Search

Real-world data (RWD) from electronic health records (EHRs) and administrative claims databases are used increasingly to generate real-world evidence (RWE). RWE is used to support clinical evidence packages for medicines that inform decision-makers. In this review of current issues in the use of RWD-derived external comparator groups to support regulatory filings, we assess a series of topics that generally apply across many disease indications. However, most of the examples and illustrations focus on the oncology clinical research setting. The topics include an overview of current uses of RWD in drug development, a discussion of regulatory filings using RWD-derived external comparators, a brief overview of guidance documents and white papers pertaining to external comparators, a summary of some limitations and methodological issues in the use of external comparator groups and finally, a look at the future of this area and recommendations.

Legislative and technological advancements over the past decade have given rise to the proliferation of healthcare data generated from routine clinical practice, often referred to as real-world data (RWD). These data have piqued the interest of healthcare stakeholders due to their potential utility in generating evidence to support clinical and regulatory decision making. In the oncology setting, studies leveraging RWD offer distinct advantages that are complementary to randomized controlled trials (RCTs). They also permit the conduct of investigations that may not be possible through prospective designs due to ethics or feasibility. Despite its promise, the use of RWD for the generation of clinical evidence remains controversial due to concerns of unmeasured confounding and other sources of bias that must be carefully addressed in the study design and analysis. To facilitate a better understanding of when RWD can provide reliable conclusions on drug effectiveness, we seek to conduct 10 RWD-based studies that emulate RCTs in oncology using a systematic, protocol-driven approach described herein. Results of this investigation will help inform clinical, scientific, and regulatory stakeholders on the applications of RWD in the context of product labeling expansion, drug safety, and comparative effectiveness in oncology.

Oncology drug development increasingly relies on single‐arm clinical trials. External controls (ECs) derived from electronic health record (EHR) databases may provide additional context. Patients from a US‐based oncology EHR database were aligned with patients from randomized controlled trials (RCTs) and trial‐specific eligibility criteria were applied to the EHR dataset. Overall survival (OS) in the EC‐derived control arm was compared with OS in the RCT experimental arm. The primary outcome was OS, defined as time from randomization or treatment initiation (EHR) to death. Cox regression models were used to obtain effect estimates using EHR data. EC‐derived hazard ratio estimates aligned closely with those from the corresponding RCT with one exception. Comparing log HRs among all RCT and EC results gave a Pearson correlation coefficient of 0.86. Properly selected control arms from contemporaneous EHR data could be used to put single‐arm trials of OS in advanced non‐small cell lung cancer into context.

Introduction The KRAS p.G12C mutation, prevalent in non–small-cell lung cancer (NSCLC), has only recently become a viable target. Here we present results of the largest retrospective observational study analyzing KRAS p.G12C in patients with advanced NSCLC. Materials and Methods Adults with advanced NSCLC (All Advanced NSCLC cohort) and subcohorts with different mutation profiles (KRAS p.G12C [G12C] and KRAS/EGFR/ALK wild type [Triple WT]) diagnosed January 2011 to March 2019 were selected from a US clinico-genomic database; treatment-related characteristics, molecular profiles, real-world overall (rwOS) and progression-free survival (rwPFS) were analyzed. Results Demographics were similar across cohorts, with more smokers and nonsquamous cell carcinoma histology in the G12C cohort. KRAS p.G12C was nearly mutually exclusive (≤1.2 %) with known actionable driver mutations, but non-driver co-mutations were common (STK11, 21.5 %; KEAP1, 7.0 %; TP53, 48.0 %). Among G12C patients, 20 % had no documentation of receiving systemic therapy. Across treated G12C patients, 67 % received immune checkpoint inhibitors; first-line usage increased from 0% (2014) to 81 % (2019). Among G12C patients, median (95 % CI) rwOS was 12.0 (9.6–15.3), 9.5 (8.1–13.1), and 6.7 (5.9–10.7) months after first, second, and third line of therapy, respectively; median (95 % CI) rwPFS was 5.0 (4.4–5.8), 4.0 (2.8–5.3), and 3.1 (2.4–4.3) months. Outcomes for the G12C subcohort were similar to those for all patients (All Advanced NSCLC cohort). Mutations in STK11/KEAP1 were associated with poorer survival across all cohorts. Conclusion The poor outcomes associated with KRAS p.G12C mutated advanced NSCLC indicate an unmet need for more effective novel treatments.

Purpose The aim of this study was to assess the impact of missing death data on survival analyses conducted in an oncology EHR‐derived database. Methods The study was conducted using the Flatiron Health oncology database and the National Death Index (NDI) as a gold standard. Three analytic frameworks were evaluated in advanced non‐small cell lung cancer (aNSCLC) patients: median overall survival [mOS]), relative risk estimates conducted within the EHR‐derived database, and “external control arm” analyses comparing an experimental group augmented with mortality data from the gold standard to a control group from the EHR‐derived database only. The hazard ratios (HRs) obtained within the EHR‐derived database (91% sensitivity) and the external control arm analyses, were compared with results when both groups were augmented with mortality data from the gold standard. The above analyses were repeated using simulated lower mortality sensitivities to understand the impact of more extreme levels of missing deaths. Results Bias in mOS ranged from modest (0.6–0.9 mos.) in the EHR‐derived cohort with (91% sensitivity) to substantial when lower sensitivities were generated through simulation (3.3–9.7 mos.). Overall, small differences were observed in the HRs for the EHR‐derived cohort across comparative analyses when compared with HRs obtained using the gold standard data source. When only one treatment arm was subject to estimation bias, the bias was slightly more pronounced, but increased substantially when lower sensitivities were simulated. Conclusions The impact on survival analysis is minimal with high mortality sensitivity with only modest impact associated within external control arm applications.

Objective To create a high-quality electronic health record (EHR)-derived mortality dataset for retrospective and prospective real-world evidence generation. Data sources/study setting Oncology EHR data, supplemented with external commercial and US Social Security Death Index data, benchmarked to the National Death Index (NDI). Study design We developed a recent, linkable, high-quality mortality variable amalgamated from multiple data sources to supplement EHR data, benchmarked against the highest completeness U.S. mortality data, the NDI. Data quality of the mortality variable version 2.0 is reported here. Principal findings For advanced non-small-cell lung cancer, sensitivity of mortality information improved from 66 percent in EHR structured data to 91 percent in the composite dataset, with high date agreement compared to the NDI. For advanced melanoma, metastatic colorectal cancer, and metastatic breast cancer, sensitivity of the final variable was 85 to 88 percent. Kaplan-Meier survival analyses showed that improving mortality data completeness minimized overestimation of survival relative to NDI-based estimates. Conclusions For EHR-derived data to yield reliable real-world evidence, it needs to be of known and sufficiently high quality. Considering the impact of mortality data completeness on survival endpoints, we highlight the importance of data quality assessment and advocate benchmarking to the NDI.

Background For many asthma medications, pregnancy safety data remains insufficient. Objective The omalizumab pregnancy registry, EXPECT, evaluates maternal, pregnancy, and infant outcomes after exposure to omalizumab, including incidence of congenital anomalies. Methods EXPECT is a prospective, observational study of pregnant women exposed to ≥1 dose of omalizumab within 8 weeks prior to conception or at any time during pregnancy. Primary outcome measures include rates of live births, elective terminations, stillbirths, and congenital anomalies. Data were collected at enrollment, each trimester, birth, and every 6 months up to 18 months post-delivery. Results As of November 2012, 188 of 191 pregnant women were exposed to omalizumab during their first trimester. Of 169 pregnancies with known outcomes (median exposure during pregnancy, 8.8 months), there were 156 live births of 160 infants (4 twin pairs), 1 fetal death/stillbirth, 11 spontaneous abortions, and 1 elective termination. Among 152 singleton infants, 22 (14.5%) were born prematurely. Of 147 singleton infants with weight data, 16 (10.9%) were small for gestational age. Among 125 singleton full-term infants, 4 (3.2%) had low birth weights. Overall, 20 infants had congenital anomalies confirmed, 7 (4.4%) of whom had 1 major defect. No pattern of anomalies was observed. Conclusions To date, proportions of major congenital anomalies, prematurity, low birth weight, and small size for gestational age observed in the EXPECT registry are not inconsistent with findings from other studies in this asthma population. Recognizing the small sample size available, no apparent increased birth prevalence of major anomalies or patterns of major anomalies has been observed.

Background EXCELS, a postmarketing observational cohort study, was a commitment to the US Food and Drug Administration to assess the long-term safety of omalizumab in an observational setting, focusing predominantly on malignancies. Objective The aim of this study was to examine a potential association between omalizumab and cardiovascular (CV)/cerebrovascular (CBV) events in EXCELS. Methods Patients (≥12 years of age) with moderate to severe allergic asthma and who were being treated with omalizumab (n = 5007) or not (n = 2829) at baseline were followed up for ≤5 years. Analyses included overall CV/CBV events, but focused on the subset of arterial thromboembolic events (ATEs), comprising CV death, myocardial infarction, ischemic stroke, transient ischemic attack, and unstable angina. A prespecified analysis of the end point of ATE was conducted to control for available potential confounders. A blinded independent expert panel adjudicated all events. Results At baseline, the 2 cohorts had similar demographic characteristics, but severe asthma was more common in the omalizumab versus the non-omalizumab group (50% vs 23%). Omalizumab-treated patients had a higher rate of CV/CBV serious adverse events (13.4 per 1,000 person years [PYs]) than did non–omalizumab-treated patients (8.1 per 1,000 PYs). The ATE rates per 1,000 PYs were 6.66 (101 patients/15,160 PYs) in the omalizumab cohort and 4.64 (46 patients/9,904 PYs) in the non-omalizumab cohort. After control for available confounding factors, the hazard ratio was 1.32 (95% CI, 0.91-1.91). Conclusion This observational study demonstrated a higher incidence rate of CV/CBV events in the omalizumab versus the non-omalizumab cohort. Differences in asthma severity between cohorts likely contributed to this imbalance, but some increase in risk cannot be excluded.

The 42-item version of the Inventory of Drinking Situations (IDS-42) assesses relative frequency of drinking behavior across eight categories of drinking situations and was originally developed as a method for identifying high-risk situations in alcoholic samples. This study was designed to examine the psychometric properties of the IDS-42 in a sample of university students in order to assess its suitability as an assessment tool in the non-clinical population. Three hundred and ninety-six students (111 M, 283 F, 2 with missing gender data) completed the IDS-42 and a well established measure of drinking motives, the Drinking Motives Questionnaire (DMQ). Confirmatory factor analysis of the IDS-42 established a hierarchical factor structure with eight lower-order factors and three higher-order factors of negatively-reinforcing situations, positively-reinforcing situations, and temptation situations. The eight lower-order IDS-42 factors demonstrated moderate to high internal consistency and excellent concurrent validity with conceptually-similar DMQ subscale scores. Non-parametric analyses revealed that male students reported a higher drinking frequency overall as compared to female students, particularly in IDS-42 situations involving Social Pressure to Drink, Pleasant Times with Others, Testing Personal Control, and Urges and Temptations. Across the entire sample of university student drinkers, a higher drinking frequency was reported in positively-reinforcing situations as compared to negatively-reinforcing situations and temptation situations, as predicted. Results suggest the IDS-42 possesses good psychometric properties and support its utility as a tool in identifying situation-specific antecedents to drinking among university students.

Background The Epidemiologic Study of Xolair (omalizumab): Evaluating Clinical Effectiveness and Long-term Safety in Patients with Moderate-to-Severe Asthma (EXCELS) assessed the long-term safety of omalizumab in a clinical practice setting as part of a phase IV US Food and Drug Administration postmarketing commitment. Objective We sought to evaluate long-term safety in omalizumab-treated and nonomalizumab-treated patients. Primary outcome measures focused on assessment of malignancies. Methods EXCELS was a prospective observational cohort study in patients (≥12 years of age) with moderate-to-severe allergic asthma. There were 2 cohorts: omalizumab (taking omalizumab at baseline) and nonomalizumab (no history of omalizumab treatment). Primary outcomes included all confirmed, incident, study-emergent primary malignancies (malignancies), including and excluding nonmelanoma skin cancer (NMSC); all malignancies were externally adjudicated. Results The omalizumab cohort had a higher proportion of patients with severe asthma compared with the nonomalizumab cohort (50.0% vs 23.0%). Median follow-up was approximately 5 years for both cohorts. Crude malignancy rates were similar in the omalizumab and nonomalizumab cohorts, with a rate ratio of 0.84 (95% CI, 0.62-1.13) for all malignancies and 0.98 (95% CI, 0.71-1.36) for all malignancies excluding NMSC. Kaplan-Meier plots of time to first confirmed study-emergent primary malignancy were similar for the 2 treatment cohorts. Cox proportional hazards modeling, adjusting for confounders and risk factors, resulted in a hazard ratio (omalizumab vs nonomalizumab) of 1.09 (95% CI, 0.87-1.38) for all malignancies and 1.15 (95% CI, 0.83-1.59) for all malignancies excluding NMSC. Conclusion Results from EXCELS suggest that omalizumab therapy is not associated with an increased risk of malignancy.