Table 1.

Objective tasks completion rate

Device	Power on (%)	Start ventilation (%)	Inspiratory flow setting (%)	Ventilatory mode modification (%)	Cycling setting (%)	NIV mode activation (%)
Carefusion Avea	92	100	100	100	75	75
Dräger V500	100	92	100	100	92	50
Covidien PB980	100	100	100	91	100	100
Philips V680	100	100	50§	91	91	40†
Hamilton S1	91	91	18§	82	91	91
GE R860	100	100	100	91	100	91
Maquet Servo-U	36*	100	0¤	100	45‡	27†

Device	Ventilator offset (%)	Alarm shut down (%)	Mode recognition (%)	Humidification system recognition (%)	Settings reading (%)	Overall
Carefusion Avea	100	100	100	75	42	87%/85%
Dräger V500	100	75	100	42	25	80%/75%
Covidien PB980	73*	64	100	73	45	86%/86%
Philips V680	100	82	100	55	18	75%/70%
Hamilton S1	100	91	100	36	27	74%/70%
GE R860	100	64	100	9‖	27	80%/76%
Maquet Servo-U	100	82	100	0‖	45	58%#/55%‖

Results are presented as the different objective tasks success rate, expressed as the percentage of successful attempts. ICU’s physicians with mechanical ventilation’s knowledge had to complete 11 specific tasks of variable clinical importance for each ventilator, 4 mainly dedicated to monitoring and 7 to setting. Overall provides results of the entire bunch of test (first value), or after powering on/switching off tasks exclusion (second value)

Among all ventilators, our reference device the Avea had the better success rate. In between our comparison of the six ventilators, Covidien PB980 had the better results and the Servo-U the worst. A minority of users could power on Servo-U and always over the predefined 120 s time range. Settings reading represented the more difficult task, whatever the device, while their terminology was highly different from one device to the other

* p < 0.01 as compared to others. ^§ p < 0.001 as compared to Avea, V500 and PB980. ^¤ p < 0.01 as compared to V680. ^‡ p < 0.05 as compared to others except Avea. ^† p < 0.05 as compared to PB980 S1 and R860. ^‖ p < 0.001 as compared to Avea and PB980. ^# p < 0.01 as compared to others, except S1 and V680