Marie C. Bradley

BACKGROUND: Fee-for-service (FFS) Medicare linkage to clinical registries has been used to assess medical device performance, quality improvement, and postmarket surveillance, but prior linkages have Medicare Advantage (MA). We assessed the accuracy and implications of linking registry data to combined Medicare FFS and MA claims. METHODS: Patients ≥65 years old in the Society of Thoracic Surgeons/American College of Cardiology Transcatheter Valve Therapy Registry (TVT) undergoing transcatheter aortic valve replacement (TAVR) or mitral transcatheter edge-to-edge repair were linked using indirect identifiers to MA and FFS claims from January 1, 2016 to December 31, 2021. Primary outcomes were the linkage rate and representativeness of the linked cohort, assessed by comparing patient characteristics and unadjusted 1-year all-cause mortality between linked and unlinked TVT Registry records before and after the addition of MA data. RESULTS: Among 364 573 eligible TAVR procedures in the TVT Registry, 206 116 (57%) were linked to FFS claims and 111 662 (31%) to MA encounter data, for a combined linkage rate of 87%. FFS linkage rates declined steadily from 2016 to 2021. Inclusion of MA data increased the number of linked procedures by 54%, preserved linkage rates over time, and increased Hispanic (2.26-fold) and Black (1.91-fold) linkage. Characteristics of the linked TAVR cohort closely matched the full TVT Registry population, with most absolute differences under 1%. Unadjusted 1-year TAVR mortality was 12.6% and 11.8% among FFS and MA-linked patients (difference, 0.8% [95% CI, 0.6–1.1]). Findings were similar in the mitral transcatheter edge-to-edge repair cohort. CONCLUSIONS: Inclusion of MA claims in Medicare linkage to TAVR or mitral transcatheter edge-to-edge repair increased cohort size, enhanced subgroup representation, and maintained linkage rates despite declining FFS enrollment. These findings support incorporation of MA data into real-world surveillance, clinical evidence generation, and quality initiatives.

Real‐world evidence (RWE) derived from real‐world data (RWD) can complement randomized controlled trials (RCTs), yet the validity of RWD relative to RCT data remains insufficiently characterized. We obtained post hoc consent and linked individual participant data from the US‐based INVESTED trial (2016–2019) with Medicare fee‐for‐service claims (2012–2020) to validate demographic factors, baseline characteristics (using 183‐, 365‐, and 730‐day lookback periods) and outcomes, and to assess post‐trial events. Among 5260 trial participants, 126 were enrolled and eligible for linkage. Among 115 participants with demographic information available from Medicare enrollment files, agreement between RCT‐ and RWD‐based demographic factors was high: only one major age discrepancy, 100% agreement for sex, and an overall agreement of 0.89 for race. Participants with Medicare claims data ( n = 65) were older and more likely to be White compared with the overall RCT population. For the 365‐day lookback period, baseline comorbidities showed high sensitivity (median 0.80) and specificity (0.89), as did medication use (sensitivity 1.00, specificity 0.88). Lengthening the lookback period to 730 days increased sensitivity but decreased specificity, whereas shortening to 183 days decreased sensitivity but increased specificity. Clinical outcomes showed high specificity (0.88–0.94) but low sensitivity (0.18–0.50). Among those with Medicare coverage beyond the trial end date ( n = 45), 22% experienced cardiopulmonary, 18% cardiovascular, and 7% heart failure (HF) hospitalizations, highlighting the value of RWD for extending RCT evidence. Proactive planning of future RCT‐RWD linkage initiatives can improve the efficiency of linkage studies, leading to more actionable results.

BACKGROUND: Direct oral anticoagulants such as apixaban and rivaroxaban, have transformed the management of thromboembolic disorders. Randomized controlled trials were initiated in 2017 (COBRRA-VTE [Comparison of Bleeding Risk Between Rivaroxaban and Apixaban for the Treatment of Acute Venous Thromboembolism]) and 2021 (COBRRA-AF [Comparison of Bleeding Risk Between Rivaroxaban and Apixaban in Patients With Atrial Fibrillation]) to directly compare apixaban and rivaroxaban on the risk of bleeding in patients with venous thromboembolism (VTE) and atrial fibrillation (AF). We aimed to emulate the design and measurements of these ongoing trials and use US health care insurance claims data to predict their findings. METHODS: In cohort studies designed to emulate COBRRA-VTE (REGISTRATION: URL: https://www.clinicaltrials.gov ; Unique identifier: NCT05264168) and COBRRA-AF (NCT05256797) in 3 US health care insurance claims databases (Medicare, Optum Clinformatics, Merative MarketScan), the risk of bleeding with apixaban was evaluated relative to the risk with rivaroxaban. Pretreatment confounding factors were adjusted for via 1:1 propensity score matching. The primary outcome was a composite of major bleeding and clinically relevant non-major bleeding (defined using International Classification of Diseases, Ninth Revision and Tenth Revision diagnosis codes) recorded in the inpatient setting. Results across databases were pooled using fixed-effects meta-analyses. Protocols were preregistered on ClinicalTrials.gov before inferential analyses were conducted. RESULTS: In the COBRRA-VTE emulation (N=14 023 matched pairs, mean age=67 years, 55% women), we observed a 31% decreased risk of bleeding for apixaban compared with rivaroxaban (incidence rate per 100 person-years: apixaban, 12.8; rivaroxaban, 19.1; hazard ratio, 0.69 [95% CI, 0.59–0.79]). For the COBRRA-AF emulation (N=176 990 matched pairs, mean age=74 years, 50% women), we similarly observed a 31% decreased risk for apixaban compared with rivaroxaban (incidence rate per 100 person-years: apixaban, 6.4; rivaroxaban, 9.2; hazard ratio, 0.69 [95% CI, 0.66–0.71]). CONCLUSIONS: In database studies that closely emulated the design and measurements of the COBRRA-VTE and COBRRA-AF trials, a substantially lower risk of bleeding with apixaban was observed, consistent with the results of COBRRA-VTE (published after completion of our analyses), and indirect comparisons from other trials.

Objective: Using national claims databases, we sought to emulate the design of the ongoing SEPRA trial and predict its findings, comparing the effects of once weekly semaglutide to SoC medications on glycemic control (A1C <7%) in type-2 diabetes mellitus (T2D). Research design and methods: Using Optum Clinformatics (July 2017 – May 2022), we identified a 1:1 propensity score-matched (PSM) cohort of adults with T2D on metformin monotherapy, who had recorded A1C and initiated either injectable semaglutide or SoC medications (dipeptidyl peptidase-4 inhibitors, sodium-glucose cotransporter-2 inhibitors, SUs, or glucagon-like peptide-1 agonists) and met eligibility criteria adapted from the SEPRA trial. The primary outcome was the proportion of patients achieving A1C <7%. The study protocol was preregistered ( NCT05577728 , ClinicalTrials.gov) before any etiologic analyses. Risk ratios and corresponding 95% CIs were estimated. Results: We identified 1,144 PSM pairs of injectable semaglutide and SoC initiators with balance in pre-exposure covariates. Semaglutide initiators were 30% (risk ratio (95% CI), 1.30 (1.16 to 1.45)) more likely to achieve glycemic control (A1C <7%) than those initiating SoC. Additionally, semaglutide initiators had a 1.3% reduction in A1C, compared with a 1.1% reduction in the SoC group. These results were consistent with interim results of the SEPRA trial, which were released after the protocol for our database study was preregistered. Conclusion: This claims database study, designed to predict the results of the SEPRA trial, found results consistent with interim trial results. Our findings support the notion that well-designed non-randomized studies using fit-for-purpose data can effectively complement pragmatic randomized controlled trials.

Real‐world evidence involving healthcare database studies is well established for making causal inferences in post‐market drug safety studies and methods, data, and research infrastructure for evaluating effectiveness have advanced in recent years. The rapidly expanding field of etiologic research using insurance claims and electronic health records databases is being evaluated for supporting effectiveness claims. One such use case to support regulatory decision‐making on effectiveness is for expanding indications beyond existing effectiveness claims. Confidence in the validity of findings from cohort studies conducted using databases (hereafter “database study”) to support indication expansions could be increased through a structured benchmarking process of an initial database study against RCT evidence followed by calibration of a subsequent database study based on differences in results observed in the initial RCT‐database pair. This paper proposes a benchmark, expand, and calibration (BenchExCal) approach to trial emulation and describes the design and process for evaluating the performance of the approach through both simulation studies; five planned empirical examples are also described. The project will provide insights regarding how a first‐stage benchmarking emulation of a completed trial for an existing indication can be used to calibrate, increase confidence, and improve interpretation of the results for a second‐stage emulation of a hypothetical trial that could potentially provide evidence for an expanded indication. Although the examples have been selected to provide a variety of learnings, five use cases do not address all clinical and data scenarios that may be encountered when seeking a supplemental indication for a marketed drug.

Background: Unmeasured confounding is often raised as a source of potential bias during the design of non-randomized studies but quantifying such concerns is challenging. Methods: We developed a simulation-based approach to assess the potential impact of unmeasured confounding during the study design stage. The approach involved generation of hypothetical individual-level cohorts using realistic parameters including a binary treatment (prevalence 25%), a time-to-event outcome (incidence 5%), 13 measured covariates, a binary unmeasured confounder (u₁, 10%), and a binary measured ‘proxy’ variable (p₁) correlated with u₁. Strength of unmeasured confounding and correlations between u₁ and p₁ were varied in simulation scenarios. Treatment effects were estimated with, a) no adjustment, b) adjustment for measured confounders (Level 1), c) adjustment for measured confounders and their proxy (Level 2). We computed absolute standardized mean differences in u₁ and p₁ and relative bias with each level of adjustment. Results: Across all scenarios, Level 2 adjustment led to improvement in balance of u₁, but this improvement was highly dependent on the correlation between u₁ and p₁. Level 2 adjustments also had lower relative bias than Level 1 adjustments (in strong u₁ scenarios: relative bias of 9.2%, 12.2%, 13.5% at correlations 0.7, 0.5, and 0.3, respectively versus 16.4%, 15.8%, 15.0% for Level 1, respectively). Conclusion: An approach using simulated individual-level data was useful to explicitly convey the potential for bias due to unmeasured confounding while designing non-randomized studies and can be helpful in informing design choices.

Background: There is a dearth of drug utilization studies for coronavirus disease 2019 (COVID-19) treatments in 2021 and beyond after the introduction of vaccines and updated guidelines; such studies are needed to contextualize ongoing COVID-19 treatment effectiveness studies during these time periods. This study describes utilization patterns for corticosteroids, interleukin-6 (IL-6) inhibitors, Janus kinase inhibitors, and remdesivir among hospitalized adults with COVID-19, over the entire hospitalization, and within hospitalization periods categorized by respiratory support requirements. Methods: This descriptive cohort study included United States adults hospitalized with COVID-19 admitted from 1 January 2021 through 1 February 2022; data included HealthVerity claims and hospital chargemaster. The number and distribution of patients were reported for the first 3 drug regimen lines initiated. Results: The cohort included 51 066 patients; the most common initial drug regimens were corticosteroids (23.4%), corticosteroids plus remdesivir (25.1%), and remdesivir (4.4%). IL-6 inhibitors and Janus kinase inhibitors were included in later drug regimens and were more commonly administered with both corticosteroids and remdesivir than with corticosteroids alone. IL-6 inhibitors were more commonly administered than Janus kinase inhibitors when patients received high-flow oxygen or ventilation. Conclusions: These findings provide important context for comparative studies of COVID-19 treatments with study periods extending into 2021 and later. While prescribing generally aligned with National Institutes of Health COVID-19 treatment guidelines during this period, these findings suggest that prescribing preference, potential confounding by indication, and confounding by prior/concomitant use of other therapeutics should be considered in the design and interpretation of comparative studies.

Objectives To examine valsartan, losartan and irbesartan usage and switching patterns in the USA, UK, Canada and Denmark before and after July 2018, when the first Angiotensin-Receptor-Blocker (ARB) (valsartan) was recalled. Design Retrospective cohort study. Setting USA, Canadian administrative healthcare data, Danish National Prescription Registry and UK primary care electronic health records. Participants Patients aged 18 years and older between January 2014 and December 2020. Intervention Valsartan, losartan and irbesartan. Main outcome Monthly percentages of individual ARB episodes, new users and switches to another ARB, ACE inhibitors (ACEI) or calcium channel blockers containing products. Results We identified 10.8, 3.2, 1.8 and 1.2 million ARB users in the USA, UK, Canada and Denmark, respectively. Overall proportions of valsartan, losartan and irbesartan use were 18.4%, 67.9% and 5.2% in the USA; 3.1%, 48.3% and 10.2% in the UK, 16.3%, 11.4% and 18.3% in Canada, 1%, 93.5% and 0.6% in Denmark. In July 2018, we observed an immediate steep decline in the proportion of valsartan use in the USA and Canada. A similar trend was observed in Denmark; however, the decline was only minimal. We observed no change in trends of ARB use in the UK. Accompanying the valsartan decline was an increase in switching to other ARBs in the USA, Canada and Denmark. There was a small increase in switching to ACEI relative to the valsartan-to-other-ARBs switch. We also observed increased switching from other affected ARBs, losartan and irbesartan, to other ARBs throughout 2019, in the USA and Canada, although the usage trends in the USA remained unchanged. Conclusion The first recall notice for valsartan resulted in substantial decline in usage due to increased switching to other ARBs. Subsequent notices for losartan and irbesartan were also associated with increased switching around the time of the recall, however, overall usage trends remained unchanged.

Purpose: Algorithms for classification of inpatient COVID‐19 severity are necessary for confounding control in studies using real‐world data.Methods: Using Healthverity chargemaster and claims data, we selected patients hospitalized with COVID‐19 between April 2020 and February 2021, and classified them by severity at admission using an algorithm we developed based on respiratory support requirements (supplemental oxygen or non‐invasive ventilation, O2/NIV, invasive mechanical ventilation, IMV, or NEITHER). To evaluate the utility of the algorithm, patients were followed from admission until death, discharge, or a 28‐day maximum to report mortality risks and rates overall and by stratified by severity. Trends for heterogeneity in mortality risk and rate across severity classifications were evaluated using Cochran‐Armitage and Logrank trend tests, respectively.Results: Among 118 117 patients, the algorithm categorized patients in increasing severity as NEITHER (36.7%), O2/NIV (54.3%), and IMV (9.0%). Associated mortality risk (and 95% CI) was 11.8% (11.6–12.0%) overall and increased with severity [3.4% (3.2–3.5%), 11.5% (11.3–11.8%), 47.3% (46.3–48.2%); p < 0.001]. Mortality rate per 1000 person‐days (and 95% CI) was 15.1 (14.9–15.4) overall and increased with severity [5.7 (5.4–6.0), 14.5 (14.2–14.9), 32.7 (31.8–33.6); p < 0.001].Conclusion: As expected, we observed a positive association between the algorithm‐defined severity on admission and 28‐day mortality risk and rate. Although performance remains to be validated, this provides some assurance that this algorithm may be used for confounding control or stratification in treatment effect studies.