Summary of findings
This study is the first to describe a national cohort of fatal and non-fatal drowning patients treated by the Danish Air Ambulance. The population had high prehospital and 30-day mortality rates of 33% and 67%, respectively, and a high incidence of prehospital critical care interventions, particularly endotracheal intubation, use of an automated chest compression device, intraosseous cannulation, and ultrasound examination.
Results compared to the existing literature
The available literature on a nationwide population of drowning patients treated by a physician-staffed HEMS is scarce. The available body of evidence focusing on physician-staffed HEMS mainly constitutes data on traumatic injuries, paediatric emergencies, cardiac arrests, or winch operations [
7,
22‐
26].
The drowning population treated by the Danish Air Ambulance had a median age of 50 years with a bimodal pattern concentrating on ages one to six and beyond 60 years. For the paediatric subpopulation below 6 years, the 30-day survival was 75% compared to only 22.7% in the elderly individuals above 65 years. The improved survival rates in children are consistent with other studies of paediatric emergencies [
7,
27,
28], and drowning-related OHCA in children may be correlated with a higher rate of bystander first aid and CPR [
29,
30]. Also, one study indicated that paediatric drowning events were more likely to occur in swimming pools [
7], which is associated with improved survival [
12]. Nevertheless, drowning remains a frequent cause of HEMS dispatch in children and a leading cause of death in children [
22,
31].
Comparisons between patient populations assessed by a physician-staffed HEMS and ground-based EMS may be inappropriate due to case-mix differences in illness severity and diagnoses, dispatch criteria, and skills of the medical staff [
22].
Proper dispatch of HEMS is essential to reduce over-triage and costs. Patients transported by ambulance without HEMS physician escort with NACA scores less than three have previously been considered to indicate HEMS over-triage [
10]. On the contrary, a high NACA score and the need for critical care interventions may indicate the relevance of HEMS dispatch [
32]. Drowning patients treated by the Danish Air Ambulance were critically ill with high NACA scores more than three and a high incidence of endotracheal intubation and other critical care interventions. As previously demonstrated, patients were predominantly assessed in decentralised areas where helicopter dispatch most likely decreased the transport time compared to ground-based EMS [
33,
34]. Yet, the median response time in this study was 21 min, emphasising the critical role of bystanders and trained lifeguards to assist with early rescue and lifesaving treatment [
35,
36]. The results from this study indicate proper dispatch of the Danish Air Ambulance for the subpopulation of drowning patients with overall high NACA scores, decentralised geographical locations with prolonged transport times, and a high incidence of prehospital interventions.
As a surrogate measure of severity, we used the number of critical care interventions performed by the physician-staffed Danish Air Ambulance. More than two-thirds of the missions conducted by the Danish Air Ambulance resulted in at least one critical care intervention, particularly endotracheal intubation, use of an automated chest compression device, intraosseous cannulation, and ultrasound examination. There was a very high intervention rate in the fatal group, with higher NACA scores indicating increased illness severity, which is consistent with previous findings [
7]. The most frequently performed critical care intervention was endotracheal intubation, which is reasonable from a pathophysiological perspective and in line with previous studies on pediatric drowning incidents [
7]. Studies show that intubation is a feasible intervention following a submersion incident despite the lack of data to support a specific airway management strategy [
5,
37]. Without data supporting an alternative strategy, adopting the Advanced Life Support Task Force recommendations to restore ventilation and circulation seems reasonable [
38]. This study found an association between 30-day mortality and intraosseous cannulation and the use of an automated chest compression device, suggesting confounding by indication through disease severity bias, as patients with severe drowning may be more likely to receive these interventions, as observed in other studies [
39]. Our results could be interpreted likewise.
According to our data, HEMS physicians need to be confident with the prehospital care of critically ill drowning patients, including airway management, intraosseous cannulation, use of an automated chest compression device, and ultrasound. Even though children accounted for a minority of the drowning population treated by the nationwide Danish Air Ambulance, this study supports the findings from Germany that HEMS physicians need to be confident with the prehospital care of drowned children of all ages [
22].
Methods compared to other studies
This study is the first to confirm using the Danish Drowning Formula developed by Breindahl et al.[
12] to identify a nationwide cohort of drowning patients from the Danish Air Ambulance database. As previously speculated, this novel method of drowning identification shows a promising potential to improve the quality of future drowning research by searching high-quality databases linked to each patient’s unique civil registration number [
12]. The two current studies involving the Danish Drowning Formula report on two distinct subpopulations of fatal and non-fatal drowning (those with OHCA and those treated by the Danish Air Ambulance, respectively) and can hardly be compared. More studies are needed to identify all fatal and non-fatal drowning incidents in Denmark.
Strengths
The Danish Air Ambulance database used in this study is well-described. It provides high data quality in uniform data reporting and a high degree of data completeness (95.8–99.9% for variables concerning injury severity, prehospital diagnostics, and critical care interventions) [
6,
10,
19]. Nationwide data inclusion from the Danish Air Ambulance provides a complete picture of the Danish subpopulation of drowning patients treated by a physician-staffed HEMS. This may allow for subgroup analyses and comparisons with other studies.
The positive predictive value of drowning incidents was 71% when the Danish Drowning Formula was applied to the missions with relevant primary or secondary diagnosis codes in the Danish Air Ambulance database. This adds to the evidence that text-search algorithms (e.g., Danish Drowning Formula) may provide a cost-effective solution towards future drowning identification. It was impossible to calculate the sensitivity, specificity, and negative predictive value of the Danish Drowning Formula in this study, as we did not validate the missions excluded by it and therefore have no information on the true prevalence. However, studies are ongoing to clarify these measures on a larger, unselected cohort and optimise the Danish Drowning Formula to increase its applicability for drowning identification in the future.
Limitations
This study has some limitations. We applied the text-search algorithm within the relevant diagnosis codes in the Danish Air Ambulance database. Therefore, some drowning patients registered with other diagnosis codes may have been lost in this process. Furthermore, drowning was only registered in 45% and 23% of primary and secondary diagnosis codes, respectively, resulting in 26 patients (32%) not having a primary or secondary diagnosis code of drowning. This adds to the evidence that the International Classification of Diseases 10th Edition’s drowning codes are inappropriate for drowning research as they are unreliable [
40‐
42].
Various limitations are linked to the retrospective design of this study. First, analyses on the effect of critical care interventions are complex due to confounding by indication, meaning the most critically ill patients may require critical care interventions more frequently. Second, despite limited missing data in the Danish Air Ambulance database, missing data may introduce a selection bias when providing lifesaving treatment and fast transport for critically ill patients. Third, drowning patients were identified based on the prehospital data in the Danish Air Ambulance database. We cannot determine the circumstances relating to the drowning event or the actual causes of death. Fourth, there were 12,360 missions without patient encounters for various reasons, including bad weather conditions in 3,015 missions. Bad weather conditions may increase the risk of drowning in natural environments [
43], and these patients may have been transported by ground-based EMS or the Royal Danish Air Force’s Search And Rescue helicopters [
9]. The latter are used for missions at sea and were not included in this study as they are reported in a separate system [
9].