Stock DZ
20.04.2018 kl 14:41
2379

Thanks Snøffelen

This link is very good in explaining randomization in clinical trials for non medical expert, it shows why unequal randomisation is valuable for safety read out which is already achieved with FimaChem, and also shows why 1:1 randomisation has cost benefit (less patients to be included) which is the case for PF2 for FimaChem

https://www.linkedin.com/pulse/why-have-unequal-randomisation-ratios-clinical-trials-joshua-brostoff/


"Background:


The classic trial randomisation is in a 1:1 ratio. Some trials have unequal ratios such as 2:1 or 3:1. Such unequal allocations need justification, but the rationale is very rarely stated in the publication.


Reasons for unequal randomisation:

Cost

In a trial one arm is usually less expensive than the other. This can be due to research, care, or drug costs. If you have a fixed budget then randomising more participants to the cheaper arm allows greater overall recruitment.
However in most cases the placebo/control arm is the less expensive one. Thus cost is only rarely a reason for unequal randomisation.
Gathering additional safety information

A larger sample size in the active group gives more power to detect adverse events.
This is most useful when the adverse event is easily related to the trial product. This includes rapidly-occurring toxicities, or ‘signature’ adverse events that are almost always a result of the product.
Unequal randomisation is helpful when obtaining safety information is important even in the absence of efficacy, for example dose ranging studies of novel products
Early phase trials

Classic 1:1 randomised “hypothesis-testing” trial is not well suited for exploring different dose regimes
Unequal randomisation allows more variables to be tested
Learning curves

Some interventions involving surgery or new technologies have a learning curve.
Allocating more patients to the new intervention reduces the effect of the learning curve on the final trial result.
Patients more likely to join a trial if they have a good chance of getting active treatment

Better recruitment may be a rationale for unequal randomisation. Patients may well prefer a trial where they have a 3:1 randomisation of active to placebo
This is usually not scientifically valid. If there is genuine uncertainty about which treatment is better (“equipoise”) then there is no good scientific or ethical reason to have unequal randomisation.
If a high drop-out rate is expected in one arm

Allocating more patients to the arm with a high drop-out rate allows greater power for a “per-protocol” analysis
It still will not interfere with an intention-to-treat (ITT) analysis


Statistical concerns

For the same power a trial that is randomised 2:1 needs 12% more patients, and a 3:1 randomisation scheme requires 33% more patients.
A statistician must look at these closely – they are not simple to design and power correctly


Problems with unequal randomisation

Published reports rarely state why unequal randomisation was chosen.
Complex statistical concerns, and harder for non-experts to assess the statistical validity of the design
No evidence that they increase recruitment
Unethical in some cases – if there is genuine uncertainty about treatment choice then it should be allocated 1:1


Summary

Unequal randomisation (e.g. 2:1 or 3:1) is done quite rarely. There are some occasions when it is justified. Such a randomisation needs to overcome scientific and ethical problems. Such trials should always publically declare why unequal randomisation was chosen".
Arkivert