On non-inferiority margin and statistical tests in active control trials

Pragmatic randomised controlled trial of guided self-help versus individual cognitive behavioural therapy with a trauma focus for post-traumatic stress disorder (RAPID)

Background

Guided self-help has been shown to be effective for other mental conditions and, if effective for post-traumatic stress disorder, would offer a time-efficient and accessible treatment option, with the potential to reduce waiting times and costs.

Objective

To determine if trauma-focused guided self-help is non-inferior to individual, face-to-face cognitive-behavioural therapy with a trauma focus for mild to moderate post-traumatic stress disorder to a single traumatic event.

Design

Multicentre pragmatic randomised controlled non-inferiority trial with economic evaluation to determine cost-effectiveness and nested process evaluation to assess fidelity and adherence, dose and factors that influence outcome (including context, acceptability, facilitators and barriers, measured qualitatively). Participants were randomised in a 1 : 1 ratio. The primary analysis was intention to treat using multilevel analysis of covariance.

Setting

Primary and secondary mental health settings across the United Kingdom's National Health Service.

Participants

One hundred and ninety-six adults with a primary diagnosis of mild to moderate post-traumatic stress disorder were randomised with 82% retention at 16 weeks and 71% at 52 weeks. Nineteen participants and ten therapists were interviewed for the process evaluation.

Interventions

Up to 12 face-to-face, manualised, individual cognitive-behavioural therapy with a trauma focus sessions, each lasting 60-90 minutes, or to guided self-help using Spring, an eight-step online guided self-help programme based on cognitive-behavioural therapy with a trauma focus, with up to five face-to-face meetings of up to 3 hours in total and four brief telephone calls or e-mail contacts between sessions.

Main outcome measures

Primary outcome: the Clinician-Administered PTSD Scale for Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, at 16 weeks post-randomisation. Secondary outcomes: included severity of post-traumatic stress disorder symptoms at 52 weeks, and functioning, symptoms of depression, symptoms of anxiety, alcohol use and perceived social support at both 16 and 52 weeks post-randomisation. Those assessing outcomes were blinded to group assignment.

Results

Non-inferiority was demonstrated at the primary end point of 16 weeks on the Clinician-Administered PTSD Scale for Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition [mean difference 1.01 (one-sided 95% CI -∞ to 3.90, non-inferiority p = 0.012)]. Clinician-Administered PTSD Scale for Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, score improvements of over 60% in both groups were maintained at 52 weeks but the non-inferiority results were inconclusive in favour of cognitive-behavioural therapy with a trauma focus at this timepoint [mean difference 3.20 (one-sided 95% confidence interval -∞ to 6.00, non-inferiority p = 0.15)]. Guided self-help using Spring was not shown to be more cost-effective than face-to-face cognitive-behavioural therapy with a trauma focus although there was no significant difference in accruing quality-adjusted life-years, incremental quality-adjusted life-years -0.04 (95% confidence interval -0.10 to 0.01) and guided self-help using Spring was significantly cheaper to deliver [£277 (95% confidence interval £253 to £301) vs. £729 (95% CI £671 to £788)]. Guided self-help using Spring appeared to be acceptable and well tolerated by participants. No important adverse events or side effects were identified.

Limitations

The results are not generalisable to people with post-traumatic stress disorder to more than one traumatic event.

Conclusions

Guided self-help using Spring for mild to moderate post-traumatic stress disorder to a single traumatic event appears to be non-inferior to individual face-to-face cognitive-behavioural therapy with a trauma focus and the results suggest it should be considered a first-line treatment for people with this condition.

Future work

Work is now needed to determine how best to effectively disseminate and implement guided self-help using Spring at scale.

Trial registration

This trial is registered as ISRCTN13697710.

Funding

This award was funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme (NIHR award ref: 14/192/97) and is published in full in Health Technology Assessment; Vol. 27, No. 26. See the NIHR Funding and Awards website for further award information.

Multiple assessments of non-inferiority trials with ordinal endpoints

Non-inferiority (NI) trials are implemented when there is a practical demand to search for alternatives to standard therapies, such as to reduce side effects. An experimental treatment is considered non-inferior to the standard treatment when it exhibits clinically non-significant loss of efficacy. Ordinal categorical responses are frequently observed in clinical trials. It has been reported that responses measured using an ordinal scale produce more informative analysis than when responses collapse into binary outcomes. We study the NI trials using ordinal endpoints. We propose a latent variable model for ordinal categorical responses. Based on the proposed latent variable model, the mean efficacy of the different treatments is denoted by the corresponding mean parameter of the underlying continuous distributions. A two-step procedure is proposed for model identification and parameter estimation. A non-inferiority analysis can then be conducted based on the latent variable model and the corresponding estimation procedure. We also develop a method and an algorithm to produce an optimal sample size configuration based on the proposed testing procedure. Two clinical examples are provided for demonstrative purposes.

How do we know a treatment is good enough? A survey of non-inferiority trials

BACKGROUND

Non-inferiority and equivalence trials aim to determine whether a new treatment is good enough (non-inferior) or as good as (equivalent to) another treatment. To inform the decision about non-inferiority or equivalence, a margin is used. We aimed to identify the current methods used to determine non-inferiority or equivalence margins, as well as the main challenges and suggestions from trialists.

METHODS

We developed an online questionnaire that included both closed and open-ended questions about methods to elicit non-inferiority or equivalence margins, underlying principles, and challenges and suggestions for improvement. We recruited trialists with experience of determining a margin by contacting corresponding authors for non-inferiority or equivalence trials. We used descriptive statistics and content analysis to identify categories in qualitative data.

RESULTS

We had forty-one responses, all from non-inferiority trials. More than half of the trials were non-pharmacological (n = 21, 51%), and the most common primary outcome was clinical (n = 29, 71%). The two most used methods to determine the margin were as follows: a review of the evidence base (n = 27, 66%) and opinion seeking methods (n = 24, 59%). From those using reviews, the majority used systematic reviews or reviews of multiple RCTs to determine the margin (n = 17, 63%). From those using opinion seeking methods, the majority involved clinicians with or without other professionals (n = 19, 79%). Respondents reported that patients' opinions on the margin were sought in four trials (16%). Median confidence in overall quality of the margin was 5 out of 7 (maximum confidence); however, around a quarter of the respondents were "completely unconfident" that the margin reflected patient's views. We identified "stakeholder involvement" as the most common category to determine respondent's confidence in the quality of the margins and whether it reflected stakeholder's views. The most common suggestion to improve the definition of margins was "development of methods to involve stakeholders," and the most common challenge identified was "communication of margins."

CONCLUSIONS

Responders highlighted the need for clearer guidelines on defining a margin, more and better stakeholder involvement in its selection, and better communication tools that enable discussions about non-inferiority trials with stakeholders. Future research should focus on developing best practice recommendations.

Efficacy and safety of temperature-sensitive acellular dermal matrix in prevention of postoperative adhesion after thyroidectomy: A randomized, multicenter, double-blind, non-inferiority study

Introduction

MegaShield^® is a newly developed temperature-sensitive anti-adhesive containing micronized acellular dermal matrix. The aim of this study was to investigate the efficacy and safety of MegaShield^® compared with Guardix-SG^® in the prevention of adhesions in patients undergoing bilateral total thyroidectomy.

Method

We conducted a multicenter trial between October 2018 and March 2020 in patients undergoing total thyroidectomy. The patients were randomly assigned to either the MegaShield^® group or the Guardix-SG^® group. The primary outcome was the esophageal movement using marshmallow six weeks after the surgery and the secondary outcome was the assessed adhesion score. The safety assessment was also evaluated.

Results

The study included 70 patients each in the MegaShield^® and control (Guardix-SG^®) groups. Baseline clinical characteristics, the mean score of marshmallow esophagography, and the sum of adhesion scores were not statistically different between the two groups. Inferiority test demonstrated that the efficacy of MegaShield^® is not inferior to that of Guardix-SG®. There were no device-related complications in both groups.

Conclusion

The efficacy and safety of MegaShield^® were not inferior than those of Guardix-SG^®. MegaShield^® demonstrated the potential of ADM as a potential future anti-adhesive agent.

Trial registration

The name of trial registry CRIS (Clinical Research Information Service) https://cris.nih.go.kr/cris/index.jsp. (The full trial protocol can be accessed) Registration number: KCT0003204.

Guided, internet based, cognitive behavioural therapy for post-traumatic stress disorder: pragmatic, multicentre, randomised controlled non-inferiority trial (RAPID)

OBJECTIVE

To determine if guided internet based cognitive behavioural therapy with a trauma focus (CBT-TF) is non-inferior to individual face-to-face CBT-TF for mild to moderate post-traumatic stress disorder (PTSD) to one traumatic event.

DESIGN

Pragmatic, multicentre, randomised controlled non-inferiority trial (RAPID).

SETTING

Primary and secondary mental health settings across the UK's NHS.

PARTICIPANTS

196 adults with a primary diagnosis of mild to moderate PTSD were randomised in a 1:1 ratio to one of two interventions, with 82% retention at 16 weeks and 71% retention at 52 weeks. 19 participants and 10 therapists were purposively sampled and interviewed for evaluation of the process.

INTERVENTIONS

Up to 12 face-to-face, manual based, individual CBT-TF sessions, each lasting 60-90 minutes; or guided internet based CBT-TF with an eight step online programme, with up to three hours of contact with a therapist and four brief telephone calls or email contacts between sessions.

MAIN OUTCOME MEASURES

Primary outcome was the Clinician Administered PTSD Scale for DSM-5 (CAPS-5) at 16 weeks after randomisation (diagnosis of PTSD based on the criteria of the Diagnostic and Statistical Manual of Mental Disorders, fifth edition, DSM-5). Secondary outcomes included severity of PTSD symptoms at 52 weeks, and functioning, symptoms of depression and anxiety, use of alcohol, and perceived social support at 16 and 52 weeks after randomisation.

RESULTS

Non-inferiority was found at the primary endpoint of 16 weeks on the CAPS-5 (mean difference 1.01, one sided 95% confidence interval -∞ to 3.90, non-inferiority P=0.012). Improvements in CAPS-5 score of more than 60% in the two groups were maintained at 52 weeks, but the non-inferiority results were inconclusive in favour of face-to-face CBT-TF at this time point (3.20, -∞ to 6.00, P=0.15). Guided internet based CBT-TF was significantly (P<0.001) cheaper than face-to-face CBT-TF and seemed to be acceptable and well tolerated by participants. The main themes of the qualitative analysis were facilitators and barriers to engagement with guided internet based CBT-TF, treatment outcomes, and considerations for its future implementation.

CONCLUSIONS

Guided internet based CBT-TF for mild to moderate PTSD to one traumatic event was non-inferior to individual face-to-face CBT-TF and should be considered a first line treatment for people with this condition.

TRIAL REGISTRATION

ISRCTN13697710.

Bayesian sample size determination in a three-arm non-inferiority trial with binary endpoints

A three-arm non-inferiority trial including a test treatment, a reference treatment, and a placebo is recommended to assess the assay sensitivity and internal validity of a trial when applicable. Existing methods for designing and analyzing three-arm trials with binary endpoints are mainly developed from a frequentist viewpoint. However, these methods largely depend on large sample theories. To alleviate this problem, we propose two fully Bayesian approaches, the posterior variance approach and Bayes factor approach, to determine sample size required in a three-arm non-inferiority trial with binary endpoints. Simulation studies are conducted to investigate the performance of the proposed Bayesian methods. An example is illustrated by the proposed methodologies. Bayes factor method always leads to smaller sample sizes than the posterior variance method, utilizing the historical data can reduce the required sample size, simultaneous test requires more sample size to achieve the desired power than the non-inferiority test, the selection of the hyperparameters has a relatively large effect on the required sample size. When only controlling the posterior variance, the posterior variance criterion is a simple and effective option for obtaining a rough outcome. When conducting a previous clinical trial, it is recommended to use the Bayes factor criterion in practical applications.

The use of real-world data/evidence in regulatory submissions

The 21st Century Cures Act passed by the United States (US) Congress in December 2016 requires the US Food and Drug Administration (FDA) shall establish a program to evaluate the potential use of real-world evidence (RWE) which is generated from real-world data (RWD) to (i) support approval of new indication for a drug approved under section 505 (c) and (ii) satisfy post-approval study requirements. RWE offers the opportunities to develop robust evidence using high-quality data and sophisticated methods for producing causal-effect estimates regardless randomization is feasible. In this article, we have demonstrated that the assessment of treatment effect (RWE) based on RWD could be biased due to the potential selection and information biases of RWD. Although fit-for-purpose RWE may meet regulatory standards under certain assumptions, it is not the same as substantial evidence (current regulatory standard in support of approval of regulatory submission). In practice, it is then suggested that when there are gaps between fit-for-purpose RWE and substantial evidence, we should make efforts to fill these gaps based on a comprehensive evaluation of the treatment effect. We also review two RWE examples to show some potential use of RWE in clinical studies.

Student Perception of Case-based Teaching by Near-Peers and Faculty during the Internal Medicine Clerkship: A Noninferiority Study

INTRODUCTION

Third-year medical students traditionally receive their didactic or small group teaching sessions from clinical faculty during clerkship rotations. Near-peer teaching is increasingly recognized as an acceptable method for teaching, however most near-peer teaching takes place during the pre-clinical curriculum. We sought to determine if fourth year medical students were noninferior to faculty in facilitating small group discussions during clerkship rotations.

METHODS

Seventy-five third-year medical students participated in a small group session focused on rheumatologic diseases during their internal medicine clerkship rotation. Students were taught by fourth-year medical students who self-selected to participate as near-peer teachers at 1 clinical site (near-peers, N = 36) and by clinical faculty at another site (N = 39). At the end of the session, third-year medical students completed a survey evaluating teacher performance and effectiveness.

RESULTS

There was no significant difference between the 2 groups on each of the 17 survey items assessing teacher performance, the total teaching performance score, and the teaching effectiveness rating (all P-values >.05). A mean between-group difference of 2% in favor of the near-peers indicated noninferiority of the near-peer teachers compared with faculty teachers on the total teaching performance score. An absolute difference of 14% in favor of the near-peers indicated noninferiority of the near-peer teachers compared with faculty teachers on the teaching effectiveness score. Near-peer teachers reported several benefits, including improving their own medical knowledge and skills as a future educator.

DISCUSSION

Our data supports the noninferiority of the perceived performance and effectiveness of near-peer teachers compared to faculty teachers in the clerkship setting. Adding near-peer teachers to the clerkship setting is feasible and can be beneficial to all stakeholders.

Randomized, controlled, clinical trial to evaluate a xenogeneic collagen matrix as an alternative to free gingival grafting for oral soft tissue augmentation: A 6- to 8-year follow-up

BACKGROUND

The purpose of this follow up study was to determine if a xenogeneic collagen matrix (CMX) is as effective as free gingival graft (FGG) in preventing further recession 6+ years following vestibuloplasty.

METHODS

This study was a single-blind (examiner), randomized, controlled, split-mouth study of 30 subjects with insufficient zones of KT (< 2 mm), associated with at least two, paired teeth. The study utilized a within subject treatment comparison to examine non-inferiority according to primary and secondary endpoints 6+ years after therapy. The original study primary efficacy endpoint was keratinized tissue width (KTw); however, in this report, prevention of recession (Rec) was also examined, along with traditional, secondary clinical measures, histopathology of mucosal biopsies and exploratory, patient reported outcomes (PROs) for pain and satisfaction.

RESULTS

A total of 23 of the 30 original, study patients were available for 6 to 8-year postoperative assessment, and these patients were representative of the original patient population. For preventing further Rec, CMX was not inferior to FGG (ΔRec = -0.07 ± 1.26 mm for CMX and -0.17 ± 0.78 mm for FGG, P = 0.710). There were no adverse results observed, and histological assessment indicated normal, keratinized gingiva for both therapies. Tissue texture and color match to surrounding, native tissues were significantly better for CMX, and patients preferred CMX over FGG therapy.

CONCLUSIONS

CMX appears to be a suitable substitute for FGG 6+ years after therapy.

Nutrient Challenge Testing Is Not Equivalent to Scintigraphy-Lactulose Hydrogen Breath Testing in Diagnosing Small Intestinal Bacterial Overgrowth

Background/Aims

Small intestinal bacterial overgrowth (SIBO) is a common condition in disorders of gut-brain interaction (DGBI). Recently, a combined scintigraphy–lactulose hydrogen breath test (ScLHBT) was described as an accurate tool diagnosing SIBO. We aim to analyze whether a lactulose nutrient challenge test (NCT), previously shown to separate DGBI from healthy volunteers, is equivalent to ScLHBT in diagnosing SIBO.

Methods

We studied data of 81 DGBI patients undergoing ScLHBT with 30 g lactulose and 300 mL water as well as NCT with 30 g lactulose and a 400 mL liquid test meal. Differences in proportion of positive SIBO diagnoses according to specified cecal load and time criteria for NCT and ScLHBT, respectively, were tested in an equivalence trial. An odds ratio (OR) range of 0.80-1.25 was considered equivalent.

Results

Diagnosis of SIBO during NCT was not equivalent to SIBO diagnosis in ScLHBT, considering a hydrogen increase before cecal load of 5.0%, 7.5%, or 10.0%, respectively ([OR, 3.76; 90% CI, 1.99-7.09], [OR, 1.87; 90% CI, 1.06-3.27], and [OR, 1.11; 90% CI, 0.65-1.89]). Considering only time to hydrogen increase as criterion, the odds of a positive SIBO diagnosis in the NCT (0.65) was lower than in ScLHBT (1.70) (OR, 0.38; 90% CI, 0.23-0.65).

Conclusions

This study could not show an equivalence of NCT and ScLHBT in diagnosing SIBO. A possible explanation might be the different transit times owing to unequal testing substances. The effect of this deviation in relation to consecutive therapy regimens should be tested in further prospective studies.

Simultaneous confidence interval for assessing non-inferiority with assay sensitivity in a three-arm trial with binary endpoints

A three-arm trial including an experimental treatment, an active reference treatment and a placebo is often used to assess the non-inferiority (NI) with assay sensitivity of an experimental treatment. Various hypothesis-test-based approaches via a fraction or pre-specified margin have been proposed to assess the NI with assay sensitivity in a three-arm trial. There is little work done on confidence interval in a three-arm trial. This paper develops a hybrid approach to construct simultaneous confidence interval for assessing NI and assay sensitivity in a three-arm trial. For comparison, we present normal-approximation-based and bootstrap-resampling-based simultaneous confidence intervals. Simulation studies evidence that the hybrid approach with the Wilson score statistic performs better than other approaches in terms of empirical coverage probability and mesial-non-coverage probability. An example is used to illustrate the proposed approaches.

A reproducibility probability-based bias-adjustment approach on the specification of non-inferiority margin using historical data

The effect of reference treatment over placebo, known as M1, is essential in the development of non-inferiority margin. We proposed a M1 adjustment approach to reduce the selection bias for collected data of historical trials. A quantitative illustration of selection bias of historical data is also defined. Simulation study shows that the proposed approaches would significantly reduce the bias when the proportion of positive studies in historical data is noticeably larger than the power of studies include in historical data. When historical data are constituted by only positive studies, the performance of the proposed method is also appreciable. However, when the proportion of positive studies is close to the power of studies included or the number of studies included is too small, the performance of the proposed approach may not be reliable. A real-data application is also presented. The proposed bias-adjustment approach is a reasonable method to reduce the over-estimate of effect size in the specification of non-inferiority margin. It could also be applied in most non-inferiority margin specification methods or be cooperate used with other bias-adjustment approaches.

Pragmatic RAndomised controlled trial of a trauma-focused guided self-help Programme versus InDividual trauma-focused cognitive Behavioural therapy for post-traumatic stress disorder (RAPID): trial protocol

BACKGROUND

There is good evidence that trauma-focused therapies for Post-Traumatic Stress Disorder are effective. However, they are not always feasible to deliver due a shortage of trained therapists and demands on the patient. An online trauma-focused Guided Self-Help (GSH) programme which could overcome these barriers has shown promise in a pilot study. This study will be the first to evaluate GSH against standard face-to-face therapy to assess its suitability for use in the NHS.

METHODS

The study is a large-scale multi-centre pragmatic randomised controlled non-inferiority trial, with assessors masked to treatment allocation. One hundred and ninety-two participants will be randomly allocated to receive either face-to-face trauma-focused cognitive behaviour therapy (TFCBT) or trauma-focused online guided self-help (GSH). The primary outcome will be the severity of symptoms of PTSD over the previous week as measured by the Clinician Administered PTSD Scale for DSM5 (CAPS-5) at 16 weeks post-randomisation. Secondary outcome measures include PTSD symptoms over the previous month as measured by the CAPS-5 at 52 weeks plus the Impact of Event Scale - revised (IES-R), Work and Social Adjustment Scale (WSAS), Patient Health Questionnaire-9 (PHQ-9), General Anxiety Disorder-7 (GAD-7), Alcohol Use Disorders Test (AUDIT-O), Multidimensional Scale for Perceived Social Support (MSPSS), short Post-Traumatic Cognitions Inventory (PTCI), Insomnia Severity Index (ISI) and General Self Efficacy Scale (GSES) measured at 16 and 52 weeks post-randomisation. Changes in health-related quality of life will be measured by the EQ-5D and the level of healthcare resource utilisation for health economic analysis will be determined by an amended version of the Client Socio-Demographic and Service Receipt Inventory European Version. The Client Satisfaction Questionnaire (CSQ) will be collected at 16 weeks post-randomisation to evaluate treatment satisfaction.

DISCUSSION

This study will be the first to compare online GSH with usual face-to-face therapy for PTSD. The strengths are that it will test a rigorously developed intervention in a real world setting to inform NHS commissioning. The potential challenges of delivering such a pragmatic study may include participant recruitment, retention and adherence, therapist retention, and fidelity of intervention delivery.

TRIAL REGISTRATION

ISRCTN13697710 registered on 20/12/2016.

Some practical considerations in three-arm non-inferiority trial design

Non-inferiority trials aim to demonstrate whether an experimental therapy is not unacceptably worse than an active reference therapy already in use. When applicable, a three-arm non-inferiority trial, including an experiment therapy, an active reference therapy, and a placebo, is often recommended to assess assay sensitivity and internal validity of a trial. In this paper, we share some practical considerations based on our experience from a phase III three-arm non-inferiority trial. First, we discuss the determination of the total sample size and its optimal allocation based on the overall power of the non-inferiority testing procedure and provide ready-to-use R code for implementation. Second, we consider the non-inferiority goal of 'capturing all possibilities' and show that it naturally corresponds to a simple two-step testing procedure. Finally, using this two-step non-inferiority testing procedure as an example, we compare extensively commonly used frequentist p -value methods with the Bayesian posterior probability approach. Copyright © 2016 John Wiley & Sons, Ltd.

Avoiding ambiguity with the Type I error rate in noninferiority trials

This review article sets out to examine the Type I error rates used in noninferiority trials. Most papers regarding noninferiority trials only state Type I error rate without mentioning clearly which Type I error rate is evaluated. Therefore, the Type I error rate in one paper is often different from the Type I error rate in another paper, which can confuse readers and makes it difficult to understand papers. Which Type I error rate should be evaluated is related directly to which paradigm is employed in the analysis of noninferiority trial, and to how the historical data are treated. This article reviews the characteristics of the within-trial Type I error rate and the unconditional across-trial Type I error rate which have frequently been examined in noninferiority trials. The conditional across-trial Type I error rate is also briefly discussed. In noninferiority trials comparing a new treatment with an active control without a placebo arm, it is argued that the within-trial Type I error rate should be controlled in order to obtain approval of the new treatment from the regulatory agencies. I hope that this article can help readers understand the difference between two paradigms employed in noninferiority trials.

Non-inferiority margins employed in clinical trials in Japan

WHAT IS KNOWN AND OBJECTIVE

Determination of the non-inferiority margin is one of the major and most difficult considerations when planning a non-inferiority clinical trial. This article aims to list the non-inferiority margins employed in recent clinical drug-development trials in Japan.

METHODS

We investigated non-inferiority margins by reviewing new drug-development dossiers for drugs approved between January 2010 and December 2012 in Japan.

RESULTS AND DISCUSSION

We identified 174 non-inferiority trials, where the efficacy of the test drug was compared to that of a control drug. We have described 70 clinical endpoints and the corresponding non-inferiority margins. In antidiabetes drug trials, a margin of 0·4% mean difference in haemoglobin A1c level was used most frequently. In trials for glaucoma and ocular hypertension, 1·5 mmHg mean difference in intra-ocular pressure value was the commonest margin. A 10% margin of proportion difference was the most frequently chosen in trials of anti-infection drugs. We have provided a short description of the methods used to determine the non-inferiority margin.

WHAT IS NEW AND CONCLUSION

We report on the non-inferiority margins used for a range of endpoints in recent drug-development trials for a number of different diseases. We hope that the details would be helpful to those appraising, reporting or designing non-inferiority trials.

Evidence-based medicine: the design and interpretation of noninferiority clinical trials in veterinary medicine

Noninferiority trials are clinical studies designed to demonstrate that an investigational drug is at least as effective as an established treatment within a predetermined margin. They are conducted, in part, because of ethical concerns of administering a placebo to veterinary patients when an established effective treatment exists. The use of noninferiority trial designs has become more common in veterinary medicine with the increasing number of established veterinary therapeutics and the desire to eliminate potential pain or distress in a placebo-controlled study. Selecting the appropriate active control and an a priori noninferiority margin between the investigational and active control drug are unique and critical design factors for noninferiority studies. Without reliable historical knowledge of the disease response in the absence of treatment and of the response to the selected active control drug, proper design and interpretation of a noninferiority trial is not possible. Despite the appeal of conducting noninferiority trials to eliminate ethical concerns of placebo-controlled studies, there are real limitations and possible ethical conundrums associated with noninferiority trials. The consequences of incorrect study conclusions because of poor noninferiority trial design need careful attention. Alternative trial designs to typical noninferiority studies exist, but these too have limitations and must also be carefully considered.

Two new covariate adjustment methods for non-inferiority assessment of binary clinical trials data

In clinical trials, examining the adjusted treatment difference has become the preferred way to establish non-inferiority (NI) in cases involving a binary endpoint. However, current methods are inadequate in the area of covariate adjustment. In this paper, we introduce two new methods, nonparametric and parametric, of using the probability and probability (P-P) curve to address the issue of unadjusted categorical covariates in the traditional assessment of NI in clinical trials. We also show that the area under the P-P curve is a valid alternative for assessing NI using the adjusted treatment difference, and we compute this area using Mann-Whitney nonparametric statistics. Our simulation studies demonstrate that our proposed methods can not only control type I error at a predefined significance level but also achieve higher statistical power than those of traditional parametric and nonparametric methods that overlook covariate adjustment, especially when covariates are unbalanced in the two treatment groups. We illustrate the effectiveness of our methodology with data from clinical trials of a therapy for coronary heart disease.

A covariate-adjustment regression model approach to noninferiority margin definition

To maintain the interpretability of the effect of experimental treatment (EXP) obtained from a noninferiority trial, current statistical approaches often require the constancy assumption. This assumption typically requires that the control treatment effect in the population of the active control trial is the same as its effect presented in the population of the historical trial. To prevent constancy assumption violation, clinical trial sponsors were recommended to make sure that the design of the active control trial is as close to the design of the historical trial as possible. However, these rigorous requirements are rarely fulfilled in practice. The inevitable discrepancies between the historical trial and the active control trial have led to debates on many controversial issues. Without support from a well-developed quantitative method to determine the impact of the discrepancies on the constancy assumption violation, a correct judgment seems difficult. In this paper, we present a covariate-adjustment generalized linear regression model approach to achieve two goals: (1) to quantify the impact of population difference between the historical trial and the active control trial on the degree of constancy assumption violation and (2) to redefine the active control treatment effect in the active control trial population if the quantification suggests an unacceptable violation. Through achieving goal (1), we examine whether or not a population difference leads to an unacceptable violation. Through achieving goal (2), we redefine the noninferiority margin if the violation is unacceptable. This approach allows us to correctly determine the effect of EXP in the noninferiority trial population when constancy assumption is violated due to the population difference. We illustrate the covariate-adjustment approach through a case study.

Mixed noninferiority margin and statistical tests in active controlled trials

In an active controlled noninferiority trial without a placebo arm, one of the major considerations is the selection of the noninferiority margin. Although the ICH E10 guideline provides general principles for the selection of appropriate noninferiority margins, there are no established rules or gold standards for the selection of noninferiority margins in active control trials. Hung et al. (2003) proposed a margin selection based on relative risk. However, with relative risk, it is difficult to adjust for covariates. On the other hand, Chow and Shao (2006) proposed a method for selecting noninferiority margins based on treatment difference. The determination of noninferiority margin based on either a test for treatment difference or a test for relative risk would be critical. In this paper, we propose a method for noninferiority testing with the use of a mixed null hypothesis. The mixed null hypothesis consists of a margin based on treatment difference and a margin based on relative risk. Both noninferiority margins will simultaneously satisfy the principles as described in the ICH E10 guideline. Statistical tests for mixed noninferiority margin are also derived. An example concerning the efficacy of a test therapy to an active control on a clinical adverse event in the target patient population with cardiovascular disease is presented to illustrate the proposed method. Simulation studies were also conducted to assess the type I error rate and the power.

Issues with Statistical Risks for Testing Methods in Noninferiority Trial Without a Placebo ARM

Noninferiority trials without a placebo arm often require an indirect statistical inference for assessing the effect of a test treatment relative to the placebo effect or relative to the effect of the selected active control treatment. The indirect inference involves the direct comparison of the test treatment with the active control from the noninferiority trial and the assessment, via some type of meta-analyses, of the effect of the active control relative to a placebo from historical studies. The traditional within-noninferiority-trial Type I error rate cannot ascertain the statistical risks associated with the indirect inference, though this error rate is of the primary consideration under the frequentist statistical framework. Another kind of Type I error rate, known as across-trial Type I error rate, needs to be considered in order that the statistical risks associated with the indirect inference can be controlled at a small level. Consideration of the two kinds of Type I error rates is also important for defining a noninferiority margin. For the indirect statistical inference, the practical utility of any method that controls only the across-trial Type I error rate at a fixed small level is limited.