Insufficient dosing could be a problem; however, the more likely explanation (IMO) for the observed efficacy (to date) being less than the 23% modeled by the company is (as noted in #msg-110961835) that the drug used in the control arm, Lactulose, is quite effective in a brute-force way.
I think there's some truth to that, but there's probably more to it since the primary endpoint isn't response rate per se, but time to response. The co's story appears to be 1) serum NH3 correlates to HE severity; 2) OCR-002 reduces NH3 levels much more rapidly than Lactulose alone, leading to shorter time to symptom abatement. Perhaps though transient blood NH3 levels aren't as well correlated to time to symptom improvement and there are other factors associated with gut clearance that have a contributory effect (see link below to PEG trial). And as JQ1234 noted, it's difficult to make sense of the phase IIa dose response data and the whole patient variability thing - it doesn't help that it's presented inconsistently between parts A and B. Would be nice for the company to show waterfall plots of change in NH3 levels.