Statistical Analysis with Missing Data

Objectives: To assess the methodological quality of intention to treat analysis as reported in randomised controlled trials in four large medical journals. Design: Survey of all reports of randomised controlled trials published in 1997 in the BMJ, Lancet, JAMA , and New England Journal of Medicine . Main outcome measures: Methods of dealing with deviations from random allocation and missing data. Results: 119 (48%) of the reports mentioned intention to treat analysis. Of these, 12 excluded any patients who did not start the allocated intervention and three did not analyse all randomised subjects as allocated. Five reports explicitly stated that there were no deviations from random allocation. The remaining 99 reports seemed to analyse according to random allocation, but only 34 of these explicitly stated this. 89 (75%) trials had some missing data on the primary outcome variable. The methods used to deal with this were generally inadequate, potentially leading to a biased treatment effect. 29 (24%) trials had more than 10% of responses missing for the primary outcome, the methodsof handling the missing responses were similar in this subset. Conclusions: The intention to treat approach is often inadequately described and inadequately applied. Authors should explicitly describe the handling of deviations from randomised allocation and missing responses and discuss the potential effect of any missing response. Readers should critically assess the validity of reported intention to treat analyses. Key messages Intention to treat gives a pragmatic estimate of the benefit of a change in treatment policy rather than of potential benefit in patients who receive treatment exactly as planned Full application of intention to treat is possible only when complete outcome data are available for all randomised subjects About half of all published reports of randomised controlled trials stated that intention to treat was used, but handling of deviations from randomised allocation varied widely Many trials had some missing data on the primary outcome variable, and methods used to deal with this were generally inadequate, potentially leading to bias Intention to treat analyses are often inadequately described and inadequately applied

https://www.bmj.com/content/bmj/319/7211/670.full.pdf

What is meant by intention to treat analysis? Survey of published randomised controlled trials

CONTEXT: Initial treatment of major depressive disorder in adolescents may include cognitive-behavioral therapy (CBT) or a selective serotonin reuptake inhibitor (SSRI). However, little is known about their relative or combined effectiveness. OBJECTIVE: To evaluate the effectiveness of 4 treatments among adolescents with major depressive disorder. DESIGN, SETTING, AND PARTICIPANTS: Randomized controlled trial of a volunteer sample of 439 patients between the ages of 12 to 17 years with a primary Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, diagnosis of major depressive disorder. The trial was conducted at 13 US academic and community clinics between spring 2000 and summer 2003. INTERVENTIONS: Twelve weeks of (1) fluoxetine alone (10 to 40 mg/d), (2) CBT alone, (3) CBT with fluoxetine (10 to 40 mg/d), or (4) placebo (equivalent to 10 to 40 mg/d). Placebo and fluoxetine alone were administered double-blind; CBT alone and CBT with fluoxetine were administered unblinded. MAIN OUTCOME MEASURES: Children's Depression Rating Scale-Revised total score and, for responder analysis, a (dichotomized) Clinical Global Impressions improvement score. RESULTS: Compared with placebo, the combination of fluoxetine with CBT was statistically significant (P =.001) on the Children's Depression Rating Scale-Revised. Compared with fluoxetine alone (P =.02) and CBT alone (P =.01), treatment of fluoxetine with CBT was superior. Fluoxetine alone is a superior treatment to CBT alone (P =.01). Rates of response for fluoxetine with CBT were 71.0% (95% confidence interval [CI], 62%-80%); fluoxetine alone, 60.6% (95% CI, 51%-70%); CBT alone, 43.2% (95% CI, 34%-52%); and placebo, 34.8% (95% CI, 26%-44%). On the Clinical Global Impressions improvement responder analysis, the 2 fluoxetine-containing conditions were statistically superior to CBT and to placebo. Clinically significant suicidal thinking, which was present in 29% of the sample at baseline, improved significantly in all 4 treatment groups. Fluoxetine with CBT showed the greatest reduction (P =.02). Seven (1.6%) of 439 patients attempted suicide; there were no completed suicides. CONCLUSION: The combination of fluoxetine with CBT offered the most favorable tradeoff between benefit and risk for adolescents with major depressive disorder. Language: en

Fluoxetine, cognitive-behavioral therapy, and their combination for adolescents with depression: Treatment for Adolescents With Depression Study (TADS) randomized controlled trial.

Randomized controlled trials often suffer from two major complications, ie, noncompliance and missing outcomes One potential solution to this problem is a statistical concept called intention-to-treat (ITT) analysis ITT analysis includes every subject who is randomized according to randomized treatment assignment It ignores noncompliance, protocol deviations, withdrawal, and anything that happens after randomization ITT analysis maintains prognostic balance generated from the original random treatment allocation In ITT analysis, estimate of treatment effect is generally conservative A better application of the ITT approach is possible if complete outcome data are available for all randomized subjects Per-protocol population is defined as a subset of the ITT population who completed the study without any major protocol violations

Intention-to-treat concept: A review.

ContextActivation of the coagulation system and depletion of endogenous anticoagulants
are frequently found in patients with severe sepsis and septic shock. Diffuse
microthrombus formation may induce organ dysfunction and lead to excess mortality
in septic shock. Antithrombin III may provide protection from multiorgan failure
and improve survival in severely ill patients.ObjectiveTo determine if high-dose antithrombin III (administered within 6 hours
of onset) would provide a survival advantage in patients with severe sepsis
and septic shock.Design and SettingDouble-blind, placebo-controlled, multicenter phase 3 clinical trial
in patients with severe sepsis (the KyberSept Trial) was conducted from March
1997 through January 2000.PatientsA total of 2314 adult patients were randomized into 2 equal groups of
1157 to receive either intravenous antithrombin III (30 000 IU in total
over 4 days) or a placebo (1% human albumin).Main Outcome MeasureAll-cause mortality 28 days after initiation of study medication.ResultsOverall mortality at 28 days in the antithrombin III treatment group
was 38.9% vs 38.7% in the placebo group (P = .94).
Secondary end points, including mortality at 56 and 90 days and survival time
in the intensive care unit, did not differ between the antithrombin III and
placebo groups. In the subgroup of patients who did not receive concomitant
heparin during the 4-day treatment phase (n = 698), the 28-day mortality was
nonsignificantly lower in the antithrombin III group (37.8%) than in the placebo
group (43.6%) (P = .08). This trend became significant
after 90 days (n = 686; 44.9% for antithrombin III group vs 52.5% for placebo
group; P = .03). In patients receiving antithrombin
III and concomitant heparin, a significantly increased bleeding incidence
was observed (23.8% for antithrombin III group vs 13.5% for placebo group; P<.001).ConclusionsHigh-dose antithrombin III therapy had no effect on 28-day all-cause
mortality in adult patients with severe sepsis and septic shock when administered
within 6 hours after the onset. High-dose antithrombin III was associated
with an increased risk of hemorrhage when administered with heparin. There
was some evidence to suggest a treatment benefit of antithrombin III in the
subgroup of patients not receiving concomitant heparin.

https://jamanetwork.com/journals/jama/articlepdf/194284/JCE10000.pdf

High-Dose Antithrombin III in Severe Sepsis: A Randomized Controlled Trial

Intention to treat in clinical trials

Importance Treatment with the anti-ERBB2 humanized monoclonal antibody trastuzumab and chemotherapy significantly improves outcome in patients with ERBB2 (HER2)–positive metastatic breast cancer; a clinically effective biosimilar may help increase access to this therapy. Objective To compare the overall response rate and assess the safety of a proposed trastuzumab biosimilar plus a taxane or trastuzumab plus a taxane in patients without prior treatment for ERBB2-positive metastatic breast cancer. Design, Setting, and Participants Multicenter, double-blind, randomized, parallel-group, phase 3 equivalence study in patients with metastatic breast cancer. From December 2012 to August 2015, 500 patients were randomized 1:1 to receive a proposed biosimilar or trastuzumab plus a taxane. Chemotherapy was administered for at least 24 weeks followed by antibody alone until unacceptable toxic effects or disease progression occurred. Interventions Proposed biosimilar (n = 230) or trastuzumab (n = 228) with a taxane. Main Outcomes and Measures The primary outcome was week 24 overall response rate (ORR) defined as complete or partial response. Equivalence boundaries were 0.81 to 1.24 with a 90% CI for ORR ratio (proposed biosimilar/trastuzumab) and −15% to 15% with a 95% CI for ORR difference. Secondary outcome measures included time to tumor progression, progression-free and overall survival at week 48, and adverse events. Results Among 500 women randomized, the intention-to-treat population included 458 women (mean [SD] age, 53.6 [11.11] years) and the safety population included 493 women. The ORR was 69.6% (95% CI, 63.62%-75.51%) for the proposed biosimilar vs 64.0% (95% CI, 57.81%-70.26%) for trastuzumab. The ORR ratio (1.09; 90% CI, 0.974-1.211) and ORR difference (5.53; 95% CI, −3.08 to 14.04) were within the equivalence boundaries. At week 48, there was no statistically significant difference with the proposed biosimilar vs trastuzumab for time to tumor progression (41.3% vs 43.0%; −1.7%; 95% CI, −11.1% to 6.9%), progression-free survival (44.3% vs 44.7%; −0.4%; 95% CI, −9.4% to 8.7%), or overall survival (89.1% vs 85.1%; 4.0%; 95% CI, −2.1% to 10.3%). In the proposed biosimilar and trastuzumab groups, 239 (98.6%) and 233 (94.7%) had at least 1 adverse event, the most common including neutropenia (57.5% vs 53.3%), peripheral neuropathy (23.1% vs 24.8%), and diarrhea (20.6% vs 20.7%). Conclusions and Relevance Among women with ERBB2-positive metastatic breast cancer receiving taxanes, the use of a proposed trastuzumab biosimilar compared with trastuzumab resulted in an equivalent overall response rate at 24 weeks. Further study is needed to assess safety and long-term clinical outcome. Trial Registration clinicaltrials.gov Identifier:NCT02472964; EudraCT Identifier: 2011-001965-42

https://jamanetwork.com/journals/jama/articlepdf/2590051/joi160140.pdf

Effect of a Proposed Trastuzumab Biosimilar Compared With Trastuzumab on Overall Response Rate in Patients With ERBB2 (HER2)-Positive Metastatic Breast Cancer: A Randomized Clinical Trial.

A frequent problem in designing a clinical trial is the need to rely on guesses of the magnitude of parameters, such as means and variances, needed to perform the calculations of required sample size. A sample size adjustment design incorporates a protocol-specified interim analysis conducted not to test a hypothesis but rather to adjust the sample size of the trial. This paper extends previous work on internal pilot studies for normally distributed outcome variables to the case of a binomial outcome variable.Through a specific example from a cancer clinical trial, the effect of the sample size adjustment process on significance level, power, and average sample size is addressed. Design issues such as timing of the interim analysis, potential for unblinding, and the effect of disparity of initial guesses of parameter values and estimates derived from interim analysis are also discussed. The sample size adjustment process is found to have negligible effect on significance level and to deliver reasonable po...

The Use of Interim Analysis for Sample Size Adjustment

Praise for the first edition:

"Given the author’s years of experience as a statistician and as a founder of the first DMC in pharmaceutical industry trials, I highly recommend this book—not only for experts because of its cogent and organized presentation, but more importantly for young investigators who are seeking information about the logistical and philosophical aspects of a DMC." 
-S. T. Ounpraseuth, The American Statistician 

　
In the first edition of this well-regarded book, the author provided a groundbreaking and definitive guide to best practices in pharmaceutical industry data monitoring committees (DMCs). Maintaining all the material from the first edition and adding substantial new material, Data and Safety Monitoring Committees in Clinical Trials, Second Edition is ideal for training professionals to serve on their first DMC as well as for experienced clinical and biostatistical DMC members, sponsor and regulatory agency staff. 

The second edition guides the reader through newly emerging DMC responsibilities brought about by regulations emphasizing risk vs benefit and the emergence of risk-based monitoring. It also provides the reader with many new statistical methods, clinical trial designs and clinical terminology that have emerged since the first edition. The references have been updated and the very popular end-of-chapter Q&A section has been supplemented with many new experiences since the first edition. 

　
New to the Second Edition:



Presents statistical methods, tables, listings and graphs appropriate for safety review, efficacy analysis and risk vs benefit analysis, SPERT and PRISMA initiatives.


Newly added interim analysis for efficacy and futility section.


DMC responsibilities in SUSARs (Serious Unexpected Serious Adverse Reactions), basket trials, umbrella trials, dynamic treatment strategies /SMART trials, pragmatic trials, biosimilar trials, companion diagnostics, etc.


DMC responsibilities for data quality and fraud detection (Fraud Recovery Plan)


Use of patient reported outcomes of safety


Use of meta analysis and data outside the trial


New ideas for training and compensation of DMC members


　
Jay Herson is Senior Associate, Biostatistics, Johns Hopkins Bloomberg School of Public Health where he teaches courses on clinical trials and drug development based on his many years experience in clinical trials in academia and the pharmaceutical industry.

Data and Safety Monitoring Committees in Clinical Trials

This is a commissioned report by a writing committee formed by the Society for Clinical Trials. The committee was formed with the objectives of 1) reviewing data monitoring guidelines for confirmatory (phase III) trials published by the National Institutes of Health, US Food and Drug Administration, Veterans Administration, and the International Conference on Harmonisation and 2) proposing corresponding guidelines for exploratory clinical trials (ie, most phase I and phase II trials and others not requiring a fully independent data monitoring committee). These trials typically involve fewer subjects and are of shorter duration than phase III trials. Nevertheless there are safety concerns, especially because these are often the first human trials for a new intervention. Recommendations are given for appropriate elements of a data monitoring plan, decision criteria for institution of a data monitoring committee (DMC), and critical elements for a DMC to consider in exploratory trials. Review and approval of ...

Jay Herson

Papers

Intention to treat in clinical trials

Effect of a Proposed Trastuzumab Biosimilar Compared With Trastuzumab on Overall Response Rate in Patients With ERBB2 (HER2)-Positive Metastatic Breast Cancer: A Randomized Clinical Trial.

The Use of Interim Analysis for Sample Size Adjustment

Data and Safety Monitoring Committees in Clinical Trials

Guidelines for data and safety monitoring for clinical trials not requiring traditional data monitoring committees