Nurses versus physician-led interhospital critical care transport: a randomized non-inferiority trial
Jan van Lieshout. Intensive Care Medicine 2016; 42: 1146–1154. DOI: 10.1007/s00134-016-4355-y
- In critically ill patients requiring interhospital ground transport, is a nurse-led transfer team non-inferior to a physician-led team, for the proportion of patients with critical events?
- Non-inferiority study
- Prospective, randomised, open-label, blinded-endpoint trial (PROBE design)
- Web-based randomisation using permuted blocks of variable size, stratified by inter-hospital distance
- Blinding of referring hospital during pre-transport stabilisation period
- Blinded assessment of electronic health record to determine adverse clinical events
- Consecutive patients
- 274 patient required to power the study at 80% to detect a non-inferiority margin of 0.03
- One-sided Z test used with a significance level targeted at 0.05
- Non-inferiority was assumed if the upper limit of the two-sided 90% confidence interval for between-group difference was below the non-inferiority margin of 3%.
- Single Academic Medical Centre in the Netherlands
- Data collected between January 2006 and March 2008.
- Inclusion criteria: Mechanically ventilated adult ICU patients to be transported by a mobile critical care unit
- Exclusion criteria: Severe hypoxia (p/F ratio < 100 with > 15 cmH20 PEEP); haemodynamic instability (MAP < 60mmHg despite vasopressor support and IV fluid resuscitation), post cardiac arrest, requiring immediate transport
- 618 patients assessed for eligibility, with 298 patients analysed after dropouts due to exclusion, insufficient data, emergency transport or transport cancellation.
- Both groups were well-matched at baseline. Comparing intervention vs control group:
- APACHE II score: 19 vs. 18
- Median transport time: 66min vs. 65min
- Nurse led inter-hospital transport (n=147)
- A dedicated transport team staffed by a critical care nurse (with at least two years ICU experience) accompanied by a paramedic (with at least 5 years pre-hospital experience and licensed in advanced cardiac life support including rescue endotracheal intubation)
- A critical care physician accompanied every transport for safety reasons but was not physically present in the patient compartment of the ambulance.
- 12 patients crossed over to the control arm post randomisation due to clinical requirement. This was prior to leaving the referring hospital.
- Physician led inter-hospital transport (n=151)
- The transport team included a critical care physician (European board-certified intensivist/anaesthesiologist, or a senior resident in anaesthesiology with at least six months ICU experience) in addition to a nurse and paramedic.
For both intervention and control groups:
- Physician and nursing staff from single Academic Medical Centre
- Both nurses and clinicians completed the extended critical care transport training of the Academic Medical Centre
- The equipment and transport ambulance used was identical.
- Primary outcome: The number of patients with critical events, which included both clinical and technical events during transport
- Clinical events were defined as a reduction in O2 sats of > 10% for > 10 minutes, a rise or fall in blood pressure of > 20mmHg for > 10 minutes, a temperature fall below 36 degrees, blood loss > 250 mls, arrhythmias with hypotension or death
- Technical events, defined as related to technical aspects of critical care transport, included ECG lead disconnections, loss of battery power or any other technical equipment failure, airway loss requiring airway manipulation or re-intubation, loss of any intravascular device, or dislodgment of any chest tube, Foley catheter, or surgical drain
- Non-inferiority of the intervention was not established
- Comparing intervention with control group there was 16.3% vs. 15.2% critical events (difference 1.1%, 90% CI [-5.9 to 8.1]; p= 0.38)
- Fragility index: -11 patients
- No haemorrhages, arrhythmias or deaths were reported in either group
- Secondary outcomes – Comparing intervention vs. control group
- Clinical incidents – no significant difference
- 13.6% vs. 14.6% (difference -1%, 90% CI [-5.2 to 8.7]; p=0.44)
- Technical incidents – no significant difference
- 2.7% vs. 0.7% (difference 2.1%, 90% CI [0.7 to 5.3]; p= 0.35)
- No difference in ventilator adjustments or changes to sedative/ vasoactive medication
- A trend towards increased fluid therapy > 1000mls in the physician group
- 5.4% vs. 11.3% (difference -5.9% 95% C.I. [-12.5 to 0.5]
- Clinical incidents – no significant difference
- This study could not establish the non-inferiority of nurse-led inter-hospital critical care transport. The number of patients with critical events during transport did not differ substantially between the physician and nurse-led groups but a larger study would be needed to definitively show this.
- Appropriate study design – randomised, controlled, non-inferiority study
- A power calculation was performed to detect a difference in critical event rate between the two group of less than 3%
- The web-based permuted block randomisation, with variable block size, ensured allocation concealment
- Randomisation was stratified to ensure similar transport times in both study groups
- Groups were well-matched at baseline
- Mean APACHE II scores in the intervention and control groups was 19 and 18 respectively, so the patients recruited were a reasonably sick cohort.
- Outcome parameters were assessed using anonymised electronic medical records to ensure blinding of the allocation group
- An electronic health record was used to record clinical data on a minute-by-minute basis, reducing clinician reporting bias.
- Single centre study, limiting external validity
- The sickest cohort of ICU patients was excluded. These were patients with severe hypoxia and haemodynamic instability and patients requiring immediate transport
- Insufficiently powered to detect non-inferiority due to a higher than anticipated critical event rate
- The primary outcome was the occurrence of critical events that were based on “expert opinion”. These were all surrogate outcomes that may not necessarily have had clinical relevance. Clinical events ranged from relatively benign outcomes such as a rise in systolic blood pressure of 20mmHg to sustained hypoxia. No breakdown was given of the clinical events that occurred in each group
- A physician was present in the ambulance vehicle in the intervention arm, which may have influenced the behaviour of the nurse/ paramedic transfer team
- The higher than anticipated event rate may have been due to the increased sensitivity of electronic health record analysis but may also have been due to suboptimal patient care during transport
- The response to abnormal clinical events, rather than their occurrence, may more accurately reflect the competence of the transferring team.
The Bottom Line
- Stable ICU patients without severe hypoxia or haemodynamic instability may have similar rates of critical events when transferred by a nurse/paramedic team compared with a physician led team, but an adequately powered study is required to establish this.
- In the hospital I currently work in, physicians are part of the team performing inter-hospital transfers of critical care patients. This paper does not provide enough evidence to change this practice.
- [article abstract] Nurses versus physician-led interhospital critical care transport: a randomized non-inferiority trial
- [article] Free full text
- [further reading] ESICM Journal Club Review
- [further listening] Scott Aberegg on non-inferiority trials