Distributed Research Network Research Articles

Utilization and patient characteristics for the trastuzumab reference and biosimilars, and other human epidermal growth factor receptor 2 inhibitors in the United States.

Trastuzumab is an antihuman epidermal growth factor receptor 2 monoclonal antibody used to treat breast and other cancers. Trastuzumab biosimilars were approved in the United States beginning in 2017. Utilization information on these biosimilars is limited. To examine utilization patterns and characteristics of patients treated with trastuzumab (biosimilars and reference) and other human epidermal growth factor receptor 2 products. We evaluated health care claims data from the Biologics and Biosimilars Collective Intelligence Consortium distributed research network, representing a large, geographically diverse US population of commercially insured individuals. We queried 4 distributed research network health plan partners to capture product usage data and patient information from October 1, 2016, to October 31, 2022. Patients were required to be continuously enrolled in their health plan for at least 365 days before their first observed trastuzumab utilization date in this study period. Data were aggregated across data partners. More than 16 million eligible health plan members representing more than 31 million person-years of data were evaluated. We identified 5,984 incident treatment episodes; 3,878 (64.8%) episodes were with the reference trastuzumab. The mean ages were consistent across trastuzumab products (60.2 to 65.1 years) and at least 80% of the episodes were among female patients. The mean comorbidity index score was 1.2 (SD = 1.9) among users of the reference vs the biosimilars (range 1.2-2.5). Other clinical characteristics (eg, diabetes, hypertension) were comparable across products. The proportion of total incident episodes of the reference trastuzumab decreased substantially over time (96% in 2016 vs 28% in 2021) as utilization of the biosimilars increased (eg, use of trastuzumab-anns increased from 2% [2019] to 36% [2021]). Similar utilization trends were seen among patients with and without metastatic breast cancer. Trastuzumab biosimilars utilization has grown since their introduction to the US market. Exploration of these biosimilars' comparative effectiveness and safety to their reference product is warranted.

Study protocol: a mixed-methods study of the implementation of doula care to address racial health equity in six state Medicaid programs

BackgroundRacial inequities in severe maternal morbidity (SMM) and mortality constitute a public health crisis in the United States. Doula care, defined as care from birth workers who provide culturally appropriate, non-clinical support during pregnancy and postpartum, has been proposed as an intervention to help disrupt obstetric racism as a driver of adverse pregnancy outcomes in Black and other birthing persons of colour. Many state Medicaid programs are implementing doula programs to address the continued increase in SMM and mortality. Medicaid programs are poised to play a major role in addressing the needs of these populations with the goal of closing the racial gaps in SMM and mortality. This study will investigate the most effective ways that Medicaid programs can implement doula care to improve racial health equity.MethodsWe describe the protocol for a mixed-methods study to understand how variation in implementation of doula programs in Medicaid may affect racial equity in pregnancy and postpartum health. Primary study outcomes include SMM, person-reported measures of respectful obstetric care, and receipt of evidence-based care for chronic conditions that are the primary causes of postpartum mortality (cardiovascular, mental health, and substance use conditions). Our research team includes doulas, university-based investigators, and Medicaid participants from six sites (Kentucky, Maryland, Michigan, Pennsylvania, South Carolina and Virginia) in the Medicaid Outcomes Distributed Research Network (MODRN). Study data will include policy analysis of doula program implementation, longitudinal data from a cohort of doulas, cross-sectional data from Medicaid beneficiaries, and Medicaid healthcare administrative data. Qualitative analysis will examine doula and beneficiary experiences with healthcare systems and Medicaid policies. Quantitative analyses (stratified by race groups) will use matching techniques to estimate the impact of using doula care on postpartum health outcomes, and will use time-series analyses to estimate the average treatment effect of doula programs on population postpartum health outcomes.DiscussionFindings will facilitate learning opportunities among Medicaid programs, doulas and Medicaid beneficiaries. Ultimately, we seek to understand the implementation and integration of doula care programs into Medicaid and how these processes may affect racial health equity.Study registration The study is registered with the Open Science Foundation (https://doi.org/10.17605/OSF.IO/NXZUF).

Time Series AI Model for Acute Kidney Injury Detection Based on a Multicenter Distributed Research Network: Development and Verification Study.

Acute kidney injury (AKI) is a marker of clinical deterioration and renal toxicity. While there are many studies offering prediction models for the early detection of AKI, those predicting AKI occurrence using distributed research network (DRN)-based time series data are rare. In this study, we aimed to detect the early occurrence of AKI by applying an interpretable long short-term memory (LSTM)-based model to hospital electronic health record (EHR)-based time series data in patients who took nephrotoxic drugs using a DRN. We conducted a multi-institutional retrospective cohort study of data from 6 hospitals using a DRN. For each institution, a patient-based data set was constructed using 5 drugs for AKI, and an interpretable multivariable LSTM (IMV-LSTM) model was used for training. This study used propensity score matching to mitigate differences in demographics and clinical characteristics. Additionally, the temporal attention values of the AKI prediction model's contribution variables were demonstrated for each institution and drug, with differences in highly important feature distributions between the case and control data confirmed using 1-way ANOVA. This study analyzed 8643 and 31,012 patients with and without AKI, respectively, across 6 hospitals. When analyzing the distribution of AKI onset, vancomycin showed an earlier onset (median 12, IQR 5-25 days), and acyclovir was the slowest compared to the other drugs (median 23, IQR 10-41 days). Our temporal deep learning model for AKI prediction performed well for most drugs. Acyclovir had the highest average area under the receiver operating characteristic curve score per drug (0.94), followed by acetaminophen (0.93), vancomycin (0.92), naproxen (0.90), and celecoxib (0.89). Based on the temporal attention values of the variables in the AKI prediction model, verified lymphocytes and calcvancomycin ium had the highest attention, whereas lymphocytes, albumin, and hemoglobin tended to decrease over time, and urine pH and prothrombin time tended to increase. Early surveillance of AKI outbreaks can be achieved by applying an IMV-LSTM based on time series data through an EHR-based DRN. This approach can help identify risk factors and enable early detection of adverse drug reactions when prescribing drugs that cause renal toxicity before AKI occurs.

Corrigendum to: Development and Verification of Time-Series Deep Learning for Drug-Induced Liver Injury Detection in Patients Taking Angiotensin II Receptor Blockers: A Multicenter Distributed Research Network Approach.

Corrigendum to: Development and Verification of Time-Series Deep Learning for Drug-Induced Liver Injury Detection in Patients Taking Angiotensin II Receptor Blockers: A Multicenter Distributed Research Network Approach.

One-shot distributed algorithms for addressing heterogeneity in competing risks data across clinical sites

ObjectiveTo characterize the interplay between multiple medical conditions across sites and account for the heterogeneity in patient population characteristics across sites within a distributed research network, we develop a one-shot algorithm that can efficiently utilize summary-level data from various institutions. By applying our proposed algorithm to a large pediatric cohort across four national Children’s hospitals, we replicated a recently published prospective cohort, the RISK study, and quantified the impact of the risk factors associated with the penetrating or stricturing behaviors of pediatric Crohn's disease (PCD). MethodsIn this study, we introduce the ODACoRH algorithm, a one-shot distributed algorithm designed for the competing risks model with heterogeneity. Our approach considers the variability in baseline hazard functions of multiple endpoints of interest across different sites. To accomplish this, we build a surrogate likelihood function by combining patient-level data from the local site with aggregated data from other external sites. We validated our method through extensive simulation studies and replication of the RISK study to investigate the impact of risk factors on the PCD for adolescents and children from four children's hospitals within the PEDSnet, A National Pediatric Learning Health System. To evaluate our ODACoRH algorithm, we compared results from the ODACoRH algorithms with those from meta-analysis as well as those derived from the pooled data. ResultsThe ODACoRH algorithm had the smallest relative bias to the gold standard method (−0.2%), outperforming the meta-analysis method (−11.4%). In the PCD association study, the estimated subdistribution hazard ratios obtained through the ODACoRH algorithms are identical on par with the results derived from pooled data, which demonstrates the high reliability of our federated learning algorithms. From a clinical standpoint, the identified risk factors for PCD align well with the RISK study published in the Lancet in 2017 and other published studies, supporting the validity of our findings. ConclusionWith the ODACoRH algorithm, we demonstrate the capability of effectively integrating data from multiple sites in a decentralized data setting while accounting for between-site heterogeneity. Importantly, our study reveals several crucial clinical risk factors for PCD that merit further investigations.

Development and Verification of Time-Series Deep Learning for Drug-Induced Liver Injury Detection in Patients Taking Angiotensin II Receptor Blockers: A Multicenter Distributed Research Network Approach.

The objective of this study was to develop and validate a multicenter-based, multi-model, time-series deep learning model for predicting drug-induced liver injury (DILI) in patients taking angiotensin receptor blockers (ARBs). The study leveraged a national-level multicenter approach, utilizing electronic health records (EHRs) from six hospitals in Korea. A retrospective cohort analysis was conducted using EHRs from six hospitals in Korea, comprising a total of 10,852 patients whose data were converted to the Common Data Model. The study assessed the incidence rate of DILI among patients taking ARBs and compared it to a control group. Temporal patterns of important variables were analyzed using an interpretable timeseries model. The overall incidence rate of DILI among patients taking ARBs was found to be 1.09%. The incidence rates varied for each specific ARB drug and institution, with valsartan having the highest rate (1.24%) and olmesartan having the lowest rate (0.83%). The DILI prediction models showed varying performance, measured by the average area under the receiver operating characteristic curve, with telmisartan (0.93), losartan (0.92), and irbesartan (0.90) exhibiting higher classification performance. The aggregated attention scores from the models highlighted the importance of variables such as hematocrit, albumin, prothrombin time, and lymphocytes in predicting DILI. Implementing a multicenter-based timeseries classification model provided evidence that could be valuable to clinicians regarding temporal patterns associated with DILI in ARB users. This information supports informed decisions regarding appropriate drug use and treatment strategies.

Pragmatic guidance for embedding pragmatic clinical trials in health plans: Large simple trials aren't so simple.

There are unique opportunities related to the design and conduct of pragmatic trials embedded in health insurance plans, which have longitudinal data on member/patient demographics, dates of coverage, and reimbursed medical care, including prescription drug dispensings, vaccine administrations, behavioral healthcare encounters, and some laboratory results. Such trials can be large and efficient, using these data to identify trial-eligible patients and to ascertain outcomes. We use our experience primarily with the National Institutes of Health Pragmatic Trials Collaboratory Distributed Research Network, which comprises health plans that participate in the US Food & Drug Administration's Sentinel System, to describe lessons learned from the conduct and planning of embedded pragmatic trials. Information is available for research on more than 75 million people with commercial or Medicare Advantage health plans. We describe three studies that have used or plan to use the Network, as well as a single health plan study, from which we glean our lessons learned. Studies that are conducted in health plans provide much-needed evidence to drive clinically meaningful changes in care. However, there are many unique aspects of these trials that must be considered in the planning, implementation, and analytic phases. The type of trial best suited for studies embedded in health plans will be those that require large sample sizes, simple interventions that could be disseminated through health plans, and where data available to the health plan can be leveraged. These trials hold potential for substantial long-term impact on our ability to generate evidence to improve care and population health.

STI Testing among Medicaid Enrollees Initiating PrEP for HIV Prevention in Six Southern States.

The purpose of this study was to measure sexually transmitted infection (STI) testing among Medicaid enrollees initiating preexposure prophylaxis (PrEP) to prevent human immunodeficiency virus. Secondary data are in the form of Medicaid enrollment and claims data in six states in the US South. Research partnerships in six states in the US South developed a distributed research network to accomplish study aims. Each state identified all first-time PrEP users in fiscal year 2017-2018 (combined N = 990) and measured the presence of STI testing for chlamydia, syphilis, and gonorrhea through 2019. Each state calculated the percentage of individuals with at least one STI test during 3-, 6-, and 12-month follow-up periods. The proportion of first-time PrEP users that received an STI test varied by state: 37% to 67% of all of the individuals in each state who initiated PrEP received a test within the first 6 months of PrEP treatment and 50% to 77% received a test within the first 12 months. Although the Centers for Disease Control and Prevention recommends STI testing at least every 6 months for PrEP users, our analysis of Medicaid data suggests that STI testing occurs less frequently than recommended in populations at elevated risk of syphilis, gonorrhea, and chlamydia.

Feasibility Study of Federated Learning on the Distributed Research Network of OMOP Common Data Model.

Since protecting patients' privacy is a major concern in clinical research, there has been a growing need for privacy-preserving data analysis platforms. For this purpose, a federated learning (FL) method based on the Observational Medical Outcomes Partnership (OMOP) common data model (CDM) was implemented, and its feasibility was demonstrated. We implemented an FL platform on FeederNet, which is a distributed clinical data analysis platform based on the OMOP CDM in Korea. We trained it through an artificial neural network (ANN) using data from patients who received steroid prescriptions or injections, with the aim of predicting the occurrence of side effects depending on the prescribed dose. The ANN was trained using the FL platform with the OMOP CDMs of Kyung Hee University Medical Center (KHMC) and Ajou University Hospital (AUH). The area under the receiver operating characteristic curves (AUROCs) for predicting bone fracture, osteonecrosis, and osteoporosis using only data from each hospital were 0.8426, 0.6920, and 0.7727 for KHMC and 0.7891, 0.7049, and 0.7544 for AUH, respectively. In contrast, when using FL, the corresponding AUROCs were 0.8260, 0.7001, and 0.7928 for KHMC and 0.7912, 0.8076, and 0.7441 for AUH, respectively. In particular, FL led to a 14% improvement in performance for osteonecrosis at AUH. FL can be performed with the OMOP CDM, and FL often shows better performance than using only a single institution's data. Therefore, research using OMOP CDM has been expanded from statistical analysis to machine learning so that researchers can conduct more diverse research.

Identification of cancer chemotherapy regimens and patient cohorts in administrative claims: challenges, opportunities, and a proposed algorithm

BackgroundReal-world evidence is a valuable source of information in healthcare. This study describes the challenges and successes during algorithm development to identify cancer cohorts and multi-agent chemotherapy regimens from claims data to perform a comparative effectiveness analysis of granulocyte colony stimulating factor (G-CSF) use.MethodsUsing the Biologics and Biosimilars Collective Intelligence Consortium’s Distributed Research Network, we iteratively developed and tested a de novo algorithm to accurately identify patients by cancer diagnosis, then extract chemotherapy and G-CSF administrations for a retrospective study of prophylactic G-CSF.ResultsAfter identifying patients with cancer and subsequent chemotherapy exposures, we observed only 12% of patients with cancer received chemotherapy, which is fewer than expected based on prior analyses. Therefore, we reversed the initial inclusion criteria to identify chemotherapy receipt, then prior cancer diagnosis, which increased the number of patients from 2,814 to 3,645, or 68% of patients receiving chemotherapy had diagnoses of interest. Additionally, we excluded patients with cancer diagnoses that differed from those of interest in the 183 days before the index date of G-CSF receipt, including early-stage cancers without G-CSF or chemotherapy exposure. By removing this criterion, we retained 77 patients who were previously excluded. Finally, we incorporated a 5-day window to identify all chemotherapy drugs administered (excluding oral prednisone and methotrexate, as these medications may be used for other non-malignant conditions) as patients may fill oral prescriptions days to weeks prior to infusion. This increased the number of patients with chemotherapy exposures of interest to 6,010. The final cohort of included patients, based on G-CSF exposure, increased from 420 from the initial algorithm to 886 using the final algorithm.ConclusionsMedications used for multiple indications, sensitivity and specificity of administrative codes, and relative timing of medication exposure must all be evaluated to identify patient cohorts receiving chemotherapy from claims data.

Abstract P4-03-33: A Population-based Analysis of Prophylactic G-CSF Biosimilar and Originator Administration over time among Patients Diagnosed with Breast Cancer

Abstract A Population-based Analysis of Prophylactic G-CSF Biosimilar and Originator Administration over time among Patients Diagnosed with Breast Cancer Pawloski PA1, McDermott CL2, Vazquez Benitez G1, DeFor T1, Mendelsohn A3, Marshall J3, Moyneur E4, Lockhart CM2; on behalf of the G-CSF Comparative Effectiveness Research Team. 1HealthPartners Institute, Bloomington MN USA 2Biologics and Biosimilars Collective Intelligence Consortium, Alexandria VA USA 3Harvard Pilgrim Healthcare Institute, Boston MA USA 4StatLog, Montreal QC Canada Objectives: To characterize G-CSF product use, including product switching, among patients diagnosed with breast cancer in the Biologics and Biosimilars Collective Intelligence Consortium’s (BBCIC) Distributed Research Network (DRN). Methods: A retrospective analysis of electronically extracted insurance claims from 2015-2019 at 4 Research Partner Sites was conducted. Patients aged >=20 years with a diagnosis of breast cancer who received chemotherapy associated with a risk of febrile neutropenia (FN) risk per National Comprehensive Cancer Network guidelines and any prophylactic granulocyte-colony stimulating factor (G-CSF), defined as before day 2 of the first cycle of chemotherapy were included. Results: A total of 11,788 patients were included; 89 (0.8%) were male sex per insurance records. The age distribution was 5,743 (49%) 50-64 years; 4,296 (36%) 20-49 years, and 1749 (15%) 65+ years. Chemotherapy regimens included cyclophosphamide/doxorubicin (n=7,377), carboplatin/docetaxel/trastuzumab/pertuzumab (n=1,862), cyclophosphamide/docetaxel (n=1,383), carboplatin/docetaxel/trastuzumab (n=430), cyclophosphamide/docetaxel/doxorubicin (n=147), and docetaxel/trastuzumab/pertuzumab (n=128). Overall, 218 patients (1.8%) developed FN during the first chemotherapy cycle. G-CSF utilization was pegfilgrastim 10,895 (92%), pegfilgrastim-cbqv 315 (3%), pegfilgrastim-jmdb 225 (2%), filgrastim 156 (1%), filgrastim-sndz 118 (1%), tbo-filgrastim 46 (< 1%), a combination of pegfilgrastim+filgrastim 26 (< 1%), and a combination of filgrastim+biosimilar 7 (< 1%) received. A total of 10,953 (93%) patients received high-risk chemotherapy and G-CSF utilization was 10,162 (93%) pegfilgrastim, 288 (3%) pegfilgrastim-cbqv, 200 (2%) pegfilgrastim-jmdb, 132 (1%) filgrastim, 101 (< 1%) filgrastim-sndz, 40 (< 1%) tbo-filgrastim, and 30 (< 1%) combination of products. Eight hundred fifteen (7%) patients received intermediate-risk chemotherapy and G-CSF utilization was, 716 (88%) pegfilgrastim, 27 (3%) pegfilgrastim-cbqv, 25 (3%) pegfilgrastim-jmdb, 23 (3%) filgrastim, 16 (2%) filgrastim-sndz, 5 (< 1%) tbo-filgrastim, and 3 (< 1%) combination. Twenty patients received low risk chemotherapy and of those 17 (85%) received pegfilgrastim. In 2019, the first full year of pegfilgrastim biosimilar availability, pegfilgrastim use was 76% pegfilgrastim, 13% pegfilgrastim-cbqv, and 9% pegfilgrastim-jmdb. Most patients who received a second cycle of chemotherapy received the same G-CSF product in the second cycle, specifically 67% received filgrastim, 71% tbo-filgrastim, 73% filgrastim-sndz, 96% pegfilgrastim, 84% pegfilgrastim-cbqv, and 81% pegfilgrastim-jmdb. Conclusions: The most common chemotherapy agents included cyclophosphamide, carboplatin, doxorubicin, docetaxel, or pertuzumab. Pegfilgrastim biosimilar uptake occurred following market availability. Within each G-CSF product, most patients received the same product during the second cycle of chemotherapy rather than switching products. Citation Format: Pamala A. Pawloski, Cara L McDermott, Gabriela Vazquez Benitez, Terese DeFor, Aaron B. Mendelsohn, James Marshall, Erick Moyneur, Catherine Lockhart. A Population-based Analysis of Prophylactic G-CSF Biosimilar and Originator Administration over time among Patients Diagnosed with Breast Cancer [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P4-03-33.

Imputing missing covariates in time-to-event analysis within distributed research networks: A simulation study.

In distributed research network (DRN) settings, multiple imputation cannot be directly implemented because pooling individual-level data are often not feasible. The performance of multiple imputation in combination with meta-analysis is not well understood within DRNs. To evaluate the performance of imputation for missing baseline covariate data in combination with meta-analysis for time-to-event analysis within DRNs, we compared two parametric algorithms including one approximated linear imputation model (Approx), and one nonlinear substantive model compatible imputation model (SMC), as well as two non-parametric machine learning algorithms including random forest (RF), and classification and regression trees (CART), through simulation studies motivated by a real-world data set. Under the setting with small effect sizes (i.e., log-Hazard ratios [logHR]) and homogeneous missingness mechanisms across sites, all imputation methods produced unbiased and more efficient estimates while the complete-case analysis could be biased and inefficient; and under heterogeneous missingness mechanisms, estimates with RF method could have higher efficiency. Estimates from the distributed imputation combined by meta-analysis were similar to those from the imputation using pooled data. When logHRs were large, the SMC imputation algorithm generally performed better than others. These findings suggest the validity and feasibility of imputation within DRNs in the presence of missing covariate data in time-to-event analysis under various settings. The performance of the four imputation algorithms varies with the effect sizes and level of missingness.

Association of polysubstance use disorder with treatment quality among Medicaid beneficiaries with opioid use disorder

IntroductionThe opioid crisis is transitioning to a polydrug crisis, and individuals with co-occurring substance use disorder (SUDs) often have unique clinical characteristics and contextual barriers that influence treatment needs, engagement in treatment, complexity of treatment planning, and treatment retention. MethodsUsing Medicaid data for 2017–2018 from four states participating in a distributed research network, this retrospective cohort study documents the prevalence of specific types of co-occurring SUD among Medicaid enrollees with an opioid use disorder (OUD) diagnosis, and assesses the extent to which different SUD presentations are associated with differential patterns of MOUD and psychosocial treatments. ResultsWe find that more than half of enrollees with OUD had a co-occurring SUD, and the most prevalent co-occurring SUD was for “other psychoactive substances”, indicated among about one-quarter of enrollees with OUD in each state. We also find some substantial gaps in MOUD treatment receipt and engagement for individuals with OUD and a co-occurring SUD, a group representing more than half of individuals with OUD. In most states, enrollees with OUD and alcohol, cannabis, or amphetamine use disorder are significantly less likely to receive MOUD compared to enrollees with OUD only. In contrast, enrollees with OUD and other psychoactive SUD were significantly more likely to receive MOUD treatment. Conditional on MOUD receipt, enrollees with co-occurring SUDs had 10 % to 50 % lower odds of having a 180-day period of continuous MOUD treatment, an important predictor of better patient outcomes. Associations with concurrent receipt of MOUD and behavioral counseling were mixed across states and varied depending on co-occurring SUD type. ConclusionsOverall, ongoing progress toward increasing access to and quality of evidence-based treatment for OUD requires further efforts to ensure that individuals with co-occurring SUDs are engaged and retained in effective treatment. As the opioid crisis evolves, continued changes in drug use patterns and populations experiencing harms may necessitate new policy approaches that more fully address the complex needs of a growing population of individuals with OUD and other types of SUD.

Outpatient follow-up and use of medications for opioid use disorder after residential treatment among Medicaid enrollees in 10 states

BackgroundFollow-up after residential treatment is considered best practice in supporting patients with opioid use disorder (OUD) in their recovery. Yet, little is known about rates of follow-up after discharge. The objective of this analysis was to measure rates of follow-up and use of medications for OUD (MOUD) after residential treatment among Medicaid enrollees in 10 states, and to understand the enrollee and episode characteristics that are associated with both outcomes. MethodsUsing a distributed research network to analyze Medicaid claims data, we estimated the likelihood of 4 outcomes occurring within 7 and 30 days post-discharge from residential treatment for OUD using multinomial logit regression: no follow-up or MOUD, follow-up visit only, MOUD only, or both follow-up and MOUD. We used meta-analysis techniques to pool state-specific estimates into global estimates. ResultsWe identified 90,639 episodes of residential treatment for OUD for 69,017 enrollees from 2018 to 2019. We found that 62.5% and 46.9% of episodes did not receive any follow-up or MOUD at 7 days and 30 days, respectively. In adjusted analyses, co-occurring mental health conditions, longer lengths of stay, prior receipt of MOUD or behavioral health counseling, and a recent ED visit for OUD were associated with a greater likelihood of receiving follow-up treatment including MOUD after discharge. ConclusionsForty-seven percent of residential treatment episodes for Medicaid enrollees are not followed by an outpatient visit or MOUD, and thus are not following best practices.

An exploratory comparative effectiveness analysis of febrile neutropenia incidence among patients with cancer receiving granulocyte colony stimulating factors.

408 Background: We conducted an exploratory comparative effectiveness analysis comparing various G-CSF (pegfilgrastim/filgrastim) products to each other in incidence of febrile neutropenia (FN) among patients with breast, lung, colon, pancreatic, ovarian cancers or non-Hodgkin’s lymphoma (NHL) in the Biologics and Biosimilars Collective Intelligence Consortium’s (BBCIC) Distributed Research Network. Methods: We included patients aged > = 20 years who, in 2015-2019, per insurance claims, received any pegfilgrastim or filgrastim products as febrile neutropenia (FN) prophylaxis during the first cycle of chemotherapy posing a high or intermediate FN risk per National Comprehensive Cancer Network guidelines. We compared the FN risk starting at day 5 following day 1 of chemotherapy receipt between products using Poisson regression model with standardized inverse probability weights and robust variance to estimate the Relative Risk (RR) and 95% Confidence intervals (CI). Results: A total of 15,941 patients received a pegfilgrastim product in cycle 1 of chemotherapy: 15,115 (95%) pegfilgrastim, 484 (3%) pegfilgrastim_cbqv, 342 (2%) pegfilgrastim_jmdb. 565 patients received a filgrastim product: 284 (50%) filgrastim, 201 (36%) filgrastim_sndz, 80 (14%) tbo-filgrastim. FN events by product were: 346 pegfilgrastim (2.3% of users), 11 pegfilgrastim_cbqv (2.3%), 8 pegfilgrastim_jmdb (2.3%), 13 filgrastim (4.6%), 5 filgrastim_sndz (2.5%), 2 tbo-filgrastim (2.5%). We found no difference in FN incidence when comparing pegfilgrastim_cbqv to pegfilgrastim (RR 0.83, 95% CI 0.41-1.69), pegfilgrastim_jmdb to pegfilgrastim (RR 1.03, 95% CI 0.56-1.92), and pegfilgrastim_jmdb to pegfilgrastim_cbqv (RR 1.11, 95% CI 0.45-2.74). Similarly, we found no difference in FN incidence when comparing filgrastim_sndz to filgrastim (RR 0.46, 95% CI 0.17-1.28), tbo-filgrastim to filgrastim (RR = 0.30, 95% CI 0.06-1.36), or tbo-filgrastim to filgrastim_sndz (RR 0.54, 95% CI 0.10-2.77). Adverse events were rare, with similar rates observed among all products. Conclusions: We observed no significant difference in FN incidence among patients when comparing various G-CSF products, including when biosimilars were compared to their reference counterparts.

Utilization, user characteristics, and adverse outcomes of bevacizumab products in oncology in a distributed research network.

400 Background: Bevacizumab (Avastin), an anti-angiogenic agent, was approved for various oncologic indications in 2004 in the US. Two biosimilars, bevacizumab-awwb (Mvasi) and bevacizumab-bvzr (Zirabev), were on the market 2017 and 2019, respectively. Real-world data on adverse events for bevacizumab biosimilars is limited. Methods: We conducted a retrospective cohort study using the Biologics and Biosimilars Collective Intelligence Consortium (BBCIC) distributed database to evaluate utilization patterns, patient characteristics, and prespecified adverse events of interest for the originator bevacizumab relative to its biosimilars in oncology among users 21 years of age and older between January 1, 2010, to December 31, 2020. Oncology indications included colon, lung, and gynecologic (cervical, uterine, and ovarian) cancers. Adverse events (AEs) of interest were arterial thromboembolism (ATE), congestive heart failure (CHF), gastrointestinal perforation, stroke and acute myocardial infarction (AMI), and venous thromboembolism (VTE). The study analyzed biosimilars collectively, which include bevacizumab-awwb and bevacizumab-bvzr. Results: Both bevacizumab originator and biosimilar users for colon cancer were mostly male (55%), had a mean age of 61 years, and a mean Charlson/Elixhauser combined comorbidity score from 7.0 to 8.0. Among patients with lung cancer users, 53% were female with an overall mean age of 64 years and comorbidity scores from 6.4 to 7.5. Patients with gynecologic cancers had mean age of 63 years, and comorbidity scores from 6.1 to 6.7. Overall, oncologic bevacizumab use increased by 19% over time. Of the 24,044 total episodes from 2010 to 2020, AEs included ATE (2,560 events, 11%), CHF (2,092 events, 9%), gastrointestinal perforation (2,858 events, 12%), stroke/AMI (399 events, 2%), and VTE (4,447 events, 19%). AE rates were comparable between the originator and biosimilars: 7% and 8% for ATE, 5% and 4% for CHF, 9% and 8% for gastrointestinal perforation, 0.6% and 1% for stroke and AMI, and 10% and 8% for VTE, respectively. From 2019 to 2020, biosimilar use increased from 6% to 49% of total bevacizumab utilization for patients with colon cancer, from 2% to 36% for patients with lung cancer, and from 2% to 38% for patients with gynecologic cancers. Conclusions: Bevacizumab utilization increased across all oncology indications evaluated during the study period. From 2019 to 2020, utilization for biosimilars relative to the originator increased. As there are limited data available on bevacizumab biosimilar use in real world settings, future research should be conducted to see if utilization for biosimilar products continues to rise as well as safety over time.

Medicaid can and should play an active role in advancing health equity.

Medicaid can and should play an active role in advancing health equity.

Inclusion, Diversity, Equity, and Accessibility (IDEA) in multi-regional data research: An approach to facilitating change in a distributed research environment.

ObjectiveA recently established data research network identified the need to address Inclusion, Diversity, Equity, and Accessibility and established a team committed to IDEA informed change. The Team identifies, develops, and implements IDEA strategies in distributed research network. We will share the process of establishing this Team and outline identified coordinated opportunities moving forward. ApproachAdministrative data and analyses are not neutral. Colonialism, racism, gender discrimination, ableism, and other forms of oppression have shaped the data that are available and the research processes that are used to analyze and evaluate data. It is imperative that health data research adopts and develops methods that embed IDEA. An informal survey of a health data research network identified four areas for action including capacity building, culture shifts, information sharing, and coordinated development of policies, practices, and tools. A multi-regional team was established to move forward with this work with an emphasis on operations and research practices. ResultsIn the fall of 2021 IDEA Team members were recruited from data centres across the network, including IDEA professionals, researchers, data collection and curation specialists, human resource professionals, and public/patient engagement specialists. The first IDEA Team meetings focused on team buildings and building shared purpose by shaping the Terms of Reference and creating principles for working together. Next steps will include an environmental audit to assess network capacity and a consensus oriented decision-making process to identify priorities. The Team includes over 20 members with a broad range of knowledge and expertise, making facilitation a key operational component. Establishing a baseline of knowledge and defining priorities will encompass the first year of the Teams work, navigating the distinct needs of a distributed network, the diverse needs of data centres, and the breadth of data that flows through the network. ConclusionThis multi-regional inter-disciplinary team is crucial to adopting, developing and embedding IDEA in a distributed data network and could serve as a model for other data research organizations. The Team will work towards capacity building, creating an internal culture shift and coordinating the development of new tools and resources.

Design, Implementation, and Evolution of the Medicaid Outcomes Distributed Research Network (MODRN).

In the US, Medicaid covers over 80 million Americans. Comparing access, quality, and costs across Medicaid programs can provide policymakers with much-needed information. As each Medicaid agency collects its member data, multiple barriers prevent sharing Medicaid data between states. To address this gap, the Medicaid Outcomes Distributed Research Network (MODRN) developed a research network of states to conduct rapid multi-state analyses without sharing individual-level data across states. To describe goals, design, implementation, and evolution of MODRN to inform other research networks. MODRN implemented a distributed research network using a common data model, with each state analyzing its own data; developed standardized measure specifications and statistical software code to conduct analyses; and disseminated findings to state and federal Medicaid policymakers. Based on feedback on Medicaid agency priorities, MODRN first sought to inform Medicaid policy to improve opioid use disorder treatment, particularly medication treatment. Since its 2017 inception, MODRN created 21 opioid use disorder quality measures in 13 states. MODRN modified its common data model over time to include additional elements. Initial barriers included harmonizing utilization data from Medicaid billing codes across states and adapting statistical methods to combine state-level results. The network demonstrated its utility and addressed barriers to conducting multi-state analyses of Medicaid administrative data. MODRN created a new, scalable, successful model for conducting policy research while complying with federal and state regulations to protect beneficiary health information. Platforms like MODRN may prove useful for emerging health challenges to facilitate evidence-based policymaking in Medicaid programs.

Real-world treatment intensities and pathways of macular edema following retinal vein occlusion in Korea from Common Data Model in ophthalmology

Despite many studies, optimal treatment sequences or intervals are still questionable in retinal vein occlusion (RVO) macular edema. The aim of this study was to examine the real-world treatment patterns of RVO macular edema. A retrospective analysis of the Observational Medical Outcomes Partnership Common Data Model, a distributed research network, of four large tertiary referral centers (n = 9,202,032) identified 3286 eligible. We visualized treatment pathways (prescription volume and treatment sequence) with sunburst and Sankey diagrams. We calculated the average number of intravitreal injections per patient in the first and second years to evaluate the treatment intensities. Bevacizumab was the most popular first-line drug (80.9%), followed by triamcinolone (15.1%) and dexamethasone (2.28%). Triamcinolone was the most popular drug (8.88%), followed by dexamethasone (6.08%) in patients who began treatment with anti-vascular endothelial growth factor (VEGF) agents. The average number of all intravitreal injections per person decreased in the second year compared with the first year. The average number of injections per person in the first year increased throughout the study. Bevacizumab was the most popular first-line drug and steroids were considered the most common as second-line drugs in patients first treated with anti-VEGF agents. Intensive treatment patterns may cause an increase in intravitreal injections.