I don’t see anything obvious in their driver that would cause refresh() to be run when debug logging is enabled, but perhaps there is some convoluted chain of calls that would make that happen. However, it should be noted that disabling debug logging will hide the “refresh()” lines from the log, perhaps masking the fact that it is still being called, and will also hide some “raw” information coming back from the device (you may not see an “info” log unless it actually gets parsed into an event, which generally necessitates a state change), at least with conventional Hubitat logging. Inovelli’s drivers don’t exactly follow this, so your experience may vary, and again, I don’t use them so I can’t speak with experience.
So…I guess I just wouldn’t be so sure it’s really gone and the fact that it isn’t is just being hidden. 
I suppose an easy way to tell would be to change the line that says if (debugEnable) log.debug "refresh()" to just log.debug "refresh()" (line 1385 in their LZW31-SN driver at the moment, but I’d just look for the code), and then you’ll always get a log entry and would know for sure. Seems odd no matter what the cause…