Variance estimation in inverse probability weighted Cox models

Uncertain about uncertainty in matching-adjusted indirect comparisons? A simulation study to compare methods for variance estimation

In health technology assessment, matching-adjusted indirect comparison (MAIC) is the most common method for pairwise comparisons that control for imbalances in baseline characteristics across trials. One of the primary challenges in MAIC is the need to properly account for the additional uncertainty introduced by the matching process. Limited evidence and guidance are available on variance estimation in MAICs. Therefore, we conducted a comprehensive Monte Carlo simulation study to evaluate the performance of different statistical methods across 108 scenarios. Four general approaches for variance estimation were compared in both anchored and unanchored MAICs of binary and time-to-event outcomes: (1) conventional estimators (CE) using raw weights; (2) CE using weights rescaled to the effective sample size (ESS); (3) robust sandwich estimators; and (4) bootstrapping. Several variants of sandwich estimators and bootstrap methods were tested. Performance was quantified on the basis of empirical coverage probabilities for 95% confidence intervals and variability ratios. Variability was underestimated by CE + raw weights when population overlap was poor or moderate. Despite several theoretical limitations, CE + ESS weights accurately estimated uncertainty across most scenarios. Original implementations of sandwich estimators had a downward bias in MAICs with a small ESS, and finite sample adjustments led to marked improvements. Bootstrapping was unstable if population overlap was poor and the sample size was limited. All methods produced valid coverage probabilities and standard errors in cases of strong population overlap. Our findings indicate that the sample size, population overlap, and outcome type are important considerations for variance estimation in MAICs.

Balancing versus modelling in weighted analysis of non-randomised studies with survival outcomes: A simulation study

Weighting methods are widely used for causal effect estimation in non-randomised studies. In general, these methods use the propensity score (PS), the probability of receiving the treatment given the covariates, to arrive at the respective weights. All of these "modelling" methods actually optimize prediction of the respective outcome, which is, in the PS model, treatment assignment. However, this does not match with the actual aim of weighting, which is eliminating the association between covariates and treatment assignment. In the "balancing" approach, covariates are thus balanced directly by solving systems of numerical equations, explicitly without fitting a PS model. To compare modelling, balancing and hybrid approaches to weighting we performed a large simulation study for a binary treatment and a survival outcome. For maximal practical relevance all simulation parameters were selected after a systematic review of medical studies that used PS methods for analysis. We also introduce a new hybrid method that uses the idea of the covariate balancing propensity score and matching weights, thus avoiding extreme weights. In addition, we present a corrected robust variance estimator for some of the methods. Overall, our simulations results indicate that balancing approach methods work worse than expected. However, among the considered balancing methods, entropy balancing consistently outperforms the variance balancing approach. All methods estimating the average treatment effect in the overlap population perform well with very little bias and small standard errors even in settings with misspecified propensity score models. Finally, the coverage using the standard robust variance estimator was too high for all methods, with the proposed corrected robust variance estimator improving coverage in a variety of settings.

Air pollution and serious bleeding events in high-risk older adults

IMPORTANCE

Despite biological plausibility, very few epidemiologic studies have investigated the risks of clinically significant bleeding events due to particulate air pollution.

OBJECTIVE

To measure the independent and synergistic effects of PM2.5 exposure and anticoagulant use on serious bleeding events.

DESIGN

Retrospective cohort study (2008-2016).

SETTING

Nationwide Medicare population.

PARTICIPANTS

A 50% random sample of Medicare Part D-eligible Fee-for-Service beneficiaries at high risk for cardiovascular and thromboembolic events.

EXPOSURES

Fine particulate matter (PM2.5) and anticoagulant drugs (apixaban, dabigatran, edoxaban, rivaroxaban, or warfarin).

MAIN OUTCOMES AND MEASURES

The outcomes were acute hospitalizations for gastrointestinal bleeding, intracranial bleeding, or epistaxis. Hazard ratios and 95% CIs for PM2.5 exposure were estimated by fitting inverse probability weighted marginal structural Cox proportional hazards models. The relative excess risk due to interaction was used to assess additive-scale interaction between PM2.5 exposure and anticoagulant use.

RESULTS

The study cohort included 1.86 million high-risk older adults (mean age 77, 60% male, 87% White, 8% Black, 30% anticoagulant users, mean PM2.5 exposure 8.81 μg/m3). A 10 μg/m3 increase in PM2.5 was associated with a 48% (95% CI: 45%-52%), 58% (95% CI: 49%-68%) and 55% (95% CI: 37%-76%) increased risk of gastrointestinal bleeding, intracranial bleeding, and epistaxis, respectively. Significant additive interaction between PM2.5 exposure and anticoagulant use was observed for gastrointestinal and intracranial bleeding.

CONCLUSIONS

Among older adults at high risk for cardiovascular and thromboembolic events, increasing PM2.5 exposure was significantly associated with increased risk of gastrointestinal bleeding, intracranial bleeding, and epistaxis. In addition, PM2.5 exposure and anticoagulant use may act together to increase risks of severe gastrointestinal and intracranial bleeding. Thus, clinicians may recommend that high-risk individuals limit their outdoor air pollution exposure during periods of increased PM2.5 concentrations. Our findings may inform environmental policies to protect the health of vulnerable populations.

Joint analysis of vaccination effectiveness and antiviral drug effectiveness for COVID-19: a causal inference approach

OBJECTIVES

This study aims to estimate the causal effects of oral antivirals and vaccinations in the prevention of all-cause mortality and progression to severe COVID-19 in an integrative setting with both antivirals and vaccinations considered as interventions.

METHODS

We identified hospitalized adult patients (i.e. aged 18 or above) in Hong Kong with confirmed SARS-CoV-2 infection between March 16, 2022, and December 31, 2022. An inverse probability-weighted (IPW) Andersen-Gill model with time-dependent predictors was used to address immortal time bias and produce causal estimates for the protection effects of oral antivirals and vaccinations against severe COVID-19.

RESULTS

Given prescription is made within 5 days of confirmed infection, nirmatrelvir-ritonavir is more effective in providing protection against all-cause mortality and development into severe COVID-19 than molnupiravir. There was no significant difference between CoronaVac and Comirnaty in the effectiveness of reducing all-cause mortality and progression to severe COVID-19.

CONCLUSIONS

The use of oral antivirals and vaccinations causes lower risks of all-cause mortality and progression to severe COVID-19 for hospitalized SARS-CoV-2 patients.

Higher methadone dose at time of release from prison predicts linkage to maintenance treatment for people with HIV and opioid use disorder transitioning to the community in Malaysia

BACKGROUND

Incarcerated people with HIV and opioid-dependence often experience poor post-release outcomes in the absence of methadone maintenance treatment (MMT). In a prospective trial, we assessed the impact of methadone dose achieved within prison on linkage to MMT after release.

METHODS

From 2010 to 2014, men with HIV (N = 212) and opioid dependence before incarceration were enrolled in MMT within 6 months of release from Malaysia's largest prison and followed for 12-months post-release. As a prospective trial, allocation to MMT was at random and later by preference design (predictive nonetheless). MMT dosing was individually targeted to minimally achieve 80 mg/day. Time-to-event analyses were conducted to model linkage to MMT after release.

FINDINGS

Of the 212 participants allocated to MMT, 98 (46 %) were prescribed higher dosages (≥80 mg/day) before release. Linkage to MMT after release occurred in 77 (36 %) participants and significantly higher for those prescribed higher dosages (46% vs 28 %; p = 0.011). Factors associated with higher MMT dosages were being married, on antiretroviral therapy, longer incarceration periods, having higher levels of depression, and methadone preference compared to randomization. After controlling for other variables, being prescribed higher methadone dosage (aHR: 2.53, 95 %CI: 1.42-4.49) was the only independent predictor of linkage to methadone after release.

INTERPRETATION

Higher doses of methadone prescribed before release increased the likelihood of linkage to MMT after release. Methadone dosing should be introduced into international guidelines for treatment of opioid use disorder in prisons and further post-release benefits should be explored.

FUNDING

National Institute of Drug Abuse (NIDA).

Immuno-Molecular Targeted Therapy Use and Survival Benefit in Patients with Stage IVB Cervical Carcinoma in Commission on Cancer®-Accredited Facilities in the United States

PURPOSE

To investigate IMT use and survival in real-world stage IVB cervical cancer patients outside randomized clinical trials.

METHODS

Patients diagnosed with stage IVB cervical cancer during 2013-2019 in the National Cancer Database and treated with chemotherapy (CT) ± external beam radiation (EBRT) ± intracavitary brachytherapy (ICBT) ± IMT were studied. The adjusted hazard ratio (AHR) and 95% confidence interval (CI) for risk of death were estimated in patients treated with vs. without IMT after applying propensity score analysis to balance the clinical covariates.

RESULTS

There were 3164 evaluable patients, including 969 (31%) who were treated with IMT. The use of IMT increased from 11% in 2013 to 46% in 2019. Age, insurance, facility type, sites of distant metastasis, and type of first-line treatment were independently associated with using IMT. In propensity-score-balanced patients, the median survival was 18.6 vs. 13.1 months for with vs. without IMT (p < 0.001). The AHR was 0.72 (95% CI = 0.64-0.80) for adding IMT overall, 0.72 for IMT + CT, 0.66 for IMT + CT + EBRT, and 0.69 for IMT + CT + EBRT + ICBT. IMT-associated survival improvements were suggested in all subgroups by age, race/ethnicity, comorbidity score, facility type, tumor grade, tumor size, and site of metastasis.

CONCLUSIONS

IMT was associated with a consistent survival benefit in real-world patients with stage IVB cervical cancer.

Statistical inference for time-to-event data in non-randomized cohorts with selective attrition

In multi-season clinical trials with a randomize-once strategy, patients enrolled from previous seasons who stay alive and remain in the study will be treated according to the initial randomization in subsequent seasons. To address the potentially selective attrition from earlier seasons for the non-randomized cohorts, we develop an inverse probability of treatment weighting method using season-specific propensity scores to produce unbiased estimates of survival functions or hazard ratios. Bootstrap variance estimators are used to account for the randomness in the estimated weights and the potential correlations in repeated events within each patient from season to season. Simulation studies show that the weighting procedure and bootstrap variance estimator provide unbiased estimates and valid inferences in Kaplan-Meier estimates and Cox proportional hazard models. Finally, data from the INVESTED trial are analyzed to illustrate the proposed method.

Efficacy of a novel patient-focused intervention aimed at increasing adherence to guideline-based preventive measures in asplenic patients: the PrePSS trial

PURPOSE

To determine whether a novel intervention improves the adherence to guideline-based preventive measures in asplenic patients at risk of post-splenectomy sepsis (PSS).

METHODS

We used a prospective controlled, two-armed historical control group design to compare a novel, health action process approach (HAPA)-based telephonic intervention involving both patients and their general practitioners to usual care. Eligible patients were identified in cooperation with the insurance provider AOK Baden-Wuerttemberg, Germany. Patients with anatomic asplenia (n = 106) were prospectively enrolled and compared to a historical control group (n = 113). Comparisons were done using a propensity-score-based overlap-weighting model. Adherence to preventive measures was quantified by the study-specific 'Preventing PSS score' (PrePSS score) which includes pneumococcal and meningococcal vaccination status, the availability of a stand-by antibiotic and a medical alert card.

RESULTS

At six months after the intervention, we estimated an effect of 3.96 (95% CI 3.68-4.24) points on the PrePSS score scale (range 0-10) with mean PrePSS scores of 3.73 and 7.70 in control and intervention group, respectively. Substantial improvement was seen in all subcategories of the PrePSS score with the highest absolute gains in the availability of stand-by antibiotics. We graded the degree of participation by the general practitioner (no contact, short contact, full intervention) and noted that the observed effect was only marginally influenced by the degree of physician participation.

CONCLUSIONS

Patients who had received the intervention exhibited a significantly higher adherence to guideline-based preventive measures compared to the control group. These data suggest that widespread adoption of this pragmatic intervention may improve management of asplenic patients. Health insurance provider-initiated identification of at-risk patients combined with a patient-focused intervention may serve as a blueprint for a wide range of other preventive efforts leading to patient empowerment and ultimately to better adherence to standards of care.

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.

Federated causal inference in heterogeneous observational data

We are interested in estimating the effect of a treatment applied to individuals at multiple sites, where data is stored locally for each site. Due to privacy constraints, individual-level data cannot be shared across sites; the sites may also have heterogeneous populations and treatment assignment mechanisms. Motivated by these considerations, we develop federated methods to draw inferences on the average treatment effects of combined data across sites. Our methods first compute summary statistics locally using propensity scores and then aggregate these statistics across sites to obtain point and variance estimators of average treatment effects. We show that these estimators are consistent and asymptotically normal. To achieve these asymptotic properties, we find that the aggregation schemes need to account for the heterogeneity in treatment assignments and in outcomes across sites. We demonstrate the validity of our federated methods through a comparative study of two large medical claims databases.

Methods for comparative effectiveness based on time to confirmed disability progression with irregular observations in multiple sclerosis

Real-world data sources offer opportunities to compare the effectiveness of treatments in practical clinical settings. However, relevant outcomes are often recorded selectively and collected at irregular measurement times. It is therefore common to convert the available visits to a standardized schedule with equally spaced visits. Although more advanced imputation methods exist, they are not designed to recover longitudinal outcome trajectories and typically assume that missingness is non-informative. We, therefore, propose an extension of multilevel multiple imputation methods to facilitate the analysis of real-world outcome data that is collected at irregular observation times. We illustrate multilevel multiple imputation in a case study evaluating two disease-modifying therapies for multiple sclerosis in terms of time to confirmed disability progression. This survival outcome is derived from repeated measurements of the Expanded Disability Status Scale, which is collected when patients come to the healthcare center for a clinical visit and for which longitudinal trajectories can be estimated. Subsequently, we perform a simulation study to compare the performance of multilevel multiple imputation to commonly used single imputation methods. Results indicate that multilevel multiple imputation leads to less biased treatment effect estimates and improves the coverage of confidence intervals, even when outcomes are missing not at random.

Effects of a health psychology-based intervention for patients with asplenia on psychological determinants of preventive behaviour: A propensity score analysis

OBJECTIVE

Patients with asplenia have an increased lifelong risk of severe infections especially post splenectomy sepsis with hospital mortality rates of 30-50%. Adherence to existing guidelines for preventive measures is low. Objective of the study is the evaluation of a novel intervention to increase health psychological outcomes in patients with asplenia resulting in better adherence to preventive measures.

METHODS

The intervention was evaluated by conducting a prospective, two-armed historical control group design via propensity score analysis. Focus are health-psychological outcomes: self-efficacy, intention, risk perception, behaviour planning, self-management, health literacy, patient involvement and disease-knowledge.

RESULTS

Patients in the intervention group (N = 110) showed a higher increase in almost all outcomes compared to a historical control group (N = 115). The strongest increase was observed in "asplenia-specific self-management" (average treatment effect: ATE 1.14 [95% CI 0.91-1.36] p < .001) and "asplenia-specific health-literacy" (ATE 1.42 [95% CI 1.18-1.65] p < .001). Significant intervention effects were also found in behaviour planning, perceived involvement and disease-knowledge.

CONCLUSION

The patient-focused intervention is effective in improving health-psychological outcomes in patients with asplenia.

PRACTICE IMPLICATIONS

The implementation of the intervention can make an important contribution to care and lead to an improvement of health-psychological outcomes that may result in a higher adherence to prevention measures.

Confounding adjustment in the analysis of augmented randomized controlled trial with hybrid control arm

The augmented randomized controlled trial (RCT) with hybrid control arm includes a randomized treatment group (RT), a smaller randomized control group (RC), and a large synthetic control (SC) group from real-world data. This kind of trial is useful when there is logistics and ethics hurdle to conduct a fully powered RCT with equal allocation, or when it is necessary to increase the power of the RCT by incorporating real-world data. A difficulty in the analysis of augmented RCT is that the SC and RC may be systematically different in the distribution of observed and unmeasured confounding factors, causing bias when the two control groups are analyzed together as hybrid controls. We propose to use propensity score (PS) analysis to balance the observed confounders between SC and RC. The possible bias caused by unmeasured confounders can be estimated and tested by analyzing propensity score adjusted outcomes from SC and RC. We also propose a partial bias correction (PBC) procedure to reduce bias from unmeasured confounding. Extensive simulation studies show that the proposed PS + PBC procedures can improve the efficiency and statistical power by effectively incorporating the SC into the RCT data analysis, while still control the estimation bias and Type I error inflation that might arise from unmeasured confounding. We illustrate the proposed statistical procedures with data from an augmented RCT in oncology.

Gender-specific outcomes in immune checkpoint inhibitor therapy for advanced or metastatic urothelial cancer: a systematic review and meta-analysis

PURPOSE

To analyze gender-specific differences in survival parameters in advanced or metastatic urothelial cancer patients undergoing immune checkpoint inhibition.

METHODS

The primary aim of this systematic review and meta-analysis was to evaluate gender-specific differences in disease-free (DFS), progression-free (PFS), cancer-specific survival (CSS), event-free survival (EFS), overall survival (OS) and objective response rate (ORR). The sources MEDLINE, Embase and Cochrane Library were systematically searched from January 2010 to June 2022. No restrictions were made concerning language, study region or publication type. A comparison of gender-specific differences in survival parameters was performed using a random-effects meta-analysis. A risk of bias assessment was done using the ROBINS-I tool.

RESULTS

Five studies were included. In a random-effect meta-analysis of the studies, PCD4989g and IMvigor 211 with both using atezolizumab, females were more likely to have better objective response rate (ORR) than men (OR 2.24; 95% CI 1.20-4.16; p = 0.0110). In addition, females had a comparable median OS to men (MD 1.16; 95% CI - 3.15-5.46; p = 0.598). In summary, comparing all results, a tendency was seen toward better response rates and survival parameters in female patients. The risk of bias assessment yielded an overall low risk of bias.

CONCLUSIONS

There is a tendency toward better outcomes in women for immunotherapy in advanced or metastatic urothelial cancer, but only for the antibody atezolizumab women have a significantly better ORR. Unfortunately, many studies fail to report gender-specific outcomes. Therefore, further research is essential when aiming for individualized medicine. This research should address immunological confounders.

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.

Adjuvant pelvic radiation versus observation in intermediate-risk early-stage cervical cancer patients following primary radical surgery: a propensity score-adjusted analysis

OBJECTIVE

To compare survival outcomes, posttreatment complications, and quality of life (QoL) of early-stage cervical cancer patients with intermediate-risk factors between those who received adjuvant pelvic radiation and those without adjuvant treatment.

METHODS

Stages IB-IIA cervical cancer patients classified as having intermediate-risk following primary radical surgery were included. After propensity score weighted adjustment, all baseline demographic and pathological characteristics of 108 women who received adjuvant radiation and 111 women who had no adjuvant treatment were compared. The primary outcomes were progression-free survival (PFS) and overall survival (OS). The secondary outcomes included treatment-related complications and QoL.

RESULTS

Median follow-up time was 76.1 months in the adjuvant radiation group and 95.4 months in the observation group. The 5-year PFS (91.6% in the adjuvant radiation group and 88.4% in the observation group, p=0.42) and OS (90.1% in the adjuvant radiation group and 93.5% in the observation group, p=0.36) were not significantly different between the groups. There was no significant association between adjuvant treatment and overall recurrence/death in the Cox proportional hazard model. However, a substantial reduction in pelvic recurrence was observed in participants with adjuvant radiation (hazard ratio=0.15; 95% confidence interval=0.03-0.71). Grade 3/4 treatment-related morbidities and QoL scores were not significantly different between the groups.

CONCLUSION

Adjuvant radiation was associated with a lower risk of pelvic recurrence. However, its significant benefit in reducing overall recurrence and improving survival in early-stage cervical cancer patients with intermediate-risk factors could not be demonstrated.

Marginal structural models for multilevel clustered data

Marginal structural models (MSMs), which adopt inverse probability treatment weighting in the estimating equations, are powerful tools to estimate the causal effects of time-varying exposures in the presence of time-dependent confounders. Motivated by the Conservation of Hearing Study (CHEARS) Audiology Assessment Arm (AAA) where repeated hearing measurements were clustered by study participants, time, and testing sites, we propose two methods to account for the multilevel correlation structure when fitting the MSMs. The first method directly models the covariance of the repeated outcomes when solving the weighted generalized estimating equations for MSMs, while the second two-stage analysis approach fits cluster-specific MSMs first and then combines the estimated parameters using mixed-effects meta-analysis. Finite sample simulation results suggest that our methods can obtain less biased and more efficient estimates of the parameters by accounting for the multilevel correlation. Moreover, we explore the effects of using fixed- or mixed-effects model to estimate the treatment probability on the parameter estimates of the MSMs in the presence of unmeasured cluster-level confounders. Lastly, we apply our methods to the CHEARS AAA data set, to estimate the causal effects of aspirin use on hearing loss.

Recommendations for the use of propensity score methods in multiple sclerosis research

BACKGROUND

With many disease-modifying therapies currently approved for the management of multiple sclerosis, there is a growing need to evaluate the comparative effectiveness and safety of those therapies from real-world data sources. Propensity score methods have recently gained popularity in multiple sclerosis research to generate real-world evidence. Recent evidence suggests, however, that the conduct and reporting of propensity score analyses are often suboptimal in multiple sclerosis studies.

OBJECTIVES

To provide practical guidance to clinicians and researchers on the use of propensity score methods within the context of multiple sclerosis research.

METHODS

We summarize recommendations on the use of propensity score matching and weighting based on the current methodological literature, and provide examples of good practice.

RESULTS

Step-by-step recommendations are presented, starting with covariate selection and propensity score estimation, followed by guidance on the assessment of covariate balance and implementation of propensity score matching and weighting. Finally, we focus on treatment effect estimation and sensitivity analyses.

CONCLUSION

This comprehensive set of recommendations highlights key elements that require careful attention when using propensity score methods.

A Review of Causal Inference for External Comparator Arm Studies

Randomized controlled trials (RCTs) are the gold standard design to establish the efficacy of new drugs and to support regulatory decision making. However, a marked increase in the submission of single-arm trials (SATs) has been observed in recent years, especially in the field of oncology due to the trend towards precision medicine contributing to the rise of new therapeutic interventions for rare diseases. SATs lack results for control patients, and information from external sources can be compiled to provide context for better interpretability of study results. External comparator arm (ECA) studies are defined as a clinical trial (most commonly a SAT) and an ECA of a comparable cohort of patients-commonly derived from real-world settings including registries, natural history studies, or medical records of routine care. This publication aims to provide a methodological overview, to sketch emergent best practice recommendations and to identify future methodological research topics. Specifically, existing scientific and regulatory guidance for ECA studies is reviewed and appropriate causal inference methods are discussed. Further topics include sample size considerations, use of estimands, handling of different data sources regarding differential baseline covariate definitions, differential endpoint measurements and timings. In addition, unique features of ECA studies are highlighted, specifically the opportunity to address bias caused by unmeasured ECA covariates, which are available in the SAT.

Statistical methods of indirect comparison with real-world data for survival endpoint under non-proportional hazards

In clinical studies that utilize real-world data, time-to-event outcomes are often germane to scientific questions of interest. Two main obstacles are the presence of non-proportional hazards and confounding bias. Existing methods that could adjust for NPH or confounding bias, but no previous work delineated the complexity of simultaneous adjustments for both. In this paper, a propensity score stratified MaxCombo and weighted Cox model is proposed. This model can adjust for confounding bias and NPH and can be pre-specified when NPH pattern is unknown in advance. The method has robust performance as demonstrated in simulation studies and in a case study.

Impact of neutropenia on clinical manifestations and outcome of Staphylococcus aureus bloodstream infection - A propensity score-based overlap weight analysis in two large, prospectively evaluated cohorts

OBJECTIVE

To investigate whether neutropenia influenced mortality and long-term outcome of Staphylococcus aureus bloodstream infection (SAB).

METHODS

Data from two prospective, multicentre cohort studies (INSTINCT and ISAC) conducted in 20 tertiary care hospitals in 6 countries between 2006 and 2015 were analysed. Neutropenic and severely neutropenic patients (defined by the proxy of total white blood cell count <1000/μl and <500/μl, respectively, at onset of SAB) were compared to a control group using a propensity score model and overlap weights to adjust for baseline characteristics. Overall survival and time to SAB-related late complications (SAB recurrence, infective endocarditis, osteomyelitis, or other deep-seated manifestations) were analysed by Cox regression and competing risk analyses, respectively.

RESULTS

Of 3,187 patients, 102 were neutropenic and 70 were severely neutropenic at onset of SAB. Applying overlap weights yielded two groups of 83 neutropenic and 220 non-neutropenic patients, respectively. Baseline characteristics of these groups were exactly balanced. In the Cox regression analysis, we observed no significant difference in survival between the two groups (death during follow-up: 36.1 % in neutropenic vs. 30.6 % in non-neutropenic patients, hazard ratio 1.21 (95 % CI 0.79-1.83)). This finding remained unchanged when we considered severely neutropenic patients (hazard ratio 1.08 [0.60; 1.94]). Competing risk analysis showed a cause-specific hazard ratio (CSHR) of 0.39 (95 % CI 0.11-1.39) for SAB-related late-complications in neutropenic patients.

CONCLUSIONS

Neutropenia was not associated with a higher survival during follow-up. The lower rate of SAB-related late complications in neutropenic patients should be validated in other cohorts.

The Impact of Left Atrial Reduction During Surgical Ablation of Atrial Fibrillation

Enlarged left atrium (LA) is a risk factor for ablation failure after atrial fibrillation (AF) surgery. It predisposes patients to thromboembolic events, even in successful ablation; therefore, concomitant resection of the LA wall during surgical ablation was introduced. This study examined the clinical impacts of LA reduction in patients undergoing concomitant ablation for AF. This study enrolled 1484 patients with enlarged LA (≥50 mm) who underwent surgical AF ablation during major cardiac surgery between January 2001 and August 2018. Among them, 876 (59%) patients underwent concomitant LA reduction (Reduction group), whereas in the remaining 608 (41%), the LA wall was unresected (Preservation group). The primary outcome of interest was overall stroke. The secondary outcomes were overall mortality, late recurrence of AF, early postoperative complications and postoperative echocardiographic parameters. Outcomes were compared after adjusting baseline characteristics with inverse probability of treatment weighting (IPTW) using propensity score. The median follow-up was 60.1 months. After IPTW adjustment, long-term mortality (P = 0.250) and AF-free rates (P = 0.196) did not significantly differ between groups. However, the Reduction group showed a decreased risk of stroke (hazard ratio 0.54; 95% confidence interval 0.32-0.90; P = 0.018). Early postoperative complications rate such as mortality or reoperation for bleeding, was not significantly different between the 2 groups. The Reduction group showed smaller LA diameter (50.6 ± 8.0 mm vs 53.6 ± 8.9 mm; P < 0.001) on follow-up echocardiography. LA reduction effectively decreased LA size and appeared to decrease the stroke risk in patients with enlarged LA undergoing ablation for AF.