedipine, or methyldopa in the first 23 weeks of pregnancy among women who did not use oral antihypertensive therapy in the period starting 90 days prior to last menstrual period (LMP) through LMP. Like CHAP, the composite outcome was preeclampsia with severe features, preterm birth < 35 weeks (PTB), placental abruption, and/or fetal or neonatal death. SGA was modeled as a safety outcome. The analysis included 2281 singleton births: 1496 labetalol initiators, 653 methyldopa initiators, and 132 nifedipine initiators. Composite outcome incidence varied by medication (labetalol: 23%; methyldopa: 19%; nifedipine: 24%). The adjusted risk ratio (aRR) comparing methyldopa versus labetalol was 0.82 (95% CI: 0.68, 1.00); nifedipine versus labetalol was 1.01 (95% CI: 0.72, 1.41); and nifedipine versus methyldopa was 1.22 (95% CI: 0.85, 1.76). For SGA, results were similar (aRR (95% CI): methyldopa vs. labetalol: 0.81 (0.62, 1.05); nifedipine vs. labetalol: 1.07 (0.65, 1.75); nifedipine vs. methyldopa: 1.32 (0.78, 2.24)). Nifedipine and labetalol were equivalent in effectiveness and safety. Methyldopa was associated with lower risks of adverse outcomes. Potential residual confounding and limited overlap between treatment groups warrant caution in interpretation of our findings. Future research is needed to clarify whether the apparent advantages of methyldopa reflect true therapeutic benefit or underlying biases. There is currently no international consensus on which drug is preferable in treating chronic hypertension in pregnancy. Thus, we aimed to compare labetalol, nifedipine, and methyldopa initiation prior to 23 weeks of pregnancy to treat chronic hypertension in pregnancy. Using California Medicaid data linked with birth and hospital records from women who delivered between 2016 and 2020, we examined risks of preterm birth before 35 weeks, preeclampsia with severe features, placental abruption, fetal or newborn death, and small‐for‐gestational‐age (SGA) birth. We found that all outcomes were similar between nifedipine and labetalol users. Contrastingly, we found that people who started methyldopa had lower risks of these complications than those who started nifedipine or labetalol, largely because of lower risk of preterm birth occurring before 35 weeks. Methyldopa users also had a lower risk of SGA infants. Although these results show methyldopa may offer advantages in preventing these outcomes, differences in patient characteristics, lack of baseline blood pressure data, and potential increased side effects mean the findings should be interpreted cautiously. Further studies with detailed clinical data, such as blood pressure readings, are needed to confirm whether methyldopa truly provides better outcomes during pregnancy.
Unrecognized contraindications pose risks for adverse drug reactions, hospitalizations or death. Clinical decision support systems (CDSS) aim to mitigate medication-related harm, particularly originating from contraindications. However, many CDSS provide limited benefit, as they focus largely on singular risk situations such as drug-drug interactions and often generate alerts of limited clinical relevance. Comprehensive integration of contraindications into CDSS may support more clinically meaningful alerts. The aim of this work was the development and analysis of machine-readable contraindication lists, including drug-clinical condition, drug-kidney function and drug-drug (group) contraindications, for integration into CDSS and real-world data analysis. We extracted and operationalized contraindications, based on Summaries of Product Characteristics (SmPCs), of the 688 most prescribed drugs in Germany, leveraging common medical coding systems. Moreover, we analyzed extracted contraindications based on operationalizability, overall frequency and frequencies within different contraindication categories. In total, we extracted 4676 contraindications, of which 2129 (45.5%) were deemed operationalizable. Of these 2129 contraindications, 1652 (77.6%) were attributed to drug-clinical condition, 83 (3.9%) to drug-kidney, 140 (6.6%) to drug-drug group and 254 (11.9%) to drug-drug. The most frequently mentioned contraindicated risk situations were 'severe liver insufficiency' (n = 74, 3.5%), 'pregnancy' (n = 66, 3.1%), 'MAO-inhibitors' (n = 44, 2.1%), and 'shock' (n = 44, 2.1%). Our results show, that drug-clinical condition contraindications are listed far more frequently in SmPCs than other contraindication categories. Focusing on clinical condition-related contraindications within CDSS could improve the detection of clinically relevant contraindications in routine data and enhance medication safety. The clinical applicability is currently being evaluated in the INTERPOLAR study.
Digital databases such as pharmacovigilance (PV) databases could provide unique opportunities to monitor trends in suspected antibiotic resistance, ineffectiveness, and misuse, extending beyond their traditional role of tracking adverse drug reactions (ADRs). This approach is potentially valuable globally but particularly advantageous in lower-middle-income countries (LMICs) where formal resistance surveillance systems are often insufficiently developed. Leveraging PV data could help generate early signals of resistance and inappropriate antibiotic use and support antimicrobial stewardship in resource-constrained settings. To explore the potential use of PV databases in monitoring suspected antibiotic resistance trends and inappropriate use in LMICs. A retrospective cross-sectional study was conducted using VigiBase. Data were extracted in October 2024 from inception to January 1, 2024. Reports involving antibacterials for systemic use, Anatomical Therapeutic Chemical (ATC) codes J01 and J04 from LMICs were included. Selected Medical Dictionary for Regulatory Activities (MedDRA) preferred terms were mapped according to RIOLE classification to the "resistance," "ineffectiveness," "off-label use," and "error" categories to identify reporting patterns. Descriptive statistics were used to summarize reports' characteristics, and associations between categorical variables were examined using chi-squared tests. A total of 1570 ICSRs from 37 LMICs were identified, yielding 2958 drug-adverse event pairs, with reporting increasing markedly after 2016. The "off-label use" (38.6%) and "ineffectiveness" (37.0%) were the dominant RIOLE categories, driven mainly by the preferred terms (PTs) of Off-label use (795; 26.4%) and Drug ineffective (751; 25.4%). Resistance-related PTs accounted for 12.7% of pairs, most frequently Drug resistance (210; 7.0%) and Pathogen resistance (132; 4.5%), while "error" category (11.7%) was led by Product use issue (60; 2.0%) and Medication error (44; 1.5%). Watch antibiotics predominated, especially azithromycin, ceftriaxone, and meropenem, with significant associations observed between RIOLE categories and age, reporter type, ATC class, reaction outcome, AWaRe category, and WHO region. These findings demonstrate that PV databases can provide valuable insights into suspected antibiotic resistance and inappropriate use patterns in LMICs, supporting their potential role as additional data sources in antimicrobial stewardship. Antibiotics sometimes fail to work as expected, either because the bacteria are resistant, the wrong antibiotic was chosen, or the medicine was used incorrectly. In this study, we analyzed reports from lower‐ middle‐income countries (LMICs) submitted to the World Health Organization's global safety database (VigiBase) to understand why antibiotic treatments go wrong. We found that most reports were related to antibiotics being used for the wrong indication or not working as intended, while only a smaller number described confirmed resistance. These patterns likely reflect challenges in LMICs' health systems, such as limited access to diagnostic tests and the need to make treatment decisions based on symptoms alone. The study also shows that safety reporting systems, originally designed to detect side effects, can provide important early warnings about possible antimicrobial resistance, especially in places where laboratory testing is limited. Strengthening antibiotic prescribing practices, improving how health workers report treatment failures, and integrating pharmacovigilance data into national antimicrobial resistance programs could help countries identify problems sooner and support safer, more effective use of antibiotics.
Oropharyngeal adverse events (O-AEs) represent a potential safety concern associated with several drugs and/or vaccines. Although often underestimated, these events may provide valuable insights into a patient's overall clinical condition, appearing initially mild but later worsening. Therefore, this study aimed to analyze O-AEs related to drugs and/or vaccines using structured safety data. Safety reports from three Italian regions were retrieved from the national pharmacovigilance database (rete nazionale di farmacovigilanza, RNF) and analyzed for the period 2022-2024. All reports were structured and analyzed according to the International Council of Harmonisation (ICH) E2B (R3) format. Additionally, reporting odds ratios (RORs) were calculated to compare the likelihood of O-AEs being reported by different categories of reporters (e.g., physicians, other healthcare professionals, or patients). Over three years, 47,664 reports were collected, of which 1,740 (3.6%) contained at least one suspected O-AE. Most patients were female (65.7%) with a median age of 55 years (IQR 38-66). The majority of reports described non-serious events (66.7%), and outcomes were favorable in most cases (75.9%). The safety reports related to drugs (90.5%) were largely more than those related to vaccines (9.5%). A total of 129 cases of medication-related osteonecrosis of the jaw (MRONJ) were identified, most of which were serious and had unfavorable outcomes. Disproportionality analysis revealed that physicians were less likely to report O-AEs than patients (ROR=0.72; 0.62-0.84; P<<0.05) and more likely than nurses (ROR=1.34; 0.98-1.87; P<0.05). Compared with 2019-2021, the main difference was the lower proportion of vaccine-related reports, which declined from 47.8% to 9.5% in 2022-2024. Only a small proportion of safety reports involved O-AEs. These findings highlight the importance of enhancing awareness among physicians (particularly dentists) regarding O-AEs, and of fostering a collaborative pharmacovigilance culture across healthcare providers, thereby improving patient safety through more timely and reliable reporting.
Men with castration-resistant prostate cancer (CRPC) who have a pre-existing history of cardiovascular disease (CVD) or other comorbidities are often excluded from clinical trials involving oral androgen receptor pathway inhibitors (ARPi). In this study, we compared all-cause and prostate cancer-specific mortality among CRPC patients, with and without a pre-existing history of CVD, receiving ARPi compared to chemotherapy. Men with CRPC were identified using the Surveillance, Epidemiology, and End Results-Medicare Linked Database from 2004 to 2015. Patients were grouped into two analytical cohorts by drug use. Inverse probability treatment weights (IPTW)-adjusted Cox models and Fine-Gray subdistribution hazards models were used to evaluate associations between ARPi and chemotherapy, and between ENZ and AA for all-cause mortality and cancer-specific mortality separately. The study cohort included 1438 men with CRPC. Nearly 54.4% of patients had a pre-existing history of CVD. Patients with a pre-existing history of CVD had a higher incidence of all-cause and prostate cancer-specific mortality compared to patients without a history of CVD (all-cause mortality: 69.7% vs. 59.3%, prostate cancer-specific mortality: 56.0% vs. 49.5%, respectively). In the pre-existing history of CVD cohort, the IPTW-adjusted Cox model showed a significantly lower all-cause mortality in patients who received APRi, enzalutamide, and abiraterone, versus chemotherapy (IPTW-adjusted hazard ratio [AHR], 0.56; 95% Confidence Interval [CI], 0.48-0.64; p-value < 0.001). Further, the IPTW-adjusted competing risk model showed significantly lower prostate cancer-specific mortality in patients who received ARPi compared with those who received chemotherapy (IPTW-AHR, 0.48; 95% CI, 0.41 to 0.57; p-value < 0.001). In the without pre-existing history of CVD cohort, the adjusted Cox model showed significantly lower all-cause mortality in patients who received APRi than those who received chemotherapy (IPTW-AHR, 0.49; 95% CI, 0.41-0.60; p-value < 0.001). Whereas the IPTW-adjusted competing risk model showed significantly lower prostate cancer-specific mortality in patients who received ARPi compared with those who received chemotherapy (IPTW-AHR, 0.52; 95% CI, 0.42-0.64; p-value < 0.001). In this population-based cohort of older men with castration-resistant prostate cancer, treatment with oral androgen receptor pathway inhibitors was associated with lower estimated risks of all-cause and prostate cancer-specific mortality compared with chemotherapy in patients with and without pre-existing CVD. These findings add real-world comparative effectiveness evidence for patient populations not well represented in randomized clinical trials; however, given the observational design and limitations of administrative data, residual confounding and unmeasured clinical differences may have influenced the results. Prostate cancer that no longer responds to hormone therapy is called castration‐resistant prostate cancer (CRPC). Two commonly used treatments for CRPC are oral androgen receptor pathway inhibitors (ARPis), which are pills, and chemotherapy, which is given by infusion. Many older men with CRPC also have cardiovascular disease (CVD) (such as heart disease or prior stroke), but these patients are often underrepresented in clinical trials. In this study, we used national Medicare data linked with cancer registry records to examine survival outcomes among 1438 older men with CRPC who received either ARPis or chemotherapy between 2012 and 2014. About half of the patients had pre‐existing CVD. After accounting for differences in patient characteristics between treatment groups, men who received ARPis had lower overall mortality and lower prostate cancer–specific mortality compared with those who received chemotherapy. These patterns were similar among patients with and without CVD. Because this was an observational study using administrative data, unmeasured differences between groups may still have influenced the results. Nonetheless, the findings provide real‐world evidence to help inform treatment decisions for older men with advanced prostate cancer.
Quantification of prescription of antimicrobial agents and use of paediatric outpatient services before, during and after the COVID-19 pandemic. We conducted a population-based study using Norwegian linked health registries and Japanese claims (2018-2023). Paediatric antibiotic prescription rates, broad-spectrum use, and proportion of antibiotic prescriptions with prior presumed bacterial infection diagnoses were analysed monthly, overall and by age groups and sex. Interrupted time series analyses were performed to evaluate pandemic-related changes, expressed in rate ratio (RR) and its CI, using March 2020 as the interruption point and the pre-pandemic trend/level as reference. Data on 5.5 million children and 19.5 million antibiotic prescriptions were analysed. Before the pandemic, antibiotic prescribing was higher in Japan (120-200/1000 children/month) than in Norway (10-20/1000). At pandemic onset, rates fell by 45% in Norway (RR = 0.55; 95% CI, 0.45-0.67) and by 53% in Japan (RR = 0.47; 95% CI, 0.41-0.55), then by 2023 had returned to expected levels. Broad-spectrum antibiotic use was much higher in Japan (70%) compared with Norway (10%) before the pandemic. However, Norway experienced a sharp 20% increase whereas Japan remained largely unchanged post-pandemic. The proportion of prescriptions with a prior presumed bacterial diagnosis was between 50% and 65% before the pandemic then decreased modestly by 5%-10% at pandemic onset, followed by gradual rebound in both countries. The COVID-19 pandemic significantly altered paediatric antibiotic prescribing in both countries. Sustained antibiotic stewardship efforts are needed to ensure appropriate paediatric antibiotic use in the post-pandemic era.
The incidence of attention-deficit hyperactivity disorder (ADHD) in children and young people has increased in recent years. Disease frequency varies according to sociodemographic characteristics. There are seasonal patterns in ADHD diagnosis and prescribing with rates falling during school holidays. COVID-19 societal restrictions may have exacerbated ADHD symptoms. Electronic health records were utilised to examine temporal trends throughout the pandemic in the diagnosis and treatment of ADHD by ethnicity and social deprivation in Greater Manchester, England. We conducted a time-series analysis of all diagnosed episodes of ADHD and associated medication prescribing among patients aged 1-24 years using the Greater Manchester Care Record (GMCR). The 60-month observation period was split into four temporal phases: Pre-pandemic (1/2019-2/2020); Pandemic Phase 1 (3/2020-6/2021); Pandemic Phase 2 (7/2021-12/2022) and Post-Pandemic (1/2023-12/2023). Rate ratios by sex, age, ethnicity, and neighbourhood-level Indices of Multiple Deprivation (IMD) quintile were modelled using negative binomial regression. Overall, ADHD incidence and medication prescribing rates increased throughout the study period. Rates of increase were much higher in females than in males. Particularly large increases in ADHD incidence and medication prescribing were observed in Asian females, with post-pandemic incidence rates being seven times higher compared to the Pre-Pandemic phase. In addition, the ADHD medication prescribing rate was 90% higher for Asian females than for White females. Results showed a large increase in incidence and prescribing rates in the least deprived group, particularly in males where incidence rates increased by 83% compared to the most deprived quintile. ADHD incidence and prescribing rates differ between sociodemographic groups, plausibly due to cultural and behavioural differences in the way ADHD symptoms are presented or perceived. It is therefore important that there is greater understanding of how different demographic subgroups exhibit ADHD behaviour to help ensure timely diagnosis and access to the required support.
The pharmaco-epidemiological research program in kidney transplantation (PERP-KT) aims to evaluate, for the principal maintenance immunosuppressive drugs (MISDs): the influence of model-informed precision dosing on graft and patient survival; long-term exposure-effects relationships; and the benefit-harm balance of their combinations and time sequences in patient groups or clinical settings not adequately evaluated in comparative randomized clinical trials. It also aims to develop a hybrid, dynamic, deep learning model capable of predicting the rate of renal graft function decline, thereby providing a platform for individualized prediction of the benefits and risks associated with MISDs. After obtaining all regulatory andd ethical approvals, de-identified extracts of three national databases were linked to create the nationwide PERP-KT dataset, which is hosted within the highly secure environment of the French Health Data Hub. CRISTAL (the exhaustive registry of the French Agence de la Biomédecine) comprises data from 49 886 kidney donors and the corresponding 47 842 transplant recipients between 2005 and 2020. Following iterative deterministic matching and extensive quality control procedures, CRISTAL data were successfully linked to: the French national Health Data System (SNDS), which records reimbursed healthcare utilization, for 30 782 kidney transplant recipients; and to ISBA, a web-based platform for Bayesian dose adjustment of MISDs that contains pharmacological data, for 17 700 kidney grafts and 17 576 recipients. More than 20 pharmaco-epidemiological studies will leverage the database's extensive follow-up, large population size and richness of clinical, healtcare-utilization, and pharmacological data. This research program on kidney transplantation (PERP‐KT) intends to evaluate: the efficacy of fine‐tuned, individual dose adjustment of the main maintenance immunosuppressive drugs (MISDs) on the longevity of the graft and on patient survival; the long term effects of different blood levels of the main MISDs; the efficacy and adverse effects of the different combinations and time sequences of MISDs in patient groups or clinical conditions not previously evaluated; and to develop an artificial intelligence tool able to predict the speed at which each kidney graft will deteriorate after transplantation. PERP‐KT leverages the CRISTAL registry (managed by the French Agence de la Biomédecine) exhaustively listing kidney transplant procedures in France (49 886 donors and 47 842 recipients between 2005 and 2020), combined with the SNDS (Système National des Données de Santé) that contains all drugs delivered and healthcare activities for 30 782 of them, and ISBA (ImmunoSuppressive Bayesian dose Adjustment website) a database about the dose and blood levels of the main MISDs for 17 700 kidney grafts and 17 576 patients. This unprecedented database will support more than 20 studies on the efficacy and adverse events of MISDs and will foster personalized medicine in kidney transplantation.
Pemafibrate is a novel and selective peroxisome proliferator-activated receptor α modulator that has demonstrated favorable efficacy and safety in phase II or III trails. Compared to fenofibrate, a conventional fibrate, pemafibrate exerts a triglyceride-lowering effect at 1/500 of the dose, and the incidence rate of adverse drug reactions is one-ninth at these doses. The aim of this study was to obtain additional information on the safety of pemafibrate. The Japanese Adverse Drug Event Report (JADER) database was analyzed, and the associations of three fibrates and six statins with rhabdomyolysis and acute renal failure were evaluated using reporting odds ratios and information components. Additionally, preclinical experiments were conducted in rats. Creatine phosphokinase (CK) levels were measured after a single administration of fibrates. The CK levels were also assessed when pemafibrate was co-administered with statins. Data mining of the JADER database suggested associations between all fibrates and statins and rhabdomyolysis; however, no signal for acute renal failure was detected for pemafibrate. Signals for rhabdomyolysis were observed for some fibrates when combined with statins, whereas no signal was detected for pemafibrate. In rats, outlier CK levels were not observed after pemafibrate administration, and no significant increase was observed. CK levels were significantly increased by pravastatin treatment. The JADER analysis suggests that pemafibrate may be a safer alternative to conventional fibrates, with a potentially lower association with rhabdomyolysis and acute renal failure, even when co-administered with statins. Preclinical animal experiments support the results of the JADER analysis. However, large-scale prospective studies are needed to clarify the risks of pemafibrate.