Things might have improved (I haven't been involved in testing recently), but my experience with syringes is they have a huge amount of variable stiction, the stop start motion, where the syringe sticks for a few minutes and suddenly jumps forward with bolus. It was also worse with some brands compared to others.
The standard says you need to put the performance data in the operation manual, but there is no criteria or requirement to select data that is representative.
As such, it was common practice among manufacturers to wait for a good result and use that one in the operation manual.
Actually this is not just for syringe pumps, I found other examples of infusion pump manufacturers selecting "optimistic" data, and it is common in other medical devices to fudge the data.
The difference is that the variability with syringe pumps can be huge and clinically significant.
For my side, the standard should require enough samples to show for example a trumpet curve that covers 95% of cases. It's an issue raised in the rationale, just not mentioned in the normative text.