Skip to main content
. 2014 Nov 6;2014(11):CD006904. doi: 10.1002/14651858.CD006904.pub3

Kollef 1997.

Methods Randomized controlled trial
Participants Setting: USA, 2 medical and 2 surgical ICUs in 2 university affiliated teaching hospitals (900 and 450‐beds). Nurse to patient ratio 1:2 and 4.0 physician hours/bed/day (Krishnan 2004)
Participants: 357 adults (intervention 179, control 178)
Conditions: postoperative; trauma; pneumonia; COPD/asthma; pulmonary oedema; respiratory failure; drug overdose; cardiac arrest/cardiogenic shock
Inclusion: mechanically ventilated. Exclusions: head/facial burns or trauma; transfer from other hospital with prior MV; brain death
Interventions Intervention: protocol entry criteria assessed, then protocol delivered by RNs and RTs consisting of:
a) ICUs 1 and 4 ‐ daily SBTs through ventilator circuit with CPAP ≤ 5 cmH2O and PS ≤ 6 cmH20 for 30‐60 minutes then extubation
b) ICU 2 ‐ stepwise reductions of 2 cmH20 in PSV until 6 cmH20 then extubation
c) ICU 3 ‐ on PEEP ≤ 5 cmH20, PS ≤ 6 cmH20, stepwise IMV reductions of 2 breaths/min until ≤4 breaths/min, then 0 breaths for 30‐60 minutes, then extubation
Control: usual practice consisting of weaning according to physician judgement
Outcomes 1. Total duration of mechanical ventilation from intubation until discontinuation of MV
2. Reintubation
3. Length of hospital stay
4. Hospital mortality rate
5. Hospital costs
6. MV time prior to weaning
7. Requiring MV for > 7 days
Notes Protocol registration not reported. Sample size calculation based on 80% power to detect a difference in weaning time of 1 (SD 3) days, α 0.05, 145 patients needed per group. Study approved by University Human Studies Committee and hospital Institutional Review Board ‐ both waived requirement for informed consent
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Seperate blocked randomization schedules
Allocation concealment (selection bias) Low risk Opaque sealed envelopes
Blinding of participants and personnel (performance bias) 
 All outcomes High risk Impossible to blind personnel to the intervention groups
Blinding of outcome assessment (detection bias) 
 All outcomes Low risk Outcome assessors were independent from the individuals administering/supervising the intervention
Incomplete outcome data (attrition bias) 
 All outcomes Low risk No missing outcome data.Recruitment & attrition data presented. Analyses performed using ITT principle
Selective reporting (reporting bias) Low risk Weaning protocol is available; all prespecified outcomes reported
Other bias Low risk Appears to be free of other sources of bias. Sample size calculation stated (based on 80% power to detect a 1 day difference in weaning time, α 0.05, 145 required for each group)