A simple Cox approach to estimating risk ratios without sharing individual-level data in multi-site studies.

Epidemiologic studies frequently use risk ratios to quantify associations between exposures and binary outcomes. When the data are physically stored at multiple data partners, it can be challenging to perform individual-level analysis if data cannot be pooled centrally due to privacy constraints. Existing methods either require multiple file transfers between each data partner and an analysis center (e.g., distributed regression) or only provide approximate estimation of the risk ratio (e.g., meta-analysis). Here we develop a practical method that requires a single transfer of eight summary-level quantities from each data partner. Our approach leverages an existing risk-set method and software originally developed for Cox regression. Sharing only summary-level information, the proposed method provides risk ratio estimates and confidence intervals identical to those that would be provided - if individual-level data were pooled - by the modified Poisson regression. We justify the method theoretically, confirm its performance using simulated data, and implement it in a distributed analysis of COVID-19 data from the U.S. Food and Drug Administration's Sentinel System.

Direct-Acting Oral Anticoagulants and Antiseizure Medications for Atrial Fibrillation and Epilepsy and Risk of Thromboembolic Events.

Importance: Direct-acting oral anticoagulants (DOACs) are commonly prescribed with antiseizure medications (ASMs) due to concurrency of and the association between atrial fibrillation (AF) and epilepsy. However, enzyme-inducing (EI) ASMs may reduce absorption and accelerate metabolism of DOACs, potentially lowering DOAC levels and elevating thromboembolism risk. Objective: To assess the rates of thromboembolic and major bleeding events in adults with AF and epilepsy dispensed DOACs and EI ASMs vs DOACs with non-EI ASMs. Design, Setting, and Participants: This active-comparator, new-user cohort study included US health care data from the Clinformatics Data Mart database from October 2010 to September 2021 for a nationally representative population of adults with AF and epilepsy. Exposure: Evaluations included episodes of contiguous coadministration of DOACs for AF with EI ASMs (exposed) or non-EI ASMs (referent) for epilepsy. Main Outcomes and Measures: Thromboembolic events (primary outcome) and major bleeding events (secondary outcome) were identified based on a series of validated, diagnosis-based coding algorithms. Data-adaptive, high-dimensional propensity score matching was used to control for observed confounders and proxies for unobserved confounders. Adjusted hazard ratios (AHRs) were estimated using Cox proportional hazards regression models with robust variance estimators to account for clustering within matched pairs. Results: This study included 14 078 episodes (median age, 74 [IQR, 67-81]; 52.4% female) and 14 158 episodes (median age, 74 [IQR, 67-81]; 52.4% female) of incident DOAC and ASM use that met eligibility criteria for assessment of thromboembolic and major bleeding outcomes, respectively. Incidence was 88.5 per 1000 person-years for thromboembolic events and 68.3 per 1000 person-years for bleeding events. Compared with use of non-EI ASMs, use of EI ASMs with DOACs was not associated with a difference in risk of thromboembolic events (AHR, 1.10; 95% CI, 0.82-1.46) but was associated with a reduction in risk of major bleeding events (AHR, 0.63; 95% CI, 0.44-0.89). Conclusions and Relevance: In this cohort study, EI ASMs were not associated with alteration in DOAC efficacy. Further research is needed on the reduction in bleeding risk associated with EI ASMs, as this may suggest that pharmacokinetic interactions are associated with lowering DOAC levels without negating therapeutic effects.

Extending the DeLong algorithm for comparing areas under correlated receiver operating characteristic curves with missing data.

A nonparametric method proposed by DeLong et al in 1988 for comparing areas under correlated receiver operating characteristic curves is used widely in practice. However, the DeLong method as implemented in popular software quietly deletes individuals with any missing values, yielding potentially invalid and/or inefficient results. We simplify the DeLong algorithm using ranks and extend it to accommodate missing data by using a mixed model approach for multivariate data. Simulation results demonstrate the validity and efficiency of our procedure for data missing at random. We illustrate our proposed procedure in SAS, Stata, and R using the original DeLong data.

Antibiotic use among extremely low birth-weight infants from 2009 to 2021: a retrospective observational study.

OBJECTIVE: To assess trends in antibiotic use across a large cohort of extremely low birth-weight (<1000 g; ELBW) infants admitted to academic and community neonatal intensive care units (NICUs) across the USA over a 13-year period. DESIGN: Repeated cross-sectional cohort study. SETTING: Premier Health Database, a comprehensive administrative database of inpatient encounters from academic and community hospitals across the US. PATIENTS: ELBW inborn infants admitted to NICUs from 1 January 2009 to 31 December 2021. INTERVENTIONS: N/A MAIN OUTCOME MEASURES: Absolute and relative changes in (1) proportion of ELBW infants with antibiotic exposure and (2) days of therapy (DOT) per 1000 patient days, over time. Average annual differences were estimated using generalised linear regression with 95% CI. Disposition trends were also measured. RESULTS: Among 36 701 infants admitted to 402 NICUs, the proportion exposed to antibiotics was essentially unchanged (89.9% in 2009 to 89.3% in 2021; absolute reduction of -0.6%); generalised linear regression estimated an annual absolute difference of -0.3% (95% CI (-0.6%) to (-0.07%); p=0.01). DOT per 1000 patient days decreased from 337 in 2009 to 210 in 2021, a 37.8% relative difference and annual relative difference of -4.3% ((-5.2%) to (-3.5%); p<0.001). Mortality was unchanged during the study period. CONCLUSIONS: We found a substantial reduction in antibiotic DOT despite no substantive change in the proportion of infants exposed to antibiotics. This suggests the success of stewardship efforts aimed at antibiotic duration and highlight the need for improved approaches to identifying ELBW infants at highest risk of infection.

Trends in C-Reactive Protein Use in Early-onset Sepsis Evaluations and Associated Antibiotic Use.

OBJECTIVE: To determine the prevalence of C-reactive protein (CRP) use in early-onset sepsis (EOS) evaluations in neonatal intensive care units (NICUs) across the US over time and to determine the association between CRP use and antibiotic use. STUDY DESIGN: A retrospective cohort study of NICUs contributing data to Premier Healthcare Database from 2009 through 2021. EOS evaluation was defined as a blood culture charge ≤ 3 days after birth. CRP use for each NICU was calculated as the proportion of infants with a CRP test obtained ≤ 3 days after birth among those undergoing an EOS evaluation and categorized as, low (<25%); medium-low (25 to < 50%), medium-high (50 to < 75%), and high (≥75%). Outcomes included antibiotic use and mortality ≤ 7 days after birth. RESULTS: Among 572 NICUs, CRP use varied widely and was associated with time. The proportion of NICUs with high CRP use decreased from 2009 to 2021 (24.7% vs 17.4%, P < .001), and those with low CRP use increased (47.9% vs 64.8%, P < .001). Compared with low-use NICUs, high-use NICUs more frequently continued antibiotics > 3 days (10% vs 25%, P < .001). This association persisted in multivariable-adjusted regression analyses (adjusted risk ratio 1.95, 95%CI 1.54, 2.48). Risk of mortality was not different in high-use NICUs (adjusted risk difference -0.02%, 95%CI -0.04%, 0.0008%). CONCLUSIONS: CRP use in EOS evaluations varied widely across NICUs. High CRP use was associated with prolonged antibiotic therapy but not mortality ≤ 7 days after birth. Reducing routine CRP use in EOS evaluations may be a target for neonatal antibiotic stewardship efforts.

Thinking Three-Dimensionally: A Self- and Externally-Controlled Approach to Screening for Drug-Drug-Drug Interactions Among High-Risk Populations.

The global rise in polypharmacy has increased both the necessity and complexity of drug-drug interaction (DDI) assessments, given the growing potential for interactions involving more than two drugs. Leveraging large-scale healthcare claims data, we piloted a semi-automated, high-throughput case-crossover-based approach for drug-drug-drug interaction (3DI) screening. Cases were direct-acting oral anticoagulant (DOAC) users with either a major bleeding event during ongoing dispensings for potentially interacting, enzyme-inhibiting antihypertensive drugs (AHDs) (Study 1), or a thromboembolic event during ongoing dispensings for potentially interacting, enzyme-inducing antiseizure medications (ASMs) (Study 2). 3DI detection was based on screening for additional drug exposures that served as acute outcome triggers. To mitigate direct effects and confounding by concomitant drugs, self-controlled estimates were adjusted using negative cases (external "control" DOAC users with the same outcomes but co-dispensings for non-interacting AHDs or ASMs). Signal thresholds were set based on P-values and false discovery rate q-values to address multiple comparisons. Study 1: 285 drugs were examined among 3,306 episodes. Self-controlled assessments with q-value thresholds yielded 9 3DI signals (cases) and 40 DDI signals (negative cases). External adjustment generated 10 3DI signals from the P-value threshold and no signals from the q-value threshold. Study 2: 126 drugs were examined among 604 episodes. Assessments with P-value thresholds yielded 3 3DI and 26 DDI signals following self-control, as well as 4 3DI signals following adjustment. No 3DI signals met the q-value threshold. The presented self- and externally-controlled approach aimed to advance paradigms for real-world higher order drug interaction screening among high-susceptibility populations with pre-existent DDI risk.

Gabapentin as a novel adjunct for postoperative irritability after superior cavopulmonary connection operation in children.

OBJECTIVES: Describing our institution's off-label use of gabapentin to treat irritability after superior cavopulmonary connection surgery and its impact on subsequent opiate and benzodiazepine requirements. METHODS: This is a single-center retrospective cohort study including infants who underwent superior cavopulmonary connection operation between 2011 and 2019. RESULTS: Gabapentin was administered in 74 subjects (74/323, 22.9%) during the observation period, with a median (IQR) starting dose of 5.7 (3.3, 15.0) mg/kg/day and a maximum dose of 10.7 (5.5, 23.4) mg/kg/day. Infants who underwent surgery in 2015-19 were more likely to receive gabapentin compared with those who underwent surgery in 2011-14 (p < 0.0001). Infants prescribed gabapentin were younger at surgery (137 versus 146 days, p = 0.007) and had longer chest tube durations (1.8 versus 0.9 days, p < 0.001), as well as longer postoperative intensive care (5.8 versus 3.1 days, p < 0.0001) and hospital (11.5 versus 7.0 days, p < 0.0001) lengths of stays. The year of surgery was the only predisposing factor associated with gabapentin administration in multivariate analysis. In adjusted linear regression, infants prescribed gabapentin on postoperative day 0-4 (n = 64) had reduced benzodiazepine exposure in the following 3 days (-0.29 mg/kg, 95% CI -0.52 - -0.06, p = 0.01) compared with those not prescribed gabapentin, while no difference was seen in opioid exposure (p = 0.59). CONCLUSIONS: Gabapentin was used with increasing frequency during the study period. There was a modest reduction in benzodiazepine requirements associated with gabapentin administration and no reduction in opioid requirements. A randomised controlled trial could better assess gabapentin's benefits postoperatively in children with congenital heart disease.

Circulating microRNA-33b levels are associated with the presence and severity of coronary heart disease.

MicroRNA-33b (miR-33b) affected various biological pathways in regulating cholesterol homeostasis which may link to the pathogenesis of atherosclerotic lesions. However, whether this marker is associated with the presence and severity of coronary heart disease (CHD) is undetermined. We aim to explore the diagnostic value of circulating miR-33b level in the presence and severity of CHD. Altogether 320 patients were enrolled, including 240 patients diagnosed with CHD while 80 were classified as controls after CAG examination. Circulating miR-33b level was analyzed in all subjects, the Gensini score was calculated to assess the severity of stenotic lesions. The association between miR-33b and the presence and severity of CHD was analyzed, and the diagnostic potential of miR-33b of CHD was performed by the receiver operating characteristic (ROC) analysis. The CHD group had higher miR-33b levels (p < 0.001), and the miR-33b content significantly elevated following an increasing Gensini score (p for trend < 0.001). After adjustments for potential risk factors, such as several blood lipid markers, miR-33b remained a significant determinant for CHD (p < 0.001). ROC analysis disclosed that the AUC was 0.931. The optimal cutoff value of miR-33b was with a sensitivity of 81.3% and a specificity of 98.7% in differentiating CHD. It can prognosticate that the higher level of miR-33b was linked to increased severity of disease in CHD patients. Thus, the application of this marker might assist in the diagnosis and classification of CHD patients. Nevertheless, additional studies with larger sample sizes will be required to verify these results.

Factors Associated with Growth in a Cohort of Children with Complex Biventricular Congenital Heart Disease.

OBJECTIVES: To evaluate patterns and determinants of longitudinal growth among children requiring complex biventricular repair for congenital heart disease, as well as to assess for associations of growth with early feeding modality, comorbidities, postoperative complications, and socioeconomic characteristics. STUDY DESIGN: A single-institution retrospective cohort study was performed in children born February 1999 to March 2009 with complex congenital heart disease who underwent biventricular repair before age 4 years, defined by Risk Adjustment in Congenital Heart Surgery-1 category 3-5. Clinical characteristics, height, weight, and body mass index (BMI) from ages 2-12 years were collected by chart review. Neighborhood-level socioeconomic data were identified using a geographic information system approach. The adjusted association of covariates with growth outcomes was estimated using multivariable linear regression models using generalized estimating equations. RESULTS: Compared with population growth curves, the cohort (n = 150) trended toward early decrease in age-adjusted weight and height. Early tube feeding was significantly associated with decreased BMI before adolescence (-0.539; 95% CI -1.02, -0.054; P = .029). In addition, other clinical and perioperative characteristics had significant associations with growth, including low birth weight, preoperative tube feeds, need for multiple bypass runs, and diagnosis of feeding disorder. CONCLUSIONS: Early childhood growth in children with complex biventricular repair may be impaired. Early tube feeding was associated with decreased BMI over the course of early childhood, which may indicate a need for continued close nutrition follow-up and support even beyond the duration of tube feeds.

Outcome of BMI2+: Motivational Interviewing to Reduce BMI Through Primary Care AAP PROS Practices.

BACKGROUND: Pediatric obesity rates in the United States remain at an all-time high. Pediatric primary care clinicians and registered dietitians can help treat childhood obesity, and motivational interviewing (MI) has shown promising effects in prior trials. METHODS: We randomized 18 pediatric primary care practices to receive the Brief Motivational Interviewing to Reduce BMI or BMI2+ intervention or continue with usual care (UC). Practices were recruited through the American Academy of Pediatrics Pediatric Research in Office Settings network. The intervention comprised 4 components1: in-person and telehealth MI counseling by pediatric clinicians; 4 recommended sessions,2 6 telephone MI counseling sessions from a registered dietitian,3 text message reminders and tailored motivational messages, and4 parent educational materials. The main outcome was the change in the percentage of the 95th percentile of BMI. The study was conducted 2017 through 2021. RESULTS: There was a significant treatment x time interaction (b = 0.017, 95% confidence interval: [0.0066-0.027]) for the main outcome, favoring the UC group, with youth in the intervention arm showing a greater relative increase in their percent of the 95th percentile. CONCLUSIONS: There was no overall benefit of the intervention and, contrary to expectations, youth in the intervention arm gained more weight, based on percent of the distance from the 95th percentile than matched youth from UC practices. The absolute excess weight gain among intervention relative to UC youth was small, approximately 0.5 BMI units and 1 kg over 2 years. We offer several potential explanations for these unexpected findings.

Causal Effects of Stochastic PrEP Interventions on HIV Incidence Among Men Who Have Sex With Men.

Antiretroviral preexposure prophylaxis (PrEP) is highly effective in preventing human immunodeficiency virus (HIV) infection, but uptake has been limited and inequitable. Although interventions to increase PrEP uptake are being evaluated in clinical trials among men who have sex with men (MSM), those trials cannot evaluate effects on HIV incidence. Estimates from observational studies of the causal effects of PrEP-uptake interventions on HIV incidence can inform decisions about intervention scale-up. We used longitudinal electronic health record data from HIV-negative MSM accessing care at Fenway Health, a community health center in Boston, Massachusetts, from January 2012 through February 2018, with 2 years of follow-up. We considered stochastic interventions that increased the chance of initiating PrEP in several high-priority subgroups. We estimated the effects of these interventions on population-level HIV incidence using a novel inverse-probability weighted estimator of the generalized g-formula, adjusting for baseline and time-varying confounders. Our results suggest that even modest increases in PrEP initiation in high-priority subgroups of MSM could meaningfully reduce HIV incidence in the overall population of MSM. Interventions tailored to Black and Latino MSM should be prioritized to maximize equity and impact.

Multiply robust causal inference of the restricted mean survival time difference.

The hazard ratio (HR) remains the most frequently employed metric in assessing treatment effects on survival times. However, the difference in restricted mean survival time (RMST) has become a popular alternative to the HR when the proportional hazards assumption is considered untenable. Moreover, independent of the proportional hazards assumption, many comparative effectiveness studies aim to base contrasts on survival probability rather than on the hazard function. Causal effects based on RMST are often estimated via inverse probability of treatment weighting (IPTW). However, this approach generally results in biased results when the assumed propensity score model is misspecified. Motivated by the need for more robust techniques, we propose an empirical likelihood-based weighting approach that allows for specifying a set of propensity score models. The resulting estimator is consistent when the postulated model set contains a correct model; this property has been termed multiple robustness. In this report, we derive and evaluate a multiply robust estimator of the causal between-treatment difference in RMST. Simulation results confirm its robustness. Compared with the IPTW estimator from a correct model, the proposed estimator tends to be less biased and more efficient in finite samples. Additional simulations reveal biased results from a direct application of machine learning estimation of propensity scores. Finally, we apply the proposed method to evaluate the impact of intrapartum group B streptococcus antibiotic prophylaxis on the risk of childhood allergic disorders using data derived from electronic medical records from the Children's Hospital of Philadelphia and census data from the American Community Survey.

Antibiotic Use Among Infants Admitted to Neonatal Intensive Care Units.

This cross-sectional study examines antibiotic exposure, days of therapy, types of antibiotics, and changes in use patterns among newborns in neonatal intensive care units (NICUs) across the US from 2009 to 2021.

Combining meta-analysis with multiple imputation for one-step, privacy-protecting estimation of causal treatment effects in multi-site studies.

Missing data complicates statistical analyses in multi-site studies, especially when it is not feasible to centrally pool individual-level data across sites. We combined meta-analysis with within-site multiple imputation for one-step estimation of the average causal effect (ACE) of a target population comprised of all individuals from all data-contributing sites within a multi-site distributed data network, without the need for sharing individual-level data to handle missing data. We considered two orders of combination and three choices of weights for meta-analysis, resulting in six approaches. The first three approaches, denoted as RR + metaF, RR + metaR and RR + std, first combined results from imputed data sets within each site using Rubin's rules and then meta-analyzed the combined results across sites using fixed-effect, random-effects and sample-standardization weights, respectively. The last three approaches, denoted as metaF + RR, metaR + RR and std + RR, first meta-analyzed results across sites separately for each imputation and then combined the meta-analysis results using Rubin's rules. Simulation results confirmed very good performance of RR + std and std + RR under various missing completely at random and missing at random settings. A direct application of the inverse-variance weighted meta-analysis based on site-specific ACEs can lead to biased results for the targeted network-wide ACE in the presence of treatment effect heterogeneity by site, demonstrating the need to clearly specify the target population and estimand and properly account for potential site heterogeneity in meta-analyses seeking to draw causal interpretations. An illustration using a large administrative claims database is presented.

Ulinastatin inhibits the metastasis of nasopharyngeal carcinoma by involving uPA/uPAR signaling.

Distant metastasis is the primary reason for treatment failure in patients with nasopharyngeal carcinoma (NPC). In this study, we investigated the effect of ulinastatin (UTI) on NPC metastasis and its underlying mechanism. Highly-metastatic NPC cell lines S18 and 58F were treated with UTI and the effect on cell proliferation, migration, and invasion were determined by MTS and Transwell assays. S18 cells with luciferase-expressing (S18-1C3) were injected into the left hind footpad of nude mice to establish a model of spontaneous metastasis from the footpad to popliteal lymph node (LN). The luciferase messenger RNA (mRNA) was measured by quantitative polymerase chain reaction (qPCR), and the metastasis inhibition rate was calculated. Key molecular members of the UTI-related uPA, uPAR, and JAT/STAT3 signaling pathways were detected by qPCR and immunoblotting. UTI suppressed the migration and infiltration of S18 and 5-8F cells and suppressed the metastasis of S18 cells in vivo without affecting cell proliferation. uPAR expression decreased from 24 to 48 h after UTI treatment. The antimetastatic effect of UTI is partly due to the suppression of uPA and uPAR. UTI partially suppresses NPC metastasis by downregulating the expression of uPA and uPAR.

Patterns in the Development of Pediatric Allergy.

OBJECTIVES: Describe clinical and epidemiologic patterns of pediatric allergy using longitudinal electronic health records (EHRs) from a multistate consortium of US practices. METHODS: Using the multistate Comparative Effectiveness Research through Collaborative Electronic Reporting EHR database, we defined a cohort of 218 485 children (0-18 years) who were observed for ≥5 years between 1999 and 2020. Children with atopic dermatitis (AD), immunoglobulin E-mediated food allergy (IgE-FA), asthma, allergic rhinitis (AR), and eosinophilic esophagitis (EoE) were identified using a combination of diagnosis codes and medication prescriptions. We determined age at diagnosis, cumulative incidence, and allergic comorbidity. RESULTS: Allergic disease cumulative (and peak age of) incidence was 10.3% (4 months) for AD, 4.0% (13 months) for IgE-FA, 20.1% (13 months) for asthma, 19.7% (26 months) for AR, and 0.11% (35 months) for EoE. The most diagnosed IgE-FAs were peanut (1.9%), egg (0.8%), and shellfish (0.6%). A total of 13.4% of children had ≥2 allergic conditions, and respiratory allergies (ie, asthma, AR) were commonly comorbid with each other, and with other allergic conditions. CONCLUSIONS: We detail pediatric allergy patterns using longitudinal, health care provider-based data from EHR systems across multiple US states and varied pediatric practice types. Our results support the population-level allergic march progression and indicate high rates of comorbidity among children with food and respiratory allergies.

Off-brand: A 6-year study of medication brand and generic name usage in a multifacility academic healthcare system.

BACKGROUND: Usage of medication brand names in electronic health records may introduce conflicts of interest, perpetuate false perceptions of brand superiority, alter prescribing practices, and cause confusion leading to errors. OBJECTIVE: We sought to identify the frequency of brand name medication usage in clinical documentation, as well as factors associated with increased usage. DESIGNS, SETTINGS, AND PARTICIPANTS: We conducted a retrospective analysis of all clinical documentation written at our healthcare system (a multifacility academic urban healthcare system) between 2015 and 2020. MAIN OUTCOMES AND MEASURES: We used string-matching and regular expressions to identify medication mentions. We conducted bivariate analyses to identify associations between brand name usage and author-, note-, and medication-level factors, and a multivariate Poisson regression to clarify independent associations between individual factors and brand usage. RESULTS: A total of 104,456,653 notes from 37,285 unique authors were included in our analysis. A total of 162,906,009 medication mentions were identified, of which 36.0% were brand name mentions with a steady year-over-year decrease. Factors associated with the usage of a brand name include: author role, years since release, length and syllabic complexity of the generic name, service type, and encounter context. Over-the-counter availability did not affect usage. There was sizable individual variation between note writers.

Trends and Persistent Disparities in Child Obesity During the COVID-19 Pandemic.

The COVID-19 pandemic has been associated with increases in pediatric obesity and widening pre-existing disparities. To better understand the pandemic's long-term impacts, we evaluated trends in obesity across different demographic groups during the pandemic through December 2022. Using a retrospective cohort design, we analyzed electronic health record data from a large pediatric primary care network. Logistic regression models fit using generalized estimating equations estimated odds ratios (ORs) for changes in the level and trajectory of obesity across 2-year month-matched periods: prepandemic (June 2017 to December 2019) and pandemic (June 2020 to December 2022). Among a cohort of 153,667 patients with visits in each period, there was a significant increase in the level of obesity at the pandemic onset [OR: 1.229, 95% confidence interval (CI): 1.211-1.247] followed by a significant decrease in the trend for obesity (OR: 0.993, 95% CI: 0.992-0.993). By December 2022, obesity had returned to prepandemic levels. However, persistent sociodemographic disparities remain.

Revisiting sample size planning for receiver operating characteristic studies: A confidence interval approach with precision and assurance.

Estimation of areas under receiver operating characteristic curves and their differences is a key task in diagnostic studies. Here we develop closed-form sample size formulas for such studies with a focus on estimation rather than hypothesis testing, by explicitly incorporating pre-specified precision and assurance, with precision denoted by the lower limit of confidence interval and assurance denoted by the probability of achieving that lower limit. For sample size estimation purposes, we introduce a normality-based variance function for valid estimation allowing for unequal variances of observations in the disease and non-disease groups. Simulation results demonstrate that the proposed formulas produce empirical assurance probability close to the pre-specified assurance probability and empirical coverage probability close to the nominal level. Compared with a frequently used existing variance function, the proposed function provides more accurate and efficient sample size estimates. For an illustration of the proposed formulas, we present real-world worked examples. To facilitate implementation, we have developed an online calculator openly available at https://dishu.page/calculator/.

Robust causal inference of drug-drug interactions.

There is growing interest in developing causal inference methods for multi-valued treatments with a focus on pairwise average treatment effects. Here we focus on a clinically important, yet less-studied estimand: causal drug-drug interactions (DDIs), which quantifies the degree to which the causal effect of drug A is altered by the presence versus the absence of drug B. Confounding adjustment when studying the effects of DDIs can be accomplished via inverse probability of treatment weighting (IPTW), a standard approach originally developed for binary treatments and later generalized to multi-valued treatments. However, this approach generally results in biased results when the propensity score model is misspecified. Motivated by the need for more robust techniques, we propose two empirical likelihood-based weighting approaches that allow for specifying a set of propensity score models, with the second method balancing user-specified covariates directly, by incorporating additional, nonparametric constraints. The resulting estimators from both methods are consistent when the postulated set of propensity score models contains a correct one; this property has been termed multiple robustness. In this paper, we derive two multiply-robust estimators of the causal DDI, and develop inference procedures. We then evaluate their finite sample performance through simulation. The results demonstrate that the proposed estimators outperform the standard IPTW method in terms of both robustness and efficiency. Finally, we apply the proposed methods to evaluate the impact of renin-angiotensin system inhibitors (RAS-I) on the comparative nephrotoxicity of nonsteroidal anti-inflammatory drugs (NSAID) and opioids, using data derived from electronic medical records from a large multi-hospital health system.