Question about intention to treat in a behavioural intervention RCT

[Jeanette Reece]

Hello,

I have a query regarding the intention-to-treat (ITT) analysis for our clinical trial, which focuses on assessing the quality of life (QoL) score before and 6 months after an online course.

Initially, a total of 100 individuals submitted a baseline survey and were randomly allocated to either the intervention group (n=50) or the active control group (n=50). The average QoL score at baseline was 20 for both groups.

Later on, 40 individuals in each group completed either the intervention course or the active control course. Subsequently, out of these, n=20 in the intervention and n=30 in the control group submitted the 6-month follow-up survey.

The average QoL score at the 6-month mark was 25 for both groups.

Could you confirm if the comparisons of samples provided below are accurate?


(1) Intention to treat (ITT): (Intervention: QoL score 25 - QoL score 20 = 5 point (n=20 follow-up vs. n=50 baseline)
vs.
Control : QoL score 25 - QoL score 20 = 5 point (n=30 follow-up vs. n=50 baseline)


(2) Per Protocol: (Intervention: QoL score 25 - QoL score X (n=20 vs. n=20 at both baseline and 6-month follow-up)
vs.
(Control: QoL score 25 - QoL score Y (n=30 vs. n=30 at both baseline and 6-month follow-up)

Thank you so much for your help in advance!

[Tim Morris]

Hi Jeanette

I don't know that Statalist is the best place for this question, but I'm interested in it so here are a couple of thoughts for what they're worth:

1. Intention-to-treat analysis says you include all 50 in the group to which they were originally randomised.
2. This is complicated in the presence of missing data. In this paper, Ian White talked about the intention-to-treat principle.
3. People are increasingly using the language of estimands: this is about specifying the precise question you are interested in, and using that to inform the analysis you will use (see the ICH E9(R1) addendum). ITT analysis sometimes but not always relates to the so-called treatment policy strategy.

Tim