I understand this caveat, but at the same time, a consistent pathomechanism model should also be seen even with a small sample size. For instance, even with 17 patients with diabetes, insulin problems should be found. Of course, I understand things are not as simple when it comes to chronic syndromes like ME/CFS.