A noninferiority trial aims to show that an experimental treatment is not clinically worse than an active comparator by more than a predetermined margin. Such trials are increasingly common and can provide evidence for cost-effective treatment alternatives. However, they require strict adherence to design parameters, including establishment of a justified noninferiority margin and sample size calculation based on this margin, in order to correctly interpret results.
Biostatistics are widely used in clinical trials to collect and organize and describe and interpret these result and then give to us proves to take appropriate clinical decisions
Biostatistics are widely used in clinical trials to collect and organize and describe and interpret these result and then give to us proves to take appropriate clinical decisions
Superiority, Equivalence, and Non-Inferiority Trial DesignsKevin Clauson
http://bit.ly/bQKcGz This lecture was presented as part of the Drug Literature Evaluation course at Nova Southeastern University. Guided notes and an audience response system were used to augment to lecture. Context for my decision to share these slides can be found at the provided link.
Clinical Research Statistics for Non-StatisticiansBrook White, PMP
Through real-world examples, this presentation teaches strategies for choosing appropriate outcome measures, methods for analysis and randomization, and sample sizes as well as tips for collecting the right data to answer your scientific questions.
In clinical trials and other scientific studies, an interim analysis is an analysis of data that is conducted before data collection has been completed. If a treatment is particularly beneficial or harmful compared to the concurrent placebo group while the study is on-going, the investigators are ethically obliged to assess that difference using the data at hand and to make a deliberate consideration of terminating the study earlier than planned.
In interim analysis, whenever a new drug shows adverse effect on human being while testing the effectiveness of several drugs, we immediately stop the trial by taking into account the fact that maximum number of patients receive most effective treatment at the earliest stage. Interim analysis is also used to possibly reduce the expected number of patients and to shorten the follow-up time needed to make a conclusion. One wouldn't have to spend extra money if he/she already have enough evidence about the outcome. In this presentation, the total sample size is divided into four equal parts to perform the analysis and decision is made based on each individual step.
Basics of Systematic Review and Meta-analysis: Part 3Rizwan S A
A 4 part lecture series on the basics of Systematic Review and Meta-analysis, Part 3 discusses the software needed and analytical techniques used for this purpose.
Randomization is the process by which allocation of subjects to treatment groups is done by chance, without the ability to predict who is in what group. It is done in clinical trials. This presentation describes the methods of randmization used in clinical trials.
Superiority, Equivalence, and Non-Inferiority Trial DesignsKevin Clauson
http://bit.ly/bQKcGz This lecture was presented as part of the Drug Literature Evaluation course at Nova Southeastern University. Guided notes and an audience response system were used to augment to lecture. Context for my decision to share these slides can be found at the provided link.
Clinical Research Statistics for Non-StatisticiansBrook White, PMP
Through real-world examples, this presentation teaches strategies for choosing appropriate outcome measures, methods for analysis and randomization, and sample sizes as well as tips for collecting the right data to answer your scientific questions.
In clinical trials and other scientific studies, an interim analysis is an analysis of data that is conducted before data collection has been completed. If a treatment is particularly beneficial or harmful compared to the concurrent placebo group while the study is on-going, the investigators are ethically obliged to assess that difference using the data at hand and to make a deliberate consideration of terminating the study earlier than planned.
In interim analysis, whenever a new drug shows adverse effect on human being while testing the effectiveness of several drugs, we immediately stop the trial by taking into account the fact that maximum number of patients receive most effective treatment at the earliest stage. Interim analysis is also used to possibly reduce the expected number of patients and to shorten the follow-up time needed to make a conclusion. One wouldn't have to spend extra money if he/she already have enough evidence about the outcome. In this presentation, the total sample size is divided into four equal parts to perform the analysis and decision is made based on each individual step.
Basics of Systematic Review and Meta-analysis: Part 3Rizwan S A
A 4 part lecture series on the basics of Systematic Review and Meta-analysis, Part 3 discusses the software needed and analytical techniques used for this purpose.
Randomization is the process by which allocation of subjects to treatment groups is done by chance, without the ability to predict who is in what group. It is done in clinical trials. This presentation describes the methods of randmization used in clinical trials.
American Journal of Emergency Medicine: Stroke and first responders strategyEmergency Live
Background
Improving access to thrombolytic therapy for patients with ischemic stroke is challenging. We assessed a prehospital process based on firemen rescuers under strict medical direction, aimed at facilitating thrombolysis of eligible patients.
Methods
This was a prospective observational study conducted over 4 months in Paris, France. Prehospital patients with suspected stroke were included after phone consultation with a physician. If the time since the onset of symptoms was less than 6 hours, patients were transported directly to a neurovascular unit (NVU), if symptom onset was over 6 hours ago; they were transported to an emergency department. Confirmation of stroke diagnosis, the rate of thrombolysis, and the time intervals between the call and hospital arrival and imaging were assessed. Comparison used Fisher's exact test.
Results
Of the 271 patients transported to an NVU, 218 were diagnosed with a stroke (166 with ischemic stroke), 69 received thrombolytic therapy, and the mean stroke-thrombolysis interval was 150 minutes. Over 64 patients admitted to ED, 36 patients suffered a stroke (ischaemic: 24). None were thrombolysed. Globally, 36% of ischaemic strokes were thrombolysed (27% of all strokes diagnosed). The mean interval call-hospital was 65 min (ED vs NVU: p=0.61). The interval call-imaging was 202 min [IQR: 105.5-254.5] for ED and 92 min [IQR: 77 116] for NVU (p<0.001).
Conclusions
The prehospital management of stroke by rescuers, under strict medical direction, seemed to be feasible and effective for selection of patients suffering from stroke in an urban environment, and may improve the access to thrombolysis.
Presentation from Denver Open Source Users Group in February 2015. http://www.meetup.com/DOSUG1/events/219099019/
AngularJS is one of today's hottest JavaScript MVC Frameworks. In this session, we'll explore many concepts it brings to the world of client-side development: dependency injection, directives, filters, routing and two-way data binding. We'll also look at its recommended testing tools and build systems. Finally, you'll learn about my experience developing several real-world applications using AngularJS, HTML5 and Bootstrap.
An overview of the ICH E9 guidance. Easy to follow, and I can provide a live presentation of this to your team! Great for those who are not familiar with statistics.
Evidenced Based Practice (EVP): GRADE Approach to Evidenced Based Guideline D...Michael Changaris
This slide show explores how to review literature and develop an understanding of the quality of the clinical evidence for a treatment modality. Reviews the development of a guideline based on evidence based GRADE process.
4. How are noninferiority trials different from other trials? Superiority Equivalence Noninferiority Objective To determine if one treatment is superior to another To determine if an experimental treatment and an active reference treatment are similar in effect within a prestated range To determine that an experimental treatment is not clinically worse than an active reference intervention by more than a small, preset margin Null Hypothesis There is no difference between the two interventions. One treatment is superior to the other The experimental treatment is equivalent to the comparator within the prestated margin Alternate Hypothesis One treatment is superior to the other One treatment is superior to the other The experimental treatment is no worse than the comparator with respect to the prestated margin Limit P (-∆ to +∆) (0 to +∆)
This is a brief overview of Noninferiority Trials. Look at WHAT they are and WHY we use them Cover the basic info needed to ANALYZE the quality of these reports and draw APPROPRIATE conclusions.
A trial where researchers want to show that the experimental treatment is not WORSE than another treatment by more than a CLINICALLY NEGLIGIBLE AMOUNT.
Like a typical RCT (aka superiority trial), these studies provide comparison info for making evidence-based decisions. So why not just use a head-to-head trial that we’re more familiar with? Consider the following: You have 2 medications in the same class. They performed so similarly against placebo that no one is willing to invest time/money in a head-to-head trial that will likely result in no clear superiority. When we expect the treatments to be the same, a Noninferiority trial can provide a justification OTHER THAN GREATER EFFICACY for choosing one medication over another (ex: decreased cost, better side effect profile, reduced pill burden). Additionally, if a clear superiority or inferiority exists, a properly executed noninferiority trial will spot it.
Essentially, Superiority & Equivalence trials are two distinctly different kinds of trials, with Noninferiority trials as a sort-of hybrid between the two. OBJECTIVE Superiority: Show one treatment is better than the other Equivalence: Show that both treatments are the same (think of bioequivalence trials – AB rating) Noninferiority: Show that the difference between the two treatments is so small, that they might as well be the same (think of therapeutic interchange) HYPOTHESES Superiority: Assume the treatments are the same & hope to show they are actually different Equivalence & Noninferiority: Assume the treatments are different & hope to show they are the same LIMIT Superiority: set an upper limit (P). Results are statistically significant when they are less than that upper limit P Equivalence: Set both an upper AND lower limit. Results are significant when they are contained entirely between the upper & lower limits. Noninferiority: BOTH a range AND an upper limit (∆) ∆=“noninferiority margin”
How likely am I to encounter a noninferiority trial? A study in September’s Pharmacotherapy journal reviewed the frequency of these studies. They found a clear trend of increasing frequency.
Why are they so popular? Placebo control inappropriate: For instance, it would be unethical to replace an active medication with a placebo when investigating a drug for use in acute asthma attack Can be used for physical interventions (ex: surgical procedures) as long as there is an accepted standard of treatment to compare it to. Risk-benefit analyses: Consider our previous example of the two treatments. These are helpful in making formulary decisions. Comparing a drug to itself: allows a manufacturer to set min/max dosing, determine the best titration strategies, and fine tune their formulations Inferior/Noninferior/Superior claims are all appropriate: Superiority trials can ONLY establish whether a superiority exists, or does not exist.
Biggest disadvantage is that these trials have very specific design & reporting requirements. Based on the studies currently available, these requirements may not be fully understood. If there is no reference treatment, if it is controversial, or inconsistent against placebo (ex: antidepressants, Alzheimer meds) then this trial is inappropriate Larger sample size can make a noninferiority trial more difficult & more expensive.
Start with the assumption that the treatments are different. Note the wording of the alternate hypothesis. It is not enough to state that one treatment is not inferior to another. Claims of noninferiority HAVE to reference the margin.
In designing these studies, we start with the standard procedures (ex: randomization, blinding & control) but the control has to meet some specific criteria It must be a well-established therapy There must be at least 1 superiority trial establishing its advantage over placebo. That superiority trial becomes the blueprint for the remaining design decisions (ex: inclusion/exclusion criteria, outcome measures & trial conduct) Like any trial, a limit to measure our results against is set. Standard RCTs state the P value. Noninferiority studies state the upper limit (AKA noninferiority margin) Unique to Noninferiority trials, is that the JUSTIFICATION for that choice MUST be given. It has to be both statistically AND clinically valid.
In setting the margin, we want the smallest value that would be clinically meaningful. The key question to consider – What would be clinically relevant to THIS situation? A clinician might be willing to accept a 5% difference in treatment for the prevention of acne, but not a 5% difference in the prevention of death-due-to-stroke. Ideally, we want a value smaller than the difference between the reference treatment and its placebo. Example: We look at the superiority trial establishing the reference treatment as better than placebo & see that the reference treatment was 12% better than placebo. If we set the margin at 7%, we are accepting that our experimental treatment may be up to 5% worse than the comparator, but at least 7% better than placebo. Once a margin is determined, we use that choice, the size of the CI and the desired study Power to calculate sample size.
Analysis in these trials can be tricky. ITT can shift the bias towards noninferiority. Per-Protocol only accounts for the patients who completed all the study parameters appropriately. It can also introduce bias – and that bias is less predictable. Noninferiority trials should USE and REPORT THE RESULTS from Both Methods. Those results can then be treated equally for a balanced conclusion, or the more pessimistic results can be used for a conservative interpretation.
Results are stated as CI. Both single-sided & two-sided intervals are valid. In any study, reporting the results without comparing them to the prestated limit is inappropriate. In superiority trials, results (typically a single value) are compared with the upper limit P. With THESE studies, the results are typically a range of values that are fully encompassed by the CI. We compare the entire CI with both the established range AND the upper limit.
The established range (AKA Zone of Inferiority) is tinted for clarity. The noninferiority margin is at point ∆. When the entire CI is less than zero (case A), researchers may appropriately claim superiority. When the upper limit of the CI is less than the noninferiority margin (case B & C), researchers can claim noninferiority. Even though part of the CI for case B falls within the “treatment is better” territory, it would be inappropriate for researchers to claim “possible superiority” as it could lead to misunderstanding. When the upper limit of the CI is greater than ∆, the results are, at best, inconclusive. However, if the entire CI is above the margin, researchers should report it as inferior. A claim of “inconclusive” would be inaccurate.
Multiple reviews of noninferiority trials have been done. Reviewers found examples of repeated design and reporting flaws. Recognizing these flaws makes us more effective at evaluating the evidence at hand
Improper or misleading terminology: The terms noninferiority and equivalence are often misused. Authors treat them synonymously. Or they choose ambiguous expressions like “not substantially lower than” or “comparable to,” leading to misinterpretation. It could be simple confusion, however, drug trial reports are now commonly used for marketing purposes. . Manufacturers don’t want to say “Our drug was no more than 5% worse than the one we compared it to.” In providing a positive spin for advertising purposes, the results are often misleading. Inappropriate noninferiority margin: The margin must be determined using both statistical rationale AND clinical judgment. Reviewers found multiple instances in which clinical judgment was discounted = margins that were too large for clinical relevance. A strategy to evaluate appropriateness: compare effect size (absolute difference in treatments ÷ by standard deviation). Researchers may NOT use this approach to set the margin, as it must be stated prior to trial initiation – NOT calculated after the fact. Confused error types: Because the hypotheses of noninferiority & superiority trials are swapped, readers may be confused when authors refer to a Type I or Type II error without giving its precise meaning.
Incomplete results analysis: Researchers commonly report EITHER Intention-To-Treat OR Per-Protocol analysis. Individually, each may lead to biased conclusions. Bias is reduced when both methods are used and ALL RESULTS are reported. Noninferiority claims from a superiority trial : Noninferiority trials can provide evidence for a claim of superiority. BUT, when superiority trials result in a finding of no superiority, we CANNOT infer equivalence or noninferiority. Reviewers found more than one case where superiority trials were wrongly reported as noninferiority or equivalence trials. Be suspicious when an study does not include a justification for the noninferiority margin, when the margin appears to have been calculated at study completion, or when the margin was not referenced in calculating sample size.
Noninferiority trials are gaining in popularity. Increased publication = greater role in influencing the FDA & other decision makers. Need to have a clear understanding of these trials to appropirately recommend or select treatment for the patients in our care.
The quality of a noninferiority study is directly dependent on its adherence to the design and reporting requirements.