What is "small", and what is a "fusion cell"?
It is easy to induce fusion with a small particle accelerator. The efficiency is not high enough to use this for a power plant, but that was not part of the question.
The short timescale of energy loss in small objects is certainly an issue in some (not all) fusion concepts.
Say I have built a small tokamak of 1:1000 scale factor to ITER, assuming somehow I shrinked it without losing any function. Would it have one thousandth efficiency?
You cannot scale a tokamak like that.
As an example, you would have to scale the magnetic field up by a factor of 1000, and that alone is completely impossible.
The size of the plasma relative to the Larmor radius would change if you do not adjust the field (or the temperature, but then you can completely forget fusion).
It gets easier for particles to escape, as that radius depends how much their position orthogonal to the magnetic field varies. In the extreme case, your plasma vessel is smaller than the radius, then the particles will hit the wall within microseconds.