help with study sample size and statistical analysis


I'm in the preparation stages of an RCT looking to assert the effectiveness of psychological therapies for a health condition.

The design will be a 3 parallel arm RCT, comparing high intensity CBT, low intensity CBT and usual care. We are thinking of collecting baseline measures (2 time points pre-treatment), 1 post-treatment, and 3 post treatment (3,6,and 12 months after).
Our main effectiveness measure will be a symptom severity measure. In the case, any of the treatment arms are effective, we are expecting to find (from previous studies) at least a mean 50 point (100 SD) significant decrease (this is clinical significance, not statistical significance). We will also have length of condition and previous psychiatric diagnosis as covariates.

Now, my question regards essentially, to what type of statistical analysis should I go for on my main outcome analysis. ANCOVA? MANOVA?
From what I gathered the type of analysis I chose will affect the sample size. I'm using GPOwer as a sample size calculator and I know that different options will lead to different group sizes.

So, could anybody give me a hand here? Really strugling with this.

Will be happy to give more information if required.

Thank you!