Menu
It is typical for policy breaches or failure to assess outcomes to occur when executing clinical studies. This article focuses on the significance of intention-to-treat analysis while outlining the difficulties involved in interpreting the findings of such investigations.
Why bring up these issues when discussing ITT vs. PP? Let’s begin with definitions and descriptions that are more general.
Contrary to ITT, per-protocol analysis refers to only including patients who have strictly followed the regimen in the study. The per-protocol analysis estimates the intervention’s efficacy or how well it worked among those who followed the assigned treatment plan.
The statistics do not represent reality as it stands and may demonstrate an inflated therapeutic effect. As per the intention-to-treat concept, every randomly assigned participant to the clinical study must participate in the primary analysis.
Patients that discontinue therapy midway do not adhere to the dosing interval or even engage in the research’s maltreatment and are, therefore, part of the primary analysis within the specific treatment category to which they were randomly assigned during the allocation. ITT and PP populations are frequently examined in Randomized Controlled Trials (RCT).
All participants examined in their assigned treatment arms are included in the ITT population, regardless of whether they have received the therapy or finished the trial.
ITT should be reviewed as a strategy for designing and interpreting an RCT rather than just a data analysis technique. An ITT analysis aims to evaluate the average impact of giving medicine to a group of participants.
The main goal of ITT assessment is to prevent selection bias brought on by treatment assignment according to patient prognosis /expected therapy response while maintaining the effect of randomization. This compares participants in trial groups at treatment initiation for known and unknown prognostic variables.
To accomplish this, individuals who break the study rules, fail to take the study medications as prescribed, or withdraw from the study prematurely are regarded as being in the intervention arm to which they were originally randomized.
Due to the “once allocated to this group, always allocated to this group” concept, the outcome of this technique is that all participants recruited in the study are taken into account in the primary analysis.
ITT assessment also has several useful benefits. The study efficacy is unhinged because no patients are left out, reflecting everyday patient care where deviance from therapeutic plans commonly occurs.
Randomization also retains the baseline harmony of risk factors between experimental groups. The PP analysis is an alternative to ITT.
The impact of receiving the designated treatment options for the duration of the follow-up session can be examined by researchers via the PP methodology. The ITT demographic of participants who concluded the trial without materially deviating from the criteria make up the PP subgroup.
For these purposes, PP analysis captures the “genuine” publicity to treatment/placebo because it does not take into account patients who break the rules, such as those who transitioned the allocation throughout the research, did not follow the prescribed course of treatment, or skipped scheduled evaluations over time.
The critical issue is that, unlike the ITT group, we cannot be sure that individuals in the two research arms were equivalent at the start of the study for both identified and unidentified risk factors in the PP subgroup. The benefits of randomization could be compromised as a result.
In superiority, Randomized Controlled Trials (RCTs), ITT evaluation is considered conservative because it ignores the mitigation of the between-group difference caused by disobedience and cross-over. As a result, innovative therapy is less likely to be recognized as beneficial by the ITT.
In superiority RCT, the ITT is considered the primary assessment technique, and the PP method is frequently used as a supplementary, supportive analysis. Although the validity of the ITT policy is widely accepted in superiority RCTs, there is less agreement on its applicability in equivalence and non-inferiority RCTs.
In equivalence and non-inferiority RCTs, which seek to show the similarities of two medications, the ITT technique frequently supports the study hypothesis as non-compliance and cross-over generally reduce the between-arms disparities. As a result, the non-inferiority and equivalence of two medications should only be asserted when the findings of the ITT and PP assessments are consistent.
However, according to the Committee for Proprietary Medicinal Products (CPMP) standards, “in a non-inferiority study, ITT and PP assessments have equal worth, and their outcomes should lead to equivalent findings for a credible interpretation.
As “ITT and PP yield equal findings, the confidence of the researcher for the research results is reinforced,” the CONSORT (Consolidated Standards for Reporting Trials) recommendations highly advise presenting the details of both estimations in clinical studies.
As it retains the benefits of randomization, the ITT evaluation should continue to be the primary analysis in this situation. Non-inferiority and equivalence studies should employ the PP analysis as a supplementary sensitivity analysis.
ITT and PP procedures are generally valid, although their ranges and interpretations vary. Researchers and medical professionals may be intrigued by a drug’s PP effect for at least two reasons. First, if they could increase adherence to the protocol, researchers might want to understand the experimental medicine results.
For instance, an ITT analysis can indicate that medication was unsuccessful, whereas it was helpful for PP group participants who strictly followed the trial protocol. In this situation, the PP results suggest that increased adherence to the studied medicine would enhance clinical outcomes.
However, it is crucial to understand that in the PP subgroup, it is impossible to determine whether the treatment response is caused entirely by adherence or by patient characteristics associated with compliance that may positively impact the result without reference to the medicine. Second, patients are most concerned about the PP impact when determining if they should take the medication.
This choice should not just be predicated on the possible advantage of being merely “assigned” to the therapy but also on the likelihood of benefit if the participant chooses “to take” the medication, as determined by the PP analysis.
To understand the above-stated comparisons, let us look at a study performed by the Women Health Initiative, WHI. The WHI researchers conducted an RCT to compare the effects of postmenopausal estrogen plus progestin hormone replacement therapy and placebo on wellbeing.
They analyzed the ITT and PP techniques to determine the association between the allocated group and the breast cancer hazard rate. In the ITT assessment, women in the active group had a 25% greater risk of developing breast cancer than those in the placebo group.
This increased risk is a result of both inadequate adherence to the prescribed course of treatment and the actual effects of hormone therapy. In actuality, only 60% of study subjects in both groups were still taking their prescribed medications after six years.
On the contrary, a PP analysis discovered that people in the active group had a 68% greater chance of acquiring breast cancer than those in the placebo group. This discrepancy indicates that continuing use of postmenopausal estrogen + progestin hormone raises the risk of breast cancer by 68%, meaning that the impact of the administered medicine was more than two times greater than the impact of the prescribed medication.
Therefore, giving women simply the findings of the ITT analysis may be deceptive, and it might be argued that women were not given all the information on the true rise in breast cancer incidence while taking the medication.
From this angle, it is evident that ITT and PP studies offer complimentary knowledge regarding the efficacy or harm of a particular medicine. PP analyses are, therefore, a helpful exercise, even though they are subject to bias, because they give an approximation of the benefit/harm of a treatment when adherence is sustained.
Incorporating the ITT study in an RCT retains the sample size, lowers bias, and answers the question, “What is the impact of administering a medicine to a patient group?”
It also maintains the original between-arms comparison as determined by randomization. The question “What is the influence of administering a medication in a patient group?” can be answered using PP analysis, which, in contrast, solely takes into account patients who strictly followed the protocol.
We ought to be cautious that when we apply a PP evaluation, we only consider “patients effectively using the medicine,” which could lead to an overly positive treatment outcome given that protocol violations could also be caused by significant side effects of the advised intervention.
ITT and PP analyses differ in that they compare the impacts of “allocating a drug” and “using a drug,” as well as the consequences of “assigning a medicine” and “using a medicine successfully without experiencing substantial adverse effects.”
Sign up here to enjoy new blog articles about biostatistics, clinical data analytics, and stat programming