There are wide variations in the physical designs and attributes between different brands of intravenous cannulae that makes product selection and purchasing difficult. In a systematic assessment to guide purchasing, we assessed two cannulae - Cannula P and I. We proposed that the results of in-vitro performance testing of the cannulae would be associated with preference after clinical comparison. We designed an observer-blinded randomised head-to-head trial between the 18, 20 and 22 gauge versions of Cannula P and I. Our primary end-point was pressure (mmHg) generated during various flow rates and our secondary end-point was the force (Newton) required to slide the catheter away from the needle. This was followed by a prospective electronic survey following a two-week clinical trial period. The mean difference in resistance between Cannula P and I was: 307 mmHg.L-1.hr-1 (95% CI: 289-325, p < 0.001) for 22G; 135 mmHg.L-1.hr-1 (95% CI: 125-144, p < 0.001) for 20G; and 27 mmHg.L-1.hr-1 (95% CI: 26-28, p < 0.001) for 18G. The mean difference in the force needed to displace the catheter away from its needle was: 1.41 N (95% CI: 1.09-1.73, p < 0.001) for 22G; 0.19 N (95% CI: -0.04-0.41, p = 0.12) for 20G; and 1.96 N (95% CI: 1.40-2.52, p < 0.001) for 18G. After a trial period, all 16 anaesthetist who had used both cannulae preferred Cannula I to P. The evaluation process described here could help hospitals improve efficient product selection and purchasing decisions for intravenous cannulae.