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.

Larynx preservation clinical trial design: summary of key recommendations of a consensus panel

An international consensus panel was convened to develop guidelines for the conduct of phase III clinical trials of larynx preservation in patients with locally advanced laryngeal and hypopharyngeal cancer. According to their recommendations, future trial populations should include patients with T2 or T3 laryngeal or hypopharyngeal squamous cell carcinoma not considered for partial laryngectomy and should exclude those with laryngeal dysfunction or aged >70 years. Baseline and post-treatment functional assessments should include speech and swallowing evaluations. Furthermore, voice should be routinely assessed with a simple, validated instrument. Regarding endpoints, the primary endpoint should capture survival and function. As a result, the panel created a new endpoint of laryngoesophageal dysfunction (LED)-free survival, which includes the events of death, local relapse, total or partial laryngectomy, tracheotomy at ≥2 years, or feeding tube at ≥2 years. Recommended secondary endpoints are freedom from LED, overall survival, progression-free survival, locoregional control, time to tracheotomy, time to laryngectomy, time to discontinuation of feeding tube, and quality of life/patient-reported outcomes. Future exploratory correlative biomarker studies should include epidermal growth factor receptor, excision repair cross-complementation group 1 gene, E-cadherin and β-catenin, epiregulin and amphiregulin, and TP53 mutation. Revised trial designs in several key areas are needed to advance the study of larynx preservation. With consistent methodologies, clinical trials can more effectively evaluate and quantify the therapeutic benefit of novel treatment options for patients with locally advanced laryngeal and hypopharyngeal cancer.

Optimism bias leads to inconclusive results-an empirical study

OBJECTIVE

Optimism bias refers to unwarranted belief in the efficacy of new therapies. We assessed the impact of optimism bias on a proportion of trials that did not answer their research question successfully and explored whether poor accrual or optimism bias is responsible for inconclusive results.

STUDY DESIGN

Systematic review.

SETTING

Retrospective analysis of a consecutive-series phase III randomized controlled trials (RCTs) performed under the aegis of National Cancer Institute Cooperative groups.

RESULTS

Three hundred fifty-nine trials (374 comparisons) enrolling 150,232 patients were analyzed. Seventy percent (262 of 374) of the trials generated conclusive results according to the statistical criteria. Investigators made definitive statements related to the treatment preference in 73% (273 of 374) of studies. Investigators' judgments and statistical inferences were concordant in 75% (279 of 374) of trials. Investigators consistently overestimated their expected treatment effects but to a significantly larger extent for inconclusive trials. The median ratio of expected and observed hazard ratio or odds ratio was 1.34 (range: 0.19-15.40) in conclusive trials compared with 1.86 (range: 1.09-12.00) in inconclusive studies (P<0.0001). Only 17% of the trials had treatment effects that matched original researchers' expectations.

CONCLUSION

Formal statistical inference is sufficient to answer the research question in 75% of RCTs. The answers to the other 25% depend mostly on subjective judgments, which at times are in conflict with statistical inference. Optimism bias significantly contributes to inconclusive results.

The FDA outlook of events reporting after ticagrelor or clopidogrel in the PLATO Trial: impact of sponsor censoring dates, drug discontinuation, and withdrawal of consent

BACKGROUND

Censoring by the study sponsor of clinical endpoint events in indication-seeking randomized trials represents a controversial approach since the reported data may be biased in favor of experimental agents due to the obvious conflict of interest. The frequency of drug discontinuation and rates of consent withdrawal may also impact the trial outcomes.

PURPOSE

To assess patterns of event reporting dependent on sponsor censoring dates, drug discontinuation and consent withdrawal in the PLATO trial.

METHODS

Analysis of the Food and Drug Administration Complete Response Review for ticagrelor.

RESULTS

Excluding adjudicated deaths, the distribution for clopidogrel appears more uniform while that for ticagrelor was skewed to the right, suggesting more events were reported after the sponsor censoring end date. PLATO investigators reported 16 unmatched primary endpoint events for ticagrelor immediately following the sponsor censoring date. Twenty-six out of 30 unreported events following early drug discontinuation occurred amongst patients using ticagrelor. More ticagrelor patients withdrew consent (Δ = 47), or were 'not willing' to complete the study (Δ = 87) when compared to clopidogrel.

CONCLUSIONS

Site-reported primary endpoints were unequally distributed for clopidogrel and ticagrelor in the PLATO trial. This pattern suggests the importance of questioning the impact of sponsor-mediated censoring on event reporting by investigators in indication-seeking trials. In PLATO, this pattern seems to have favored the experimental drug and may require further assessment.

Phase III clinical trial development: a process of chutes and ladders

PURPOSE

The Institute of Medicine report on cooperative groups and the National Cancer Institute (NCI) report from the Operational Efficiency Working Group both recommend changes to the processes for opening a clinical trial. This article provides evidence for the need for such changes by completing the first comprehensive review of all the time and steps required to open a phase III oncology clinical trial and discusses the effect of time to protocol activation on subject accrual.

METHODS

The Dilts and Sandler method was used at four cancer centers, two cooperative groups, and the NCI Cancer Therapy Evaluation Program. Accrual data were also collected.

RESULTS

Opening a phase III cooperative group therapeutic trial requires 769 steps, 36 approvals, and a median of approximately 2.5 years from formal concept review to study opening. Time to activation at one group ranged from 435 to 1,604 days, and time to open at one cancer center ranged from 21 to 836 days. At centers, group trials are significantly more likely to have zero accruals (38.8%) than nongroup trials (20.6%; P < 0.0001). Of the closed NCI Cancer Therapy Evaluation Program-approved phase III clinical trials from 2000 to 2007, 39.1% resulted in <21 accruals.

CONCLUSIONS

The length, variability, and low accrual results demonstrate the need for the NCI clinical trials system to be reengineered. Improvements will be of only limited effectiveness if done in isolation; there is a need to return to the collaborative spirit with all parties creating an efficient and effective system. Recommendations put forth by the Institute of Medicine and Operational Efficiency Working Group reports, if implemented, will aid this renewal.

Survey of investigators' opinions on the acceptability of interactions with patients participating in clinical trials

INTRODUCTION

There is growing concern about the ability of clinical trials to reliably detect differences between active drugs and placebo. To date, little attention has focused on how interactions between clinical trial investigators and patients may influence study outcomes. We sought to explore what types of interactions with patients investigators considered to be appropriate during placebo-controlled pharmacotherapy studies of major depressive disorder.

METHODS

Questionnaires describing 26 specific types of clinician-patient interactions were administered to principal investigators (PIs) attending an investigators meeting for a phase 3 clinical trial of an antidepressant medication. Principal investigators were asked to rate the acceptability of each intervention. They were also asked to report the mean time they spent with patients at a midstudy visit. Principal investigators were grouped according to previous trial experiences (participation in <20 or > or =20 prior trials).

RESULTS

Principal investigators generally agreed that physical health recommendations and nonspecific interactions with study patients were acceptable. Relating the investigator's personal experiences and siding with the patient on interpersonal conflicts were consistently rated as unacceptable. Less-experienced PIs were significantly more likely to view as acceptable cognitive, behavioral, and emotionally supportive interventions compared with more-experienced PIs. Forty-two percent of PIs reported spending at least 20 minutes with patients at midstudy visits.

CONCLUSIONS

There is significant variability between PIs in what are considered to be appropriate interactions with patients participating in clinical trials. Greater standardization of these interactions is required to reduce placebo response rates and to strengthen the ethical conduct of clinical trials.

The uncontrolled clinical trial: scientific, ethical, and practical reasons for being

According to principles of clinical trial design, the demonstration of efficacy of a new treatment is based on comparing the response in the treated group with that of a control group receiving placebo or another active treatment. The need for a control group is also recommended by the major international institutions that govern the ethics and the practice of clinical research. Despite these principles and recommendations, inspection of a purposive sample of ongoing clinical trials listed in the NIH registry ( http://ClinicalTrials.gov ) reveals that as many as one-third of trials are uncontrolled. Since these trials were approved through a formal evaluation by ethics committees, the lack of adequate control was not perceived as a major deficiency in the study design. Most uncontrolled trials belong to the oncology/hematology area. If two extreme disease conditions for nature and progression are analyzed, such as acute myeloid leukemia (AML) and chronic heart failure (CHF), the difference in the prevalence of uncontrolled trials is very striking. The number of uncontrolled trials is only 13% in CHF, whereas it reaches 66% in the AML group. I believe that the underlying disease condition plays a primary role in orienting the design of the study: oncology and hematology may be fields in which uncontrolled studies are common, whereas in other fields, e.g., cardiology, this phenomenon can be reduced. Within the limitations of the selection process of the examined trials, the current analysis indicates that the clinical trial reality does not strictly follow experimental design theory and official recommendations.

Methodological and ethical quality in phase III--breast cancer trials

OBJECTIVE

The gold standard of Evidence Based Medicine remains the randomised controlled trial (RCT), which is the only tool that allows an approach to the "therapeutic truth". To reach credible conclusions, these trials need to be perfect in methodological and ethical quality. The purpose of this study is to evaluate methodological quality (MQ), ethical quality (EQ) and compliance with ethical requirements in phase III randomized clinical trials of breast cancer treatment.

STUDY DESIGN

MQ was evaluated by the Jadad-scale and EQ by the Berdeu-score for all the randomised controlled clinical trials (RCT) (n = 137), published between January 2001 and December 2005 in 11 international journals.

RESULTS

Mean MQ was 9.88 +/- 1.43. MQ was insufficient (Jadad score pound 9) for 49 RCT (35,8%). Mean EQ was 0.45 +/- 0.12. Mean EQ for RCT with insufficient MQ (n = 49) was 0.43 +/- 0.12; Mean EQ for RCT with good MQ (Jadad score > 9)(n = 88) was 0.46 +/- 0.11. There was significant improvement in MQ depending on the year in which the study was started (p = 0.002). EQ was independent of the year of study's start (p = 0.134).There was no relationship between MQ or EQ and the number of patients included in the study (p = 0.53 and p = 0.1). There was a tendency towards correlation between MQ and EQ (p = 0.052), but the correlation between these two variables could not be considered as significant (r = 0.67). Informed consent from patients (ICP) was not obtained in 5.8% (n = 8) of the RCTs and the approval of a research ethics committee (REC) was not mentioned in 26.3% (n = 36) of the RCTs.

CONCLUSIONS

Good MQ and reporting of ethical requirements (EQ) reflects the respect shown to the patients during the whole research process. There are still deficiencies in EQ and MQ. Quality improvement requires education and appropriation by the scientific community, in particular, medical staff, of methodological and ethical basic rules concerning trials involving human beings.

Evaluation and recommendations on good clinical laboratory practice guidelines for phase I-III clinical trials

Marcella Sarzotti-Kelsoe and colleagues harmonize various approaches to Good Clinical Laboratory Practice for clinical trials into a single set of recommendations.

[Principle and practice of clinical phase III studies]

Randomized phase III studies compare new treatments with standard therapy according to defined guidelines and legal rules. Large international randomized phase III studies are some of the most complex and expensive medical experiments. The results of such trials will decide about the future of new drugs and are the basis of evidence-based medicine and the development of clinical guidelines. This contribution discusses randomization, endpoints, inclusion and exclusion criteria of phase III trials as well as further challenges when developing and conducting phase III studies in oncology.

[Interim analyses]

The methodological principles for the planning of interim analyses in a phase III clinical trial are presented in this article. The case for superiority, non-inferiority and futility, and the roles of Data Monitoring Committees are summarized. Several examples are presented to illustrate the methodology and to help investigators by better understanding and planning interim analyses in a phase III clinical trial.

A new approach to comparing efficacy results from clinical trials of topical acne vulgaris treatments

Clinicians must evaluate a voluminous amount of information from clinical trials when choosing among the growing number of topical acne treatments. This article describes a simple way to evaluate and broadly compare the efficacy results of well-controlled phase III clinical trials of topical acne treatments by taking into account the placebo effect (or active control effect). Key efficacy results are drawn from the package inserts of 7 primary topical acne treatments. To account for placebo effect, the mean percent reduction with placebo was subtracted from that of the active treatment for each type of lesion and also calculated as an average of these differences for all lesions. Based upon the principles of research, this method accounts for within study variances and offers a quick assessment of product efficacy. Clinicians should be cautioned that this method only allows for broad comparisons and does not establish definitive differences between treatments.