Pharmacoepidemiology Research Articles

Identifying Hemolytic Disease of the Fetus and Newborn within a Large Integrated Health Care System.

This study aims to identify hemolytic disease of the fetus and newborn (HDFN) pregnancies using electronic health records (EHRs) from a large integrated health care system. A retrospective cohort study was performed among pregnant patients receiving obstetrical care at Kaiser Permanente Southern California health care system between January 1, 2008, and June 30, 2022. Using structured (diagnostic/procedural codes, medication, and laboratory records) and unstructured (clinical notes analyzed via natural language processing) data abstracted from EHRs, we extracted HDFN-specific "indicators" (maternal positive antibody test and abnormal antibody titer, maternal/infant HDFN diagnosis and blood transfusion, hydrops fetalis, infant intravenous immunoglobulin [IVIG] treatment, jaundice/phototherapy, and first administrated Rho[D] Immune Globulin) to identify potential HDFN pregnancies. Chart reviews and adjudication were then performed on select combinations of indicators for case ascertainment. HDFN due to ABO alloimmunization alone was excluded. The HDFN frequency and proportion of each combination were fully analyzed. Among the 464,711 eligible pregnancies, a total of 136 pregnancies were confirmed as HDFN pregnancies. The percentage of the HDFN-specific indicators ranged from 0.02% (infant IVIG treatment) to 34.53% (infant jaundice/phototherapy) among the eligible pregnancies, and 32.35% (infant IVIG treatment) to 100% (maternal positive antibody test) among the 136 confirmed HDFN pregnancies. Four combination groups of four indicators, four combination groups of five indicators, and the unique combination of six indicators showed 100% of HDFN pregnancies, while 80.88% of confirmed HDFN pregnancies had the indicator combination of maternal positive antibody test, maternal/infant HDFN diagnosis, and infant jaundice/phototherapy. We successfully identified HDFN pregnancies by leveraging a combination of medical indicators extracted from structured and unstructured data that may be used in future pharmacoepidemiologic studies. Traditional indicators (positive antibody test results, high titers, and clinical diagnosis codes) alone did not accurately identify HDFN pregnancies, highlighting an unmet need for improved practices in HDFN coding. · A case ascertainment method was developed to identify HDFN from structured and unstructured data.. · The method used in this study may be used in future pharmacoepidemiologic studies.. · The study highlighted an unmet need for improved practices in HDFN coding..

2001-2021 Comparative Persistence of Oral Antipsychotics in Patients Initiating Treatment: Superiority of Clozapine in Time-to-Treatment Discontinuation.

Continuous antipsychotic (AP) therapy is crucial for managing psychotic disorders, and its early interruption reflects the drug's failure. Real-world epidemiological research is essential for confirming experimental data and generating new research hypotheses. The persistence of oral APs in a large population sample from 2000 to 2021 was analyzed by comparing AP prescriptions over this period across four Italian provinces, using dispensing data linked via a record-linkage procedure among regional healthcare utilization databases. We calculated personalized daily dosages and assessed time-to-treatment discontinuation over a 3-month period for patients initiating AP treatment. Treatment persistence was evaluated using Kaplan-Meier curves and Cox regression, with adjustments for age and sex. Second-generation antipsychotics (SGAs) were favored over first-generation antipsychotics (FGAs), with olanzapine as the most prescribed. Within the study time frame, 42,434 individuals were prescribed a new continuous AP regimen. The analysis revealed 24 significant differences within 28 comparisons. As a class, SGAs demonstrated better treatment persistence than FGAs (HR: 0.76; 95%CI: 0.73, 0.79). Clozapine stood out for its superior persistence, surpassing all other SGAs, notably olanzapine (HR: 0.85; 95%CI: 0.79-0.91) and risperidone (HR: 0.80; 95%CI: 0.74-0.87). Olanzapine and aripiprazole showed better results than both risperidone and quetiapine. Quetiapine showed inferior 3-month persistence in all pairwise comparisons. The study results provide insight into the performance dynamics among SGAs: clozapine, despite being one of the less frequently dispensed APs in our sample, emerged as a significant prescription choice. The significance of pharmacoepidemiological studies in complementing experimental findings is also underscored.

A Mother-Child data linkage approach using data from the information System for the development of research in primary care (SIDIAP) in Catalonia

BackgroundLarge-scale clinical databases containing routinely collected electronic health records (EHRs) data are a valuable source of information for research studies. For example, they can be used in pharmacoepidemiology studies to evaluate the effects of maternal medication exposure on neonatal and pediatric outcomes. Yet, this type of studies is infeasible without proper mother–child linkage. MethodsWe leveraged all eligible active records (N = 8,553,321) of the Information System for Research in Primary Care (SIDIAP) database. Mothers and infants were linked using a deterministic approach and linkage accuracy was evaluated in terms of the number of records from candidate mothers that failed to link. We validated the mother–child links identified by comparison of linked and unlinked records for both candidate mothers and descendants. Differences across these two groups were evaluated by means of effect size calculations instead of p-values. Overall, we described our data linkage process following the GUidance for Information about Linking Data sets (GUILD) principles. ResultsWe were able to identify 744,763 unique mother–child relationships, linking 83.8 % candidate mothers with delivery dates within a period of 15 years. Of note, we provide a record-level category label used to derive a global confidence metric for the presented linkage process. Our validation analysis showed that the two groups were similar in terms of a number of aggregated attributes. ConclusionsComplementing the SIDIAP database with mother–child links will allow clinical researchers to expand their epidemiologic studies with the ultimate goal of improving outcomes for pregnant women and their children. Importantly, the reported information at each step of the data linkage process will contribute to the validity of analyses and interpretation of results in future studies using this resource.

Analysis of methotrexate use practice in patients with rheumatoid arthritis and psoriatic arthritis

Background. Pharmacoepidemiological studies play a key role in optimizing pharmacotherapy for various diseases. In particular, significant progress in the treatment of rheumatoid arthritis (RA) and psoriatic arthritis (PsA) necessitates a detailed analysis of the consumption of disease-modifying antirheumatic drugs. The modern therapeutic concept for these conditions focuses on achieving and maintaining long-term remission, which positions methotrexate (MTX) as a first-line agent, characterized by high effectiveness, safety and advantageous pharmacoeconomic profiles.Objective: to investigate the characteristics of MTX use in real clinical practice among patients with RA and PsA during outpatient medical care.Material and methods. The study comprised two parts: a retrospective cross-sectional analysis of MTX consumption dynamics in 2018, 2020, and 2023 via ATC/DDD (Anatomical Therapeutic Chemical classification, ATC; defined daily dose, DDD) methodology, and a longitudinal non-interventional pharmacoepidemiological study conducted in 2023 using the medical information system “BARS. Healthcare”.Results. The ATC/DDD analysis demonstrated a significant increase in MTX consumption among patients with RA and PsA: from 302,428.6 to 319,114.3 DDDs per 1,000 patients per year for RA, and from 28,157.1 to 310,771.6 DDDs per 1,000 patients per year for PsA in 2018 and 2023, respectively. The most frequently prescribed dose of MT was 15 mg/week. The primary therapeutic combination for 55.8% of patients was “MT + nonsteroidal anti-inflammatory drugs”, with 32.7% of patients also receiving glucocorticoids (GCs) as part of this combination. The study identified a high frequency of prescriptions for proton pump inhibitors, GCs, and cholecalciferol, highlighting the activity of diseases under consideration, potential complications, and the necessity for the prevention of clinically significant drug interactions.Conclusions. The conducted study confirmed that MTX remains the main drug for the treatment of RA and PsA. The ATC/DDD analysis demonstrated a significant increase in MTX consumption in recent years, correlating with its therapeutic effectiveness and accessibility. The high frequency of concomitant prescriptions underscores the complexity of treating RA and PsA and the need for an interdisciplinary approach to ensure safety and efficacy of the therapy.

Analyzing missingness patterns in real-world data using the SMDI toolkit: application to a linked EHR-claims pharmacoepidemiology study

BackgroundMissing data in confounding variables present a frequent challenge in generating evidence using real-world data, including electronic health records (EHR). Our objective was to apply a recently published toolkit for characterizing missing data patterns and based on the toolkit results about likely missingness mechanisms, illustrate the decision-making process for analyses in an empirical case example.MethodsWe utilized the Structural Missing Data Investigations (SMDI) toolkit to characterize missing data patterns in the context of a pharmacoepidemiology study comparing cardiovascular outcomes of initiating sodium-glucose-cotransporter-2 inhibitors (SGLT2i) and dipeptidyl peptidase‐4 inhibitors (DPP‐4i) among older adults. The study used a linked EHR-Medicare claims dataset from Duke Health patients (2015–2017), focusing on partially observed confounders from EHR data (HbA1c lab and body mass index [BMI] values). Our analysis incorporated SMDI's descriptive functions and diagnostic tests to explore missingness patterns and determine missingness mitigation approaches. We used findings from these investigations to inform estimation of adjusted hazard ratios comparing the two classes of medications.ResultsHigh levels of missingness were noted for important confounding variables including HbA1c (63.6%) and BMI (16.5%). Diagnostic tests resulted in output that described: 1) the distributions of patient characteristics, exposure, and outcome between patients with or without an observed value of the partially observed covariate, 2) the ability to predict missingness based on observed covariates, and 3) estimate if the missingness of a partially observed covariate is differential with respect to the outcome. There was evidence that missingness could be sufficiently described using observed data, which allowed multiple imputation by chained equations using random forests to address missing confounder data in estimating treatment effects. Multiple imputation resulted in improved alignment of effect estimates with previous studies.ConclusionsWe were able to demonstrate the practical application of the SMDI toolkit in a real-world setting. Application of the SMDI toolkit and the resulting insights of potential missingness patterns can inform the choice of appropriate analytic methods and increase transparency of research methods in handling missing data. This type of approach can inform analytic decision making and may increase our ability to generate evidence from real-world data.

Hepatotoxicity of antipsychotics: an exploratory pharmacoepidemiologic and pharmacodynamic study integrating FAERS data and in vitro receptor-binding affinities.

Antipsychotic psychopharmacotherapy is associated with the risk of drug-induced liver injury (DILI). However, understanding specific risk factors remains challenging due to limited data. This study investigates the relationship between receptor binding affinities and occupancies of antipsychotics and their associated hepatotoxic risks. A disproportionality analysis with calculation of the Reporting Odds Ratio (ROR) and the Information Component (IC) was conducted using data fromthe FDA Adverse Event Reporting System (FAERS) to identify signals related to the Standardised MedDRA Query "drug-related hepatic disorders", which served as a proxy for drug-induced hepatotoxicity. This was followed by a pharmacoepidemiologic-pharmacodynamic approach to investigate the relationship between the ROR and substance-related receptor binding affinities and occupancy, which was estimated based on in vitro receptor-binding profiles. Significant signals were identified for several antipsychotics, including chlorpromazine, loxapine, olanzapine, and quetiapine, with chlorpromazine and loxapine showing the highest RORs for DILI. Gender-specific analysis revealed a higher frequency of signals in female patients. Statistically significant negative correlations were identified between the ROR for drug-related hepatic disorders and the affinity for serotonin receptor 5-HT1A (r (17) = -0.68, p = 0.0012), while a positive correlation was observed for cholinergic receptors (r (17) = 0.46, p = 0.048). No significant correlations were found related to other receptors or drug properties. Our findings suggest that the serotonin and probably the cholinergic system may play a role in the development of DILI related to antipsychotic medications. The identification of antipsychotics with a higher association with DILI, such as chlorpromazine, underscores the need for careful monitoring in clinical practice. However, our findings need further longitudinal studies to confirm causality. A better understanding of the associations may inform clinical decision-making, particularly in patients with an increased susceptibility to liver damage.

Treatment of Subclinical Hyperthyroidism and Incident Atrial Fibrillation.

Treating overt hyperthyroidism prevents atrial fibrillation (AF). Though subclinical hyperthyroidism (SH) has been associated with AF, it is unknown whether treating SH prevents AF. We aimed to identify the association between treating SH and incident AF. In a pharmacoepidemiologic retrospective cohort study, patients diagnosed with SH between 2000 and 2021 were followed. Outpatients ≥ 18 years with biochemical SH and without prior AF, hypothyroidism, thyroid cancer, pituitary disease, or pregnancy were included. The primary outcome was incident AF. Secondary outcomes were ECG and echocardiographic features associated with AF. Of 2169 patients screened, 360 (131 treated and 229 untreated) were followed up for a mean of 4.27 years. In the treated and untreated groups, AF occurred in 4 (3.1%) and 15 (6.6%) patients (p = 0.15), and AF incidence was 0.8% and 1.4%/year (p = 0.31), respectively. The hazard ratio (HR) for treatment as a time-dependent variable was 0.60 (95% CI 0.19-1.92; p = 0.39). As some cases of AF were documented nearly simultaneously with SH treatment, a sensitivity analysis was performed reassigning two patients diagnosed with AF < 30 days after starting SH treatment to the untreated group. Here, in the treated and untreated groups, AF occurred in 1.6% and 7.4% (p = 0.02), and AF incidence was 0.4% and 1.8%/year (p = 0.02), respectively. The HR was 0.25 (0.06-1.13; p = 0.07). There were no differences in ECG or echocardiographic features. There was an overall trend towards lower incidence and prevalence of AF following treatment of SH, supporting the need for larger scale studies.

Keep Your Guard Up: The Potential Impact of Drug Shortages on Pharmacoepidemiological Studies.

Keep Your Guard Up: The Potential Impact of Drug Shortages on Pharmacoepidemiological Studies.

Core Concepts in Pharmacoepidemiology: Quantitative Bias Analysis.

Pharmacoepidemiological studies provide important information on the safety and effectiveness of medications, but the validity of study findings can be threatened by residual bias. Ideally, biases would be minimized through appropriate study design and statistical analysis methods. However, residual biases can remain, for example, due to unmeasured confounders, measurement error, or selection into the study. A group of sensitivity analysis methods, termed quantitative bias analyses, are available to assess, quantitatively and transparently, the robustness of study results to these residual biases. These approaches include methods to quantify how the estimated effect would be altered under specified assumptions about the potential bias, and methods to calculate bounds on effect estimates. This article introduces quantitative bias analyses for unmeasured confounding, misclassification, and selection bias, with a focus on their relevance and application to pharmacoepidemiological studies.

Identifying Elective Induction of Labor Among a Diverse Pregnant Population from Electronic Health Records within a Large Integrated Healthcare System.

Distinguishing between medically indicated induction of labor (iIOL) and elective induction of labor (eIOL) is a daunting process for researchers. We aimed to develop a Natural Language Processing (NLP) algorithm to identify eIOLs from electronic health records (EHR) within a large integrated healthcare system. We used structured and unstructured data from Kaiser Permanente Southern California's EHR of patients who were <35 years old and had singleton deliveries between 37 and 40 gestational weeks. Induction of labor (IOL) pregnancies were identified if there was evidence of an IOL diagnosis code, procedure code, or documentation in a delivery flowsheet or progress note. A comprehensive NLP algorithm was developed and refined through an iterative process of chart reviews and adjudications, where IOL-associated reasons (medically indicated versus elective induction) were reviewed. The final algorithm was applied to discern the indications of IOLs performed during the study period. A total of 332,163 eligible pregnancies were identified between 01/01/2008-12/31/2022. Of these eligible pregnancies, 68,541 (20.6%) were IOL, of which 6,824 (10.0%) were eIOL. Validation of the NLP process against 300 randomly selected pregnancies (100 eIOL, iIOL, and non-IOL cases each) yielded a positive predictive value of 83.0% and 88.0% for eIOL and iIOL, respectively. The rates of eIOL among the maternal age groups ranged between 9.6% -10.3%, except for the < 20 years group (12.2%). Non-Hispanic White individuals had the highest rate of eIOL (13.2%), while non-Hispanic Asian/Pacific Islanders had the lowest rate of eIOL (7.8%). The rate of eIOL increased from 1.0% in the 37 weeks gestational age group to 20.6% the 40 weeks gestational age group. Findings suggest that the developed NLP algorithm effectively identifies eIOL. It can be utilized to support eIOL-related pharmaco-epidemiological studies, filling in knowledge gaps and providing content more relevant to researchers.

The Challenges of Measuring Socioeconomic Inequality in Pharmacoepidemiology Studies.

Social conditions like socioeconomic status (SES) are critical sources of health disparities. In pharmacoepidemiology research, our ability to measure SES in retrospective, real world clinical data remains challenged by a lack of patient-reported data. Some broadly accepted concepts can be measured at the individual level, such as income, poverty, and education. Community-level measures such as the Centers for Disease Control and Prevention's Social Vulnerability Index (SVI) also exist. After reflecting on these existing measures and discussing the challenges for leveraging them with real world data, we offer three recommendations that we believe could improve the ability of pharmacoepidemiologists to better measure and interrogate the effect of SES in their own research. These recommendations include a greater collection of patient-reported metrics, reduced reliance on ZIP Codes and ZIP Code Tabulation Areas for creating community-level measures of deprivation, and the inclusion of GIS and demography specialists within pharmacoepidemiology teams.

A discovery and verification approach to pharmacovigilance using electronic healthcare data.

Pharmacovigilance is vital for drug safety. The process typically involves two key steps: initial signal generation from spontaneous reporting systems (SRSs) and subsequent expert review to assess the signals' (potential) causality and decide on the appropriate action. We propose a novel discovery and verification approach to pharmacovigilance based on electronic healthcare data. We enhance the signal detection phase by introducing an ensemble of methods which generated signals are combined using Borda count ranking; a method designed to emphasize consensus. Ensemble methods tend to perform better when data is noisy and leverage the strengths of individual classifiers, while trying to mitigate some of their limitations. Additionally, we offer the committee of medical experts with the option to perform an in-depth investigation of selected signals through tailored pharmacoepidemiological studies to evaluate their plausibility or spuriousness. To illustrate our approach, we utilize data from the German Pharmacoepidemiological Research Database, focusing on drug reactions to the direct oral anticoagulant rivaroxaban. In this example, the ensemble method is built upon the Bayesian confidence propagation neural network, longitudinal Gamma Poisson shrinker, penalized regression and random forests. We also conduct a pharmacoepidemiological verification study in the form of a nested active comparator case-control study, involving patients diagnosed with atrial fibrillation who initiated anticoagulant treatment between 2011 and 2017. The case study reveals our ability to detect known adverse drug reactions and discover new signals. Importantly, the ensemble method is computationally efficient. Hasty false conclusions can be avoided by a verification study, which is, however, time-consuming to carry out. We provide an online tool for easy application: https://borda.bips.eu.

Concomitant Administration of Psychotropic and Prostate Cancer Drugs: A Pharmacoepidemiologic Study Using Drug-Drug Interaction Databases.

Prostate cancer (PC) represents the second most common diagnosed cancer in men. The burden of diagnosis and long-term treatment may frequently cause psychiatric disorders in patients, particularly depression. The most common PC treatment option is androgen deprivation therapy (ADT), which may be associated with taxane chemotherapy. In patients with both PC and psychiatric disorders, polypharmacy is frequently present, which increases the risk of drug-drug interactions (DDIs) and drug-related adverse effects. Therefore, this study aimed to conduct a pharmacoepidemiologic study of the concomitant administration of PC drugs and psychotropics using three drug interaction databases (Lexicomp®, drugs.com®, and Medscape®). This study assayed 4320 drug-drug combinations (DDCs) and identified 814 DDIs, out of which 405 (49.63%) were pharmacokinetic (PK) interactions and 411 (50.37%) were pharmacodynamic (PD) interactions. The most common PK interactions were based on CYP3A4 induction (n = 275, 67.90%), while the most common PD interactions were based on additive torsadogenicity (n = 391, 95.13%). Proposed measures for managing the identified DDIs included dose adjustments, drug substitutions, supplementary agents, parameters monitoring, or simply the avoidance of a given DDC. A significant heterogenicity was observed between the selected drug interaction databases, which can be mitigated by cross-referencing multiple databases in clinical practice.

Association between the antibiotics use and recurrence in patients with resected colorectal cancer: EVADER-1, a nation-wide pharmaco-epidemiologic study

BackgroundThe impact of antibiotics (ATBs) on the risk of colorectal cancer (CRC) recurrence after curative resection remains unknown. MethodsUsing the French nation-wide database of cancer patients, all newly diagnosed non-metastatic CRC patients resected between 01/2012 and 12/2014 were included. The perioperative ATB intake (from 6 months before surgery until 1 year after) was classified according to the class, the period of use (pre- vs post-resection), the disease stage (localized and locally advanced), and the primary tumor location (colon and rectum/junction). The primary endpoint was the 3-year disease-free survival (DFS). The impact of ATB was assessed using time-dependent multivariate Cox models. ResultsA total of 35,496 CRC patients were included. Seventy-nine percent of patients had at least one ATB intake. Outpatient ATB intake after surgery was associated with unfavorable 3-year DFS. The ATBs associated with decreased 3-year DFS were cephalosporins, streptogramins, quinolones, penicillin A with beta-lactamase inhibitors, and antifungals with differential effects according to the primary tumor location and disease stage. ConclusionThese findings suggest that ATBs modulate the risk of recurrence after early CRC resection with a differential impact of the ATB classes depending on disease stage and tumor site.

Contemporary Practice and Considerations for Real-World Data Source Identification and Feasibility Assessment.

There has been rapid growth in the variety and number of real-world data (RWD) sources, as well as the number of regulatory documents that provide guidance for assessing the suitability of RWD sources for pharmacoepidemiology studies. This study aims to assess differences in RWD guidance and variability in current practice for identifying and assessing RWD for studies with regulatory purpose. Key criteria for feasibility assessment were mapped against relevant regulatory guidance documents across US, EU, and Asia-Pacific regions. An online survey was designed and deployed to International Society for Pharmacoepidemiology members to understand current practice. Findings were summarized and used to inform key considerations and recommendations. Eleven RWD guidance documents were identified and mapped against 14 RWD assessment criteria. Variability was seen across these documents in guidance for these criteria. Between December 2022 and January 2023, 37 survey respondents reported having used RWD for post-marketing commitments (34, 92%) and/or background epidemiology (28, 76%). RWD were mostly identified through literature (33, 89%) and data landscaping (26, 70%); guidance documents referenced included: Food and Drug Administration (20, 54%), European Network for Centres for Pharmacoepidemiology and Pharmacovigilance (17, 46%), European Medical Agency (16, 43%), and Structured Process to Identify Fit-For-Purpose Data (11, 30%). Challenges for conducting feasibility assessments included RWD accessibility, ability to complete validation, and RWD provider responsiveness. Existing guidelines are used extensively by researchers, but key criteria for RWD identification and feasibility assessment are not reflected consistently and challenges remain. Recommendations have been made reflecting study findings.

Validity assessment of self-reported medication use in a pharmacoepidemiologic study by comparison with prescription record review

Objectives Although self-reported questionnaires are widely used to collect information on medication use in epidemiological studies, their validity for studies involving older adults has not been sufficiently assessed. This study evaluated the validity of self-reported medication use using questionnaires in comparison with drug notebooks.Methods The study enrolled 370 older community dwellers who participated in an aging sub-study survey of the Tsuruoka Metabolomics Cohort Study between April 2019 and March 2021. Medication use was assessed by comparing self-reported questionnaire data with drug notebook records. We analyzed medications used for hypertension, dyslipidemia, myocardial infarction, angina, diabetes, rheumatism, osteoporosis/metabolic bone disease, constipation, anxiety/depression, dementia, asthma, allergy, thrombosis, and thyroid disease. Moreover, gastrointestinal (GI) medications, steroids, and antipyretic analgesics were assessed, and data on injectable medications for osteoporosis/metabolic bone disease was collected. Using drug notebook records, we identified regular medication users by assessing whether they had received oral medication prescriptions covering over 28 days and took the medication within the 90 days preceding the day of their survey. To define medication categories, we used Anatomical Therapeutic Chemical (ATC) classification codes. Sensitivity, specificity, and kappa statistics were calculated for each medication using drug notebooks as standards. Those who did not bring their drug notebooks on the day of the survey were defined as non-medication users.Results The mean age (standard deviation) of the 370 participants (146 men and 224 women) was 73.3 (4.0) years. The sensitivity and specificity for each medication were as follows: hypertension (0.97, 0.97), dyslipidemia (0.93, 0.98), myocardial infarction (0.24, 0.99), diabetes (0.94, 1.00), rheumatism (1.00, 1.00), osteoporosis/metabolic bone disease (0.82, 0.99), constipation (0.71, 0.98), GI conditions (0.63, 0.97), anxiety/depression (0.36, 1.00), dementia (0.67, 1.00), asthma (0.67, 0.98), allergy (0.57, 0.99), thrombosis (0.88, 0.98), steroids (0.80, 0.99), thyroid disease (1.00, 1.00) and antipyretic analgesics (0.75, 0.96).Conclusions Although sensitivity and specificity differed by medication categories, the results of our population-based cohort study suggested that self-reported questionnaires on medication use among older adults are valid, especially for medications with high sensitivity (≥ 0.8).

Dicloxacillin is an inducer of intestinal P-glycoprotein but neither dicloxacillin nor flucloxacillin increases the risk of stroke/systemic embolism in direct oral anticoagulant users.

We aimed to assess if dicloxacillin/flucloxacillin reduces the therapeutic efficacy of direct oral anticoagulants (DOACs) and the underlying molecular mechanism. In a randomized, crossover study, we assessed whether dicloxacillin reduces oral absorption of drugs through P-glycoprotein (P-gp) during 10 and 28 days of treatment. To study the impact of dicloxacillin/flucloxacillin on intestinal and hepatic expression of P-gp in vitro, we usd LS174T cells and 3D spheroids of primary human hepatocytes. Finally, we used nationwide Danish health registries and the UK's Clinical Practice Research Datalink to estimate hazard ratios (HRs) for the risk of stroke and systemic embolism following dicloxacillin/flucloxacillin exposure among DOAC users, using phenoxymethylpenicillin and amoxicillin as active comparators. Dicloxacillin reduced the area under the curve of dabigatran to a geometric mean ratio 10days of 0.67 (95% confidence interval [CI]: 0.42-1.1) and geometric mean ratio 28 days of 0.72 (95% CI: 0.39-1.4), suggesting reduced oral absorption via increased P-gp expression. In vitro, dicloxacillin raised P-gp expression in both intestinal and liver cells, while flucloxacillin only affected liver cells. In the pharmacoepidemiologic study, dicloxacillin and flucloxacillin were not associated with increased risk of stroke/systemic embolism (dicloxacillin vs. phenoxymethylpenicillin HR: 0.93, 95% CI: 0.72-1.2; flucloxacillin vs. amoxicillin HR: 0.89, 95% CI: 0.51-1.5). Dicloxacillin increases expression of intestinal P-gp, leading to reduced oral absorption of dabigatran. However, concomitant use of dicloxacillin/flucloxacillin was not associated with stroke and systemic embolism among DOAC users, suggesting no clinical impact from the drug-drug interaction between dicloxacillin/flucloxacillin and DOACs.

High-Dimensional Propensity Score and Its Machine Learning Extensions in Residual Confounding Control

“The use of health care claims datasets often encounters criticism due to the pervasive issues of omitted variables and inaccuracies or mis-measurements in available confounders. Ultimately, the treatment effects estimated using such data sources may be subject to residual confounding. Digital electronic administrative records routinely collect a large volume of health-related information; and many of which are usually not considered in conventional pharmacoepidemiological studies. A high-dimensional propensity score (hdPS) algorithm was proposed that uses such information as surrogates or proxies for mismeasured and unobserved confounders in an effort to reduce residual confounding bias. Since then, many machine learning and semi-parametric extensions of this algorithm have been proposed to better exploit the wealth of high-dimensional proxy information. In this tutorial, we will (i) demonstrate logic, steps and implementation guidelines of hdPS using an open data source as an example (using reproducible R codes), (ii) familiarize readers with the key difference between propensity score versus hdPS, as well as the requisite sensitivity analyses, (iii) explain the rationale for using the machine learning and double robust extensions of hdPS, and (iv) discuss advantages, controversies, and hdPS reporting guidelines while writing amanuscript.

Improving the quality of pharmacoepidemiological studies using the target trial emulation framework.

Improving the quality of pharmacoepidemiological studies using the target trial emulation framework.

The Population-level Effect of Adjuvant Therapies on Breast Cancer Recurrence: Application of the Trend-in-Trend Design.

Breast cancer has an average 10-year relative survival reaching 84%. This favorable survival is due, in part, to the introduction of biomarker-guided therapies. We estimated the population-level effect of the introduction of two adjuvant therapies-tamoxifen and trastuzumab-on recurrence using the trend-in-trend pharmacoepidemiologic study design. We ascertained data on women diagnosed with nonmetastatic breast cancer who were registered in the Danish Breast Cancer Group clinical database. We used the trend-in-trend design to estimate the population-level effect of the introduction of (1) tamoxifen for postmenopausal women with estrogen receptor (ER)-positive breast cancer in 1982, (2) tamoxifen for premenopausal women diagnosed with ER-positive breast cancer in 1999, and (3) trastuzumab for women <60 years diagnosed with human epidermal growth factor receptor 2-positive breast cancer in 2007. For the population-level effect of the introduction of tamoxifen among premenopausal women diagnosed with ER-positive breast cancer in 1999, the risk of recurrence decreased by nearly one-half (OR = 0.52), consistent with evidence from clinical trials; however, the estimate was imprecise (95% confidence interval [CI] = 0.25, 1.85). We observed an imprecise association between tamoxifen use and recurrence from the time it was introduced in 1982 (OR = 1.24 95% CI = 0.46, 5.11), inconsistent with prior knowledge from clinical trials. For the introduction of trastuzumab in 2007, the estimate was also consistent with trial evidence, though imprecise (OR = 0.51; 95% CI = 0.21, 22.4). We demonstrated how novel pharmacoepidemiologic analytic designs can be used to evaluate the routine clinical care and effectiveness of therapeutic advancements in a population-based setting while considering some limitations of the approach.