It depends. For some parameters it definitely matters and I wouldn't pool positions if I can avoid it, but the real question is what was specified/asked for in the first place. If you didn't tell the assessor they should measure a specific position, you might always get a mix and possibly not even have something that qualifies for baseline in the desired position. In that case it could be preferable to use the 'wrong' position rather than not having a baseline at all. Or your medic may not care about (differences in) position at all in the first place, and you can just ignore it. Or there might've been a specific instruction to collect several positions at each time point and you have to keep these separate in the analysis, or...
I would ask the clinician in charge.