Key Take Away :
This webinar will provide details on how to incorporate the population groups into the protocol and statistical analysis plan, as well as which variables and endpoint should be used in the ITT population and which ones are best investigated inside the PP population.
Overview :
No study is perfect. Subjects drop out, or may not complete all visits. Data may also be missing due to errors at the site or lab. Intention to treat (ITT) analysis, includes every subject who is randomized into the study in the data analysis, regardless of what happens afterwards.
Analysis with an ITT population gives good estimates of an “assignment effect”. However, using the ITT population in a study with missing-ness may result in inflated Type II error. There are better ways to measure the effects on study completers, also called the adherence effect.
This webinar is a presentation of ITT analysis and alternatives. (ITT), is favoured by CONSORT and FDA as a way to estimate treatment effects. IT is favored because all subjects are included regardless of compliance, which avoids the possibility of bias in treatment outcomes due to removal of subject records. However, ITT analysis of study effects is often too conservative and can increase Type II error (not seeing significant effects that are truly present).
Alternatives to the adherence to strict ITT analysis will be presented and include identification of ITT and per protocol (PP) populations and modified intention to treat analysis (mITT).
Why Should You Attend :
Subjects fail to complete a study for any number of reasons. And most studies have missing data due to deviations or non-compliance to protocol. Building a solid statistical analysis plan that takes subject attrition and protocol deviations into account at the planning stage of a study can allow for more accurate estimates of treatment effects.
Bias is always of concern in any research. ITT analysis is preferred by FDA and CONSORT to control for prognostic differences in a randomized controlled trial. However, ITT, commonly referred to as “once randomized, always analyzed”, can result in bias towards “no effect” and increase Type II error.
Why spend time and money on a research without putting necessary checks into place for the inevitable protocol deviations? Proper planning during the design phases of a study can help researchers to see efficacy endpoints as relates to study completers without the watering down of findings by non-adherence of subjects or by protocol deviations.
This webinar is a must see for researchers who want to build a more accurate and more informative study.
Areas Covered In This Webinar :
Overview of ITT
Intention to treat (ITT) analysis includes all patients who were enrolled and randomly allocated to treatment group, i.e. “once randomized, always analyzed”
Inclusion in the ITT population regardless of any deviations that may happen after randomization, including:
protocol violations (subjects received the wrong dose or treatment, the laboratory was unable to analyze a blood sample, etc.)
lost to follow up
voluntary withdrawals from the study
adverse events
Subject non-compliance
Preferred by FDA and CONSORT because ITT analysis preserves the prognostic balance generated by the planned random allocation to treatment.
The Good
Again, FDA and CONSORT support ITT
Since non-compliance and protocol violations are common in a study, ITT gives unbiased estimates of true treatment effectiveness by replicating what happens in the ‘real world’
Reduces bias from estimates using only “completers”
Preserves baseline balance between groups in randomized trials
Since all subjects are kept in the analysis regardless of study completion, ITT helps to guard against “cherry picking” of subject records for analysis.
Minimizes Type 1 errors (seeing significant effects when they are not really there)
Preserves sample size and study power
The Bad
Estimate of treatment effect is conservative because of dilution due to noncompliance and more prone to Type 2 errors (‘false negatives’)
Heterogeneity is introduced when non-compliant, dropouts and compliant subjects are mixed together
Does not assess treatment efficacy accurately unless there are negligible protocol violations, etc.
Protocol violations and poorly conducted trials may cause the results obtained from two different treatment groups to appear similar so ITT analysis alone is inappropriate for non-inferiority trials
The Alternatives
Use of two populations: ‘Per protocol’ and “Safety”
PP only includes individuals who adhered to the clinical trial instructions as specified in the study protocol
Use of a PP population for efficacy endpoints and an ITT group for safety endpoints can allow for more accurate estimates of treatment effects while also providing better views of safety outcomes modified ITT’ (mITT) analysis
Excludes subjects according to criteria which are planned for in the protocol and statistical analysis plan
So many ways to modify that consistency in application is hard to achieve. Everyone has their idea of what can and cannot be modified
Since it is so subjective, mITT can introduce bias in estimation of study outcomes
Learning Objectives :
Understand regulations and guidelines requiring ITT analysis
Viable adjustments and alternatives to the approach
Who Will Benefit :
Clinical Trail Sponsors,
Investigators/ Clinical Personnel who handle CRF and Data Collection,
Statisticians new to the field of Clinical Research
Speakers Profile :
Elaine Eisenbeisz
Elaine Eisenbeisz is a private practice statistician and owner of Omega Statistics, a statistical consulting firm based in Southern California. Elaine has over 30 years of experience in creating data and information solutions for industries ranging from governmental agencies and corporations, to start-up companies and individual researchers.
In addition to her technical expertise, Elaine possesses a talent for conveying statistical concepts and results in a way that people can intuitively understand.
Elaine’s love of numbers began in elementary school where she placed in regional and state-wide mathematics competitions. She attended University of California, Riverside, as a National Science Foundation scholar, where she earned a B.S. in Statistics with a minor in Quantitative Management, Accounting. Elaine received her Master’s Certification in Applied Statistcs from Texas A&M, and is currently finishing her graduate studies at Rochester Institute of Technology. Elaine is a member in good standing with the American Statistical Association as well as many other professional organizations. She is also a member of the Mensa High IQ Society. Omega Statistics holds an A+ rating with the Better Business Bureau.
