Optimizing collection of adverse event data in cancer clinical trials supporting supplemental indications

Streamlining the conduct of cancer clinical trials: new standard data collection practices for National Cancer Institute late-phase clinical studies

The increase in the complexity of cancer clinical trials over the past several decades has led to a dramatic growth in trial cost and operational burden. The extent and frequency of data collection, particularly in late-phase trials that enroll many participants, have been major contributors to this problem. The Clinical Trials and Translational Research Advisory Committee of the National Cancer Institute (NCI) recently assessed the effect of these stressors on the NCI National Clinical Trials Network (NCTN) and recommended that data collection in late-phase NCTN trials be limited to data elements essential to address the primary and secondary objectives of the trial. The purpose of this commentary is to describe the rationale for this recommendation, progress toward implementation, and the development of new streamlined standard practices for data collection for late-phase NCTN trials effective January 1, 2025.

Healthcare utilization disparities among children with high-risk neuroblastoma treated on Children's Oncology Group clinical trials

INTRODUCTION

Disparities in relapse and survival from high-risk neuroblastoma (HRNBL) persist among children from historically marginalized groups even in highly standardized clinical trial settings. Research in other cancers has identified differential treatment toxicity as one potential underlying mechanism. Whether racial and ethnic disparities in treatment-associated toxicity exist in HRNBL is poorly understood.

METHODS

This is a retrospective study utilizing a previously assembled merged cohort of children with HRNBL on Children's Oncology Group (COG) post-consolidation immunotherapy trials ANBL0032 and ANBL0931 at Pediatric Health Information System (PHIS) centers from 2005 to 2014. Race and ethnicity were categorized to reflect historically marginalized populations as Hispanic, non-Hispanic Black (NHB), non-Hispanic other (NHO), and non-Hispanic White (NHW). Associations between race-ethnicity and intensive care unit (ICU)-level care utilization as a proxy for treatment-associated toxicity were examined with log binomial regression and summarized as risk ratio (RR) and corresponding 95% confidence interval (CI).

RESULTS

The analytic cohort included 370 children. Overall, 88 (23.8%) patients required ICU-level care for a median of 3.0 days (interquartile range [IQR]: 1.0-6.5 days). Hispanic children had nearly three times the risk of ICU-level care (RR 3.1, 95% CI: 2.1-4.5; fully adjusted RR [aRR] 2.5, 95% CI: 1.6-3.7) compared to NHW children and the highest percentage of children requiring cardiovascular-driven ICU-level care.

CONCLUSION

Children of Hispanic ethnicity with HRNBL receiving clinical trial-delivered therapy were more likely to experience ICU-level care compared to NHW children. These data suggest that further investigation of treatment-related toxicity as a modifiable mechanism underlying outcome disparities is warranted.

Safety-Related Drug Label Changes Following Large Post-Marketing Cardiometabolic Trials: A Review of European Public Assessment Reports

Selective safety data collection may simplify late-stage clinical trials and improve their feasibility. However, the impact on increasing overall drug safety knowledge is unknown. The aim of this study is to evaluate how much safety information is added to the drug label based on large trials after initial authorization. Changes made to the "undesirable effects" section of the drug label of cardiometabolic agents approved between 2000 and 2020 based on the results of large (> 1,000 patient) clinical trials submitted to the European Medicines Agency (EMA) were evaluated. The study focused on glucose lowering, antithrombotic, and lipid-modifying agents. The primary outcome was the number of changes in adverse drug reactions in the drug label. The EMA reviewed 55 large trials concerning 25 cardiometabolic agents after the initial marketing authorization, which included 402,444 patients. Ultimately, 38 trials (69%) resulted in a safety section update, whereas 17 trials (31%) did not. Changes in listed adverse drug reactions were made following 19 trials (35%) for 12 agents: 77 adverse drug reactions were added, 11 were deleted, and the frequencies of 43 were changed. Most changes in adverse drug reactions arose from trials with antithrombotic agents (88%) and trials performed in a new population (92%). Large trials for cardiometabolic agents reported after authorization add limited new safety information on adverse drug reactions, especially when performed in the population studied prior to approval. This suggests that selective safety data collection does not reduce learnings from late stage cardiometabolic trials in populations comprehensively studied before.

Estimation of a Suitable Number of Patients for Selective Safety Data Collection (ICH E19 Draft Guideline): When is the Safety Profile of a Drug Well Characterized?

PURPOSE

We propose methods to estimate a suitable number of patients for implementing selective safety data collection (SSDC) in clinical investigations based on a confidence interval of the incidence rate or risk difference using Monte Carlo simulation.

METHODS

The incidence rates and risk differences of adverse events (AEs) were based on the safety outcome measures. A suitable number of patients for implementing SSDC was estimated based on the probability that the half-width of the two-sided 95% confidence interval of incidence rate or risk difference was equal to or less than a pre-specified cut-off value (0.5-3.0%). Monte Carlo simulation was used to estimate the suitable number of patients at probabilities of 70%, 80%, and 90%. The applicability of our proposed method for estimating a suitable number of patients for SSDC implementation was confirmed based on the incidence rates or risk differences from actual clinical trial data for panitumumab.

RESULTS

We demonstrated the performance of our proposed method in estimating a suitable number of patients to implement SSDC in several situations. Furthermore, according to the safety datasets of three phase III clinical trials, the number of suitable patients for implementing SSDC using incidence rates or risk differences of common AEs with panitumumab could confirm the applicability of our proposed method.

CONCLUSION

A suitable number of patients estimated based on our proposed method may be one of the foundations for implementing SSDC, as additional data accrual may not impact the precision of the estimates of the frequency of common AEs.

Risk of Persistent Opioid Use following Major Surgery in Matched Samples of Patients with and without Cancer

BACKGROUND

The opioid crisis has reached epidemic proportions, yet risk of persistent opioid use following curative intent surgery for cancer and factors influencing this risk are not well understood.

METHODS

We used electronic health record data from 3,901 adult patients who received a prescription for an opioid analgesic related to hysterectomy or large bowel surgery from January 1, 2013, through June 30, 2018. Patients with and without a cancer diagnosis were matched on the basis of demographic, clinical, and procedural variables and compared for persistent opioid use.

RESULTS

Cancer diagnosis was associated with greater risk for persistent opioid use after hysterectomy [18.9% vs. 9.6%; adjusted OR (aOR), 2.26; 95% confidence interval (CI), 1.38-3.69; P = 0.001], but not after large bowel surgery (28.3% vs. 24.1%; aOR 1.25; 95% CI, 0.97-1.59; P = 0.09). In the cancer hysterectomy cohort, persistent opioid use was associated with cancer stage (increased rates among those with stage III cancer compared with stage I) and use of neoadjuvant or adjuvant chemotherapy; however, these factors were not associated with persistent opioid use in the large bowel cohort.

CONCLUSIONS

Patients with cancer may have an increased risk of persistent opioid use following hysterectomy.

IMPACT

Risks and benefits of opioid analgesia for surgical pain among patients with cancer undergoing hysterectomy should be carefully considered.

Streamlining Safety Data Collection in Hospital-Acquired Bacterial Pneumonia and Ventilator-Associated Bacterial Pneumonia Trials: Recommendations of the Clinical Trials Transformation Initiative Antibacterial Drug Development Project Team

BACKGROUND

Resistant bacteria are one of the leading causes of hospital-acquired/ventilator-associated bacterial pneumonia (HABP/VABP). HABP/VABP trials are complex and difficult to conduct due to the large number of medical procedures, adverse events, and concomitant medications involved. Differences in the legislative frameworks between different regions of the world may also lead to excessive data collection. The Clinical Trials Transformation Initiative (CTTI) seeks to advance antibacterial drug development (ABDD) by streamlining clinical trials to improve efficiency and feasibility while maintaining ethical rigor, patient safety, information value, and scientific validity.

METHODS

In 2013, CTTI engaged a multidisciplinary group of experts to discuss challenges impeding the conduct of HABP/VABP trials. Separate workstreams identified challenges associated with current data collection processes. Experts defined "data collection" as the act of capturing and reporting certain data on the case report form as opposed to recording of data as part of routine clinical care. The ABDD Project Team developed strategies for streamlining safety data collection in HABP/VABP trials using a Quality by Design approach.

RESULTS

Current safety data collection processes in HABP/VABP trials often include extraneous information. More targeted strategies for safety data collection in HABP/VABP trials will rely on optimal protocol design and prespecification of which safety data are essential to satisfy regulatory reporting requirements.

CONCLUSIONS

A consensus and a cultural change in clinical trial design and conduct, which involve recognition of the need for more efficient data collection, are urgently needed to advance ABDD and to improve HABP/VABP trials in particular.

Accuracy of Adverse Event Ascertainment in Clinical Trials for Pediatric Acute Myeloid Leukemia

PURPOSE

Reporting of adverse events (AEs) in clinical trials is critical to understanding treatment safety, but data on AE accuracy are limited. This study sought to determine the accuracy of AE reporting for pediatric acute myeloid leukemia clinical trials and to test whether an external electronic data source can improve reporting.

METHODS

Reported AEs were evaluated on two trials, Children's Oncology Group AAML03P1 and AAML0531 arm B, with identical chemotherapy regimens but with different toxicity reporting requirements. Chart review for 12 AEs for patients enrolled in AAML0531 at 14 hospitals was the gold standard. The sensitivity and positive predictive values (PPV) of the AAML0531 AE report and AEs detected by review of Pediatric Health Information System (PHIS) billing and microbiology data were compared with chart data.

RESULTS

Select AE rates from AAML03P1 and AAML0531 arm B differed significantly and correlated with the targeted toxicities of each trial. Chart abstraction was performed on 204 patients (758 courses) on AAML0531. AE report sensitivity was < 50% for eight AEs, but PPV was > 75% for six AEs. AE reports for viridans group streptococcal bacteremia, a targeted toxicity on AAML0531, had a sensitivity of 78.3% and PPV of 98.1%. PHIS billing data had higher sensitivity (> 50% for nine AEs), but lower PPV (< 75% for 10 AEs). Viridans group streptococcal detection using PHIS microbiology data had high sensitivity (92.3%) and PPV (97.3%).

CONCLUSION

The current system of AE reporting for cooperative oncology group clinical trials in pediatric acute myeloid leukemia underestimates AE rates. The high sensitivity and PPV of PHIS microbiology data suggest that using external data sources may improve the accuracy of AE reporting.

Data collection in cancer clinical trials: Too much of a good thing?

Background

Substantial staff time and costs are incurred in the collection of data for cancer clinical trials. Anecdotal experience suggests that much of these data are never used in the analysis or reporting of a trial.

Purpose

To quantify data items collected in cancer clinical trials and calculate what percentage is used in subsequent published manuscripts.

Methods

Cancer clinical trials completed by the Ontario Clinical Oncology Group (OCOG) between 2003 and 2012 and the corresponding primary outcome publication were identified. The number of data items collected on each trial’s case report form (CRF) was counted and sorted into 18 categories including eligibility, baseline characteristics, medical history, toxicity, and recurrence. The data items were then counted within the corresponding published manuscripts to determine percent of data used overall and within each section.

Results

In all, 8 trials, with 9 corresponding publications, were evaluated. The CRF analysis revealed that the total collected items per subject ranged from 186 to 1035 per trial with a median of 599. Across all the publications, a median of 96 data items (18%) were reported in each manuscript, ranging from 11% to 27% per trial. In 8 of the 18 categories, 4% or less of collected data items were used.

Limitations

The number of trials reviewed is small and were conducted from a single clinical trial coordinating centre. The main outcome of the number of data items used in the published manuscript is a surrogate for trial information considered valuable by investigators. Some data may be deemed important by investigators but not included in manuscripts.

Conclusions

In this analysis of publications from 8 clinical trials, a small amount of data collected was ultimately used in peer-reviewed journal manuscripts. A large amount of data collected in cancer trials appears to go unused and could be omitted from CRFs, thus simplifying data collection and improving trial efficiency.

Induction mortality and resource utilization in children treated for acute myeloid leukemia at free-standing pediatric hospitals in the United States

BACKGROUND

Clinical trials in pediatric acute myeloid leukemia (AML) determine induction regimen standards. However, these studies lack the data necessary to evaluate mortality trends over time and differences in resource utilization between induction regimens. Moreover, these trials likely underreport the clinical toxicities experienced by patients.

METHODS

The Pediatric Health Information System database was used to identify children treated for presumed de novo AML between 1999 and 2010. Induction mortality, risk factors for induction mortality, and resource utilization by induction regimen were estimated using standard frequentist statistics, logistic regression, and Poisson regression, respectively.

RESULTS

A total of 1686 patients were identified with an overall induction case fatality rate of 5.4% that decreased from 9.8% in 2003 to 2.1% in 2009 (P = .0023). The case fatality rate was 9.0% in the intensively timed DCTER (dexamethasone, cytarabine, thioguanine, etoposide, and rubidomycin [daunomycin]/idarubicin) induction and 3.8% for ADE (cytarabine, daunomycin, and etoposide) induction (adjusted odds ratio = 2.2, 95% confidence interval = 1.1-4.5). Patients treated with intensively timed DCTER regimens had significantly greater antibiotic, red cell/platelet transfusion, analgesic, vasopressor, renal replacement therapy, and radiographic resource utilization than patients treated with ADE regimens. Resource utilization was substantially higher than reported in published pediatric AML clinical trials.

CONCLUSIONS

Induction mortality for children with AML decreased significantly as ADE use increased. In addition to higher associated mortality, intensively timed DCTER regimens had a correspondingly higher use of health care resources. Using resource utilization data as a proxy for adverse events, adverse event rates reported on clinical trials substantially underestimated the clinical toxicities of all pediatric AML induction regimens.

The impact of non-drug-related toxicities on the estimation of the maximum tolerated dose in phase I trials

The rate of observed dose-limiting toxicities (DLT) determines the maximum tolerated dose (MTD) in phase I trials. There are cases in which non-drug-related toxicities or other-cause toxicities (OCT) are flagged as DLTs, or vice versa, due to attribution errors. We aim to assess the impact of such errors on the final estimate of MTD. We compared the impact of attribution errors using 2 trial designs-the "3+3" dose-escalation scheme and the continual reassessment method (CRM). Two attribution errors are considered: when a DLT is classified as an OCT (type A error) and when an OCT is misclassified as a DLT (type B error). The impact of these errors on accuracy, patient safety, sample size, and study duration was evaluated by varying the probability of occurrence of each error through simulated trials. Under no errors, CRM is on average 35% more accurate than 3+3 in finding the true MTD. This improved accuracy is maintained in the presence of errors. At a 15% type B error rate, CRM recommends a dose within 2 levels of the true MTD 68% of the time, compared with 17% of the time using the 3+3 method. A DLT must be attributed as an OCT 30% of the time to increase the accuracy of 3+3; otherwise the method recommends a wrong dose approximately 75% of the time. CRM is more robust to toxicity attribution errors compared with the 3+3 as it uses information from all treated patients, leading to a more accurate MTD estimation at the frequency of attribution errors anticipated in phase I clinical trials.

Merging of the National Cancer Institute-funded cooperative oncology group data with an administrative data source to develop a more effective platform for clinical trial analysis and comparative effectiveness research: a report from the Children's Oncology Group

PURPOSE

The National Cancer Institute-funded cooperative oncology group trials have improved overall survival for children with cancer from 10% to 85% and have set standards of care for adults with malignancies. Despite these successes, cooperative oncology groups currently face substantial challenges. We are working to develop methods to improve the efficiency and effectiveness of these trials. Specifically, we merged data from the Children's Oncology Group (COG) and the Pediatric Health Information Systems (PHIS) to improve toxicity monitoring, to estimate treatment-associated resource utilization and costs, and to address important clinical epidemiology questions.

METHODS

COG and PHIS data on patients enrolled on a phase III COG trial for de novo acute myeloid leukemia at 43 PHIS hospitals were merged using a probabilistic algorithm. Resource utilization summary statistics were then tabulated for the first chemotherapy course based on PHIS data.

RESULTS

Of 416 patients enrolled on the phase III COG trial at PHIS centers, 392 (94%) were successfully matched. Of these, 378 (96%) had inpatient PHIS data available beginning at the date of study enrollment. For these, daily blood product usage and anti-infective exposures were tabulated and standardized costs were described.

CONCLUSIONS

These data demonstrate that patients enrolled in a cooperative group oncology trial can be successfully identified in an administrative data set and that supportive care resource utilization can be described. Further work is required to optimize the merging algorithm, map resource utilization metrics to the National Cancer Institute Common Toxicity Criteria for monitoring toxicity, to perform comparative effectiveness studies, and to estimate the costs associated with protocol therapy.

Funding Oncology Clinical Trials: Are Cooperative Group Trials Sustainable?

Purpose

Many oncology clinical trials departments (CTDs) are in serious fiscal deficit and their sustainability is in jeopardy. This study investigates whether the payment models used to fund industry versus cooperative group trials contribute to the fiscal deficit of a CTD.

Methods

We examined the lifetime costs of all cooperative group and industry trials activated in the CTD of a cancer center between 2007 and 2011. A trial's lifetime is defined as being from the date the first patient was accrued until the last patient's actual or projected final follow-up visit. For each trial, we calculated the lifetime monthly net income, which was defined as monthly revenue minus monthly costs. Data sources included study protocols, trial budgets, and accrual data.

Results

Of the 97 trials analyzed, 64 (66%) were cooperative group trials. The pattern of lifetime net income for cooperative group trials has a positive peak during patient accrual followed by a negative trough during follow-up. In contrast, the pattern for industry trials resembled an “l” shape. The patterns reflect the differing payment models: upfront lump-sum payments (cooperative group) versus milestone payments (industry).

Conclusion

The negative trough in the lifetime net income of a cooperative group trial occurs because follow-up costs are typically not funded or are underfunded. CTDs accrue more patients in new trials to offset that deficit. The CTD uses revenue from accrual to existing trials to cross-subsidize past trials in follow-up. As the number of patients on follow-up increases, the fiscal deficit grows larger each year, perpetuating the cycle.