This retrospective database analysis demonstrated that a multi-faceted advanced airway management quality improvement program resulted in a significant improvement in first pass success rates of endotracheal intubation in the setting of paramedic-delivered critical care transport. DASH-1A rates improved to a lesser degree, and this change was not statistically significant. In an exploratory analysis, the use of paralytics as well as intubation in a healthcare facility (as opposed to intubation during scene calls) were associated with higher FPS and DASH-1A rates. The use of VL was associated with higher FPS rates but not higher DASH-1A rates.

We believe that the increase in FPS rate from 60 to 86% (for the pre- and post-intervention period, respectively) proves the effectiveness of our organization’s AAM QIP. While the post-intervention results are encouraging, they are also comparable to previously published FPS rates from other prehospital or retrieval services [10, 18,19,20]. Our explorative analysis suggests that the organizational direction from DFI towards RSI and the introduction of VL might have contributed to increasing FPS rates. Paralysis with neuromuscular blockers has been shown to increase FPS rates during intubations in intensive care units and emergency departments, when compared to DFI [21, 22]. To our knowledge, there is no research which directly compares intubation FPS rates with and without paralysis in the prehospital and retrieval setting. A randomized controlled trial comparing intubation with Midazolam and Etomidate, without paralytics in both groups, demonstrated a low overall intubation success rate of only 76% [23]. This contrasts with multiple large case series of prehospital RSIs, where overall intubation success rates trend towards 100% [12, 18, 20]. Prehospital RSI is the recommended strategy for prehospital intubations (other than in cardiac arrest or peri-arrest situations) by the Association of Anaesthetists of Great Britain and Ireland [2].

The best available evidence for the use of VL, and specifically the C-MAC video-laryngoscope used in our organization, is a randomized controlled trial from a German physician-staffed helicopter emergency medical service (HEMS) [24]. FPS rates in this service were 95% and 79% with VL and DL, respectively (p = 0.007). A final factor associated with FPS in our organization was the location in which AAM occurred. FPS rates were significantly lower when intubations were undertaken during scene calls. The majority of these cases were trauma patients and intubation usually occurred in the back of a land ambulance. The lower FPS rate is likely a reflection of the underlying severity and acuity of these cases, limited space around the patient, and the challenges of intubation of trauma patients, when compared to patients with medical critical illness inside a healthcare facility [1, 18].

For our second primary outcome, DASH-1A success rates, our data showed a non-statistically significant increase from 45 to 55% (pre- and post-intervention, respectively). While the use of paralytics and the type of call were associated with higher FPS and DASH-1a rates, the increased FPS rates seen with the use of VL did not translate into higher DASH-1A rates. It is worth noting that most of the interventions of the QIP, particularly the use of a bougie and VL, were aimed at overcoming anatomical challenges of intubation. While this process has led to higher FPS rates over time, it did not directly address some of the physiological challenges of AAM, in particular avoidance of hypoxia and hypotension [25]. This highlights the importance of clinical governance with ongoing monitoring and training for prehospital AAM programs [9, 17]. In our organization, the 2022 annual training program will include simulation specifically focused on physiologically difficult intubations, particularly pre-intubation resuscitation of the severely hypoxic, hypotensive, or acidotic patient [25].

An important challenge that our organization faces in assuring safety and excellence in AAM is the relatively infrequent occurrence of AAM. On average, our paramedics are only involved in little more than one case of AAM per year, with some paramedics only being involved in AAM approximately once every 5 years. While the exact number of intubations or AAM cases required to maintain competence after initial training continues to be debated, many high-performing HEMS provide in excess of 100 prehospital RSIs per base per year [12, 18, 26]. In contrast, almost half of intubations in our dataset were undertaken by paramedics working at bases with less than 10 intubations per year. As part of the AAM QIP, paramedics are now required to participate in a minimum of four to six simulated AAM scenarios per year. To further maximise learning from these infrequent cases, structured debriefs of all DFIs and RSIs as well as regular clinical governance meetings with presentation of AAM cases were introduced to our organization mid-2021.

This before and after study cannot clearly prove cause and effect of the AAM QIP and improved intubation FPS rates. However, the effect size is considerable and in keeping with previous literature. An important limitation is that the multiple aspects of the QIP were introduced in parallel, and due to the relative infrequence of intubations in our service, there is a likely delay between intervention and effect. As such, we were unable to clearly attribute changes in FPS or DASH-1A rate to individual interventions. Importantly, some of the changes made during the QIP, such as standardized equipment, the use of a pre-intubation checklist, or the use of a bougie, was not documented in individual patient charts. We were therefore unable to assess the effectiveness of these interventions individually. Finally, data completion and accuracy are well-documented challenges in retrospective database analyses. We undertook manual reviews of charts were required to minimise this issue, and the data entry and data retrieval process was the same for the pre- and post-intervention period.