From the Guidelines
Modified Intention-To-Treat (mITT) analysis can lead to overestimation of efficacy in clinical trials under certain circumstances, particularly when exclusions are related to treatment response or adverse events. When participants are excluded from analysis after randomization, the results may no longer reflect the true comparative effectiveness of the interventions being studied. This overestimation occurs because mITT analyses often remove non-compliant patients, those lost to follow-up, or those who didn't receive the full treatment, which tends to retain patients with better outcomes. For example, in a trial comparing a new antihypertensive medication to standard therapy, excluding patients who discontinued treatment due to side effects might artificially inflate the apparent efficacy of the new drug, as noted in studies discussing the importance of intention-to-treat analysis 1.
The magnitude of this bias depends on the specific exclusion criteria used, the proportion of excluded participants, and whether exclusions differ between treatment arms. To minimize this risk, researchers should pre-specify mITT criteria in the study protocol, ensure balanced exclusions across treatment groups, and ideally report both ITT and mITT results for transparency, as suggested by guidelines for improving the design and analysis of clinical trials 2, 3.
Some key points to consider in the analysis of clinical trials include:
- The choice of analysis population, with consideration of how missing data will be handled and the potential impact of subgroup analyses and multiple comparisons 2.
- The importance of preserving the protections of randomization from bias, which can be compromised by exclusions after randomization, especially if these are not balanced across treatment arms 3.
- The role of intention-to-treat (ITT) analysis in providing an unbiased estimate of treatment effect, as it includes all randomized participants regardless of protocol adherence, making it the gold standard for clinical trials 1.
In practice, the decision to use mITT analysis should be made with caution, considering the potential for overestimation of efficacy and the importance of transparency in reporting both ITT and mITT results. The most recent and highest quality studies, such as those published in prestigious journals like the British Journal of Pharmacology 1, emphasize the need for careful consideration of analysis strategies to ensure the validity and generalizability of clinical trial results.
From the Research
Overview of Modified Intention-To-Treat (mITT) Analysis
- The mITT analysis approach excludes participants who do not initiate treatment from the analysis, which can potentially lead to biased estimates of treatment effect if not properly justified 4.
- However, studies have shown that mITT analysis can be an unbiased estimator for the principal stratum estimand under certain assumptions, such as when the intercurrent event is not affected by the assigned treatment arm 4.
Comparison of mITT and Intention-To-Treat (ITT) Analysis
- A study found that mITT trials were more likely to report post-randomization exclusions and were strongly associated with industry funding and authors' conflicts of interest, but there was no association between mITT reporting and favourable results 5.
- Another study found that the treatment effect of active intervention compared to control was comparable between mITT and ITT analysis populations, suggesting that the mITT approach does not introduce bias compared to ITT in certain contexts 6.
Potential Biases and Limitations of mITT Analysis
- The use of mITT analysis can lead to overestimation of efficacy if the excluded participants would have had a different outcome than those who were included in the analysis 7.
- The mITT approach assumes that the intercurrent event is not affected by the assigned treatment arm, which may not always be the case 4.
- The decision to use mITT analysis should be carefully considered and justified, taking into account the specific research question and study design 8.
Key Considerations for Using mITT Analysis
- Researchers should carefully evaluate the assumptions underlying the mITT approach and consider the potential for bias and limitations 4.
- The use of mITT analysis should be transparently reported and justified, with clear explanations for the exclusions made and the potential impact on the results 5.
- Further research is needed to fully understand the implications of using mITT analysis in different contexts and to develop guidelines for its appropriate use 8.