Critical care paramedics: where is the evidence? a systematic review

Objectives Paramedic-delivered prehospital critical care is an established concept in a number of emergency medical services around the world and, more recently, has been introduced to the UK. This review identifies and describes the available evidence relating to paramedics who routinely provide prehospital critical care as primary scene response (critical care paramedics, or CCP). Methods A systematic search of electronic databases was performed: CENTRAL, EMBASE, MEDLINE (through EMBASE and Web of Knowledge) and Web of Science (through Web of Knowledge). Results The search identified 12 relevant publications, one of which was a randomised controlled trial. The remaining 11 were retrospective studies. Five studies compared CCPs with physician-led care. Three of these publications demonstrated improved outcomes with physician care, while two showed no difference. Four further publications examined CCPs versus non-physician-led care and found improved outcomes (two studies), mixed effects (one study) and no difference (one study) for CCPs. Finally, three publications addressed the addition of skills to CCP competencies. A randomised controlled trial of CCP rapid sequence induction (RSI) and tracheal intubation demonstrated improved neurologic outcomes. CCP tube thoracostomy was shown to have similar complication rates to the same procedure performed in the emergency department, while addition of a non-invasive ventilation protocol to CCP practice had no effect on long-term mortality. Conclusions There is limited evidence to support the concept of paramedic-delivered prehospital critical care. The best available evidence suggests a benefit from prehospital RSI carried out by CCPs in patients with severe traumatic brain injury, but the impact of CCPs remains unclear for many conditions. Further high-quality research in this area would be welcome.


INTRODUCTION
The UK has seen a steady increase in emergency ambulance calls over the past decade, with the call volumes approaching 8 million in 2009-2010. 1 Only a small proportion of these incidents require prehospital critical care interventions, such as advanced airway management, cardiopulmonary resuscitation or inotropic support. 2 The average prehospital provider is, therefore, only rarely able to carry out such interventions 3 which are often complex and carry the risk of serious complications. 4 This leaves emergency medical services (EMS) and prehospital providers faced with the challenge of how to provide appropriate yet efficient care for the few severely ill or injured patients in urgent need of prehospital critical care. 5 The merit of advanced or critical care skills in prehospital care has been debated, 6 but there is evidence that extended procedural capacity and decision making benefits many patient groups. 7 8 The question of who should be providing prehospital critical care is an ongoing controversy. 9 10 Concerns regarding the risk of detrimental effects have led to the conclusion that only physicians should undertake certain interventions, such as rapid sequence induction (RSI) of anaesthesia and tracheal intubation. 4 11 12 Others would argue that it is not the professional group of the prehospital provider that determines capability but clinical competency, and therefore, well trained and experienced paramedics should be able to provide an equivalent level of prehospital critical care to physicians. 13 14 Paramedics subspecialising in the delivery of prehospital critical care have become established in North America, mainly as flight paramedics on helicopters, providing interhospital transfers and/or primary scene responses. 15 In parts of Australia, mobile intensive care ambulance (MICA) paramedics are dispatched to patients with suspected major trauma, either by car or helicopter. 14 Similar models exist in South Africa and New Zealand, and efforts to improve prehospital critical care in the UK have resulted in an increasing number of critical care paramedics (UK-CCPs) over recent years. 16 UK-CCPs work either with prehospital doctors, other UK-CCPs, paramedics or on their own (author's unpublished data, 2012). While there is no national training programme or curriculum for UK-CCPs, most UK-CCPs have additional competencies beyond that of UK paramedics, such as procedural sedation, joint or fracture reduction, and invasive interventions such as thoracostomy (authors' unpublished data, 2012). By contrast with some of their colleagues in Australia 14 or North America, 15 UK-CCPs are currently not able to undertake RSI independently. 12 Within the authors' institution (South Western Ambulance Service), UK-CCPs are dispatched to severely ill or injured patients only, either by helicopter or car, and regular participation in clinical governance measures is mandatory. 17 This review identifies and describes the available evidence relating to paramedics who routinely provide prehospital critical care as a primary scene response. For the purpose of clarity, we refer to paramedics acting in this capacity as 'CCPs', since their titles and scope of practice vary widely across different emergency medical systems.

Literature retrieval
We performed a systematic search of electronic databases: CENTRAL, EMBASE, MEDLINE (through EMBASE and Web of Knowledge) and Web of Science (through Web of Knowledge). No language limitation was applied. Due to significant developments in the practice of prehospital care, such as RSI 18 and the use of pulse oximetery 19 in the early 1990s, we decided to exclude papers published before 1990. This restriction was not applied to hand searches of citations in relevant reviews and manuscripts of potential interest. The searches aimed to include the general concept of paramedic-delivered critical care as well as specific critical care competencies. The fact that many CCPs work on helicopters (Helicopter Emergency Medical Services: HEMS) or fixed-wing aircrafts is also reflected in the search terms (see table 1).

Selection of eligible studies
Two independent reviewers ( JvV-F and JW) scanned all titles followed by all abstracts of potentially relevant manuscripts. Each reviewer applied inclusion criteria to the abstract, as outlined in the box below. Full manuscripts were retrieved if inclusion criteria were met, inclusion or exclusion could not be determined with certainty or if the reviewers disagreed. The same process was repeated for the full manuscripts, except reviewer disagreement at this stage was resolved by reference to a senior reviewer ( JB). One reviewer ( JvV-F) then read all eligible publications in detail and discussed the findings with the senior reviewer ( JB). The strength of evidence presented by  20 Data regarding quality of description of prehospital provider competencies, intervention studied, risks of bias and study outcomes were extracted.

Presentation of results
All eligible studies are presented in a comprehensive results table. Due to the anticipated paucity of high-level evidence from randomised trials we planned a narrative analysis.

RESULTS
The search identified 3871 titles of which 609 where potentially relevant. A review of the abstracts identified 122 manuscripts for possible inclusion. This was reduced to 49 after duplications were removed. All 49 manuscripts were retrieved for further assessment.
After review of the full text publications, 12 eligible papers remained (see table 2). Reasons for exclusion of the other 37 publications were investigation of non-paramedic prehospital providers (16/37), or paramedics not working in a critical care capacity (7/37), study designs such as editorials, reviews, case series or descriptive studies without a comparative element (9/37), no reporting of clinical outcomes (4/37) and investigation of interhospital transfers only (1/37). One article did not provide information on the helicopter service studied. 21 As we were unsuccessful in contacting the author for clarification, this article was excluded. No previous review article that addresses this review question was found.

Included publications
There was one randomised, controlled trial (level II evidence). 22 The remaining 11 studies were of level III evidence, 10 of which were retrospective cohort studies [23][24][25][26][27][28][29][30][31][32] and one a quasirandomised cohort study which did not specify whether it was prospective or retrospective. 33 Five studies compared CCP-staffed HEMS with physician-staffed HEMS. 23 28 30 32 33 Outcomes for HEMS CCPs compared to HEMS paramedics with basic competencies were reported in one publication, 31 while one study compared CCP/nurse HEMS with nurse/nurse HEMS. 27 Two further publications compared HEMS paramedics with ground paramedics. 25 29 Finally, three studies examined the effects of CCPs delivering specific interventions, such as tube thoracostomy, 24 non-invasive ventilation 26 and RSI. 22 The most commonly represented countries (5/12) were the USA 23-25 27 33 and Australia (3/12). 22 28 30 One study each originated from Sweden, 26 Canada 29 and Afghanistan. 31 One paper compared data from the USA and Germany. 32

EVIDENCE REVIEW
The articles identified by this systematic search can be arranged into three distinct questions. How does CCP-led care compare to physician-led care? How does CCP-led care compare to other non-physician-led care? What is the effect of adding specific skills to existing CCP competencies?

CCPs versus physician-led care
Five studies addressed this question, of which three 28 32 33 showed a mortality benefit for physician-staffed HEMS and two showed no difference. 23 30 Physicians were at least senior residents, or faculty-level, in four of the studies, 23 28 32 33 the fifth study 30 did not specify the level of training of the emergency physicians. Baxt and Moody 33 found significantly improved mortality for a nurse/physician HEMS crew compared to a nurse/CCP HEMS crew. The paramedic crewmember in the CCP group had less procedural capacity compared to the HEMS nurses and physicians, and interventions were undertaken more often or more aggressively in the physician group. Garner et al 28 compared a dual CCP HEMS crew and a physician/paramedic HEMS crew in Australia and showed improved survival of the physician group over the CCP group. This study also found significantly more interventions delivered in the physician group, including the administration of blood products and neuromuscular-blocking drugs, both of which were outside the paramedic scope of practice. Schmidt et al 32 compared paramedic-staffed HEMS in USA with physician-staffed HEMS in Germany and found significantly more interventions delivered and less early deaths in the German group. Further differences were the absence of penetrating trauma and significantly shorter response times in the one-tier German EMS system. Hamman's before-and-after study 23 included CCPs with at least 2 years of critical care experience and also an unspecified number of nurse/nurse HEMS crew missions in the nonphysician group. This study found significantly improved survival for both groups compared with the Major Trauma Outcome Study (MTOS) population. There was no difference between the physician and non-physician group. Finally, Cameron et al 30 compared staffing and dispatch of an Australian HEMS by either physicians or intensive care paramedics. While there was a significantly higher rate of discharges from the receiving emergency department (ED) during the time when CCPs were responsible for HEMS dispatch, revised trauma score, mortality, length of stay and rate of secondary transfers from the receiving hospital to another facility remained the same.

CCPs versus non-physician-led care
Of the four studies addressing this question, one compared CCP HEMS crews with paramedic HEMS crews and found a significantly better survival rate for CCP HEMS. 31 Two publications compared CCP HEMS crews with ground paramedics, with one showing improvement in mortality 29 in the CCP group and the other reporting mixed results. 25 The last study compared CCP/ nurse HEMS crews and nurse/nurse HEMS crews and found no difference in outcome. 27 Mabry's publication 31 compared two different military HEMS crew configurations: a dual CCP crew Box Inclusion criteria following PICOS acronym P-Patient: All patients, including trauma, medical, paediatric or unselected groups. I-Intervention: Prehospital care or interventions delivered by one or more paramedics working at critical care level. This was defined as targeted dispatch to critically unwell patients (as in many helicopter emergency medical services) and the ability to deliver any of the following interventions: rapid sequence induction, surgical airway, non-invasive ventilation, thoracostomy, thoracotomy, inotropic support, or procedural sedation. C-Comparator: Any other prehospital provider or specific intervention not delivered by critical care paramedics. O-Outcome: Any clinical outcome, including but not limited to early or late mortality and morbidity, length of stay or complications. S-Study design: Any study with a comparative element, such as randomised controlled trial, before-after or case-control designs.  and a single paramedic who was supported by a flight surgeon on an unspecified number of missions. Both groups attended to civilian and military trauma patients, with one crew relieving the other at the end of their rotation. After logistic regression analysis, mortality was significantly less in the CCP group. Mitchell et al 29 compared CCP HEMS and ground transport of trauma patients to a Canadian tertiary centre. This study reported significantly improved survival in the CCP HEMS group which also included a higher number of secondary transfers compared to the ground paramedic group. Kerr et al 25 undertook a large database analysis, comparing CCP HEMS crews to paramedic ground transport. While the direct comparison for patients with Injury Severity Score (ISS) <31 showed a small but statistically significant higher mortality in the HEMS group, this was reversed for patients with ISS ≥31. Differences between the patient groups included a higher percentage of penetrating trauma in the ground group. Wirtz et al 27 compared a CCP/nurse HEMS crew with a nurse/nurse HEMS crew, with nurses and paramedics having critical care competencies. A Trauma and ISS-based analysis showed no difference in mortality between the groups.

Additional critical care skills for CCPs
We identified three studies which examined the effect of specific interventions delivered by CCPs, of which one showed a significant improvement 22 and two showed no difference in their respective outcomes. 24 26 The first study by Bernard et al 22 was a randomised controlled trial of MICA paramedics undertaking prehospital RSI versus basic airway manoeuvres prehospital followed by RSI in the receiving ED. There was no significant difference in the primary outcome with a median Extended Glasgow Outcome Scale (GOSe) score of five and three ( p=0.28) for CCP RSI and ED RSI, respectively. The a priori defined secondary outcome of 'good neurologic recovery' (GOSe 5-8) was achieved significantly more often in the CCP RSI group (51% vs 39%, respectively, p=0.046). York et al 24 compared the complication rates and mortality of trauma patients receiving a tube thoracostomy either by a CCP/nurse, HEMS crew, or by physicians in the receiving hospital after ground transport. Complication rates were equal between the groups, with ISS and unadjusted mortality higher in the HEMS group. Gardtman et al 26 examined the effect of adding a protocol for prehospital heart failure treatment, including noninvasive ventilation, to the competencies of a CCP/nurse team on a mobile coronary care unit (MCCU). While the intervention improved the clinical picture at hospital admission, mortality at discharge and at 1 year were unchanged.

DISCUSSION CCP versus physician-led care
Baxt's study 33 has been frequently cited as a justification for including physicians in HEMS crews. 34 35 The quasirandomisation (not achieved by other studies) and the fact that the nurse crewmember in the CCP and the physician group were capable of the same procedures as the physicians makes this a compelling argument. However, Baxt relates the differences in outcome partially to the fact that the paramedic in the nurse/CCP HEMS crew had significantly less competencies, which excluded thoracostomy, cricothyrotomy or extended medications available only to the nurses and physicians. The other potential factor was the positive effect of physician decision making beyond rigid prehospital protocols. 33 Similarly, the CCPs scope of practice in Garner's study 28 did not include potentially life-saving interventions, such as RSI or blood transfusions, which were received by 42% of patients in the physician group. While recognising these differences, Garner argues that even when paramedics and physicians were able to achieve an intervention such as IV fluid replacement in hypotensive trauma patients, this was done more aggressively by the physicians. However, this comparison is again influenced by the availability of blood products, and is an interesting finding in view of the current trend towards limiting intravenous fluids in trauma patients. 36 Both these studies have been cited in support of physician staffing of helicopters. 35 However, one could argue that they point to better outcomes with increased procedural skills, training and experience rather than the inherent superiority of a particular professional group. This view would be supported by the publications from Hamman et al 23 and Cameron et al, 30 which explicitly state that their HEMS CCPs have significant critical care experience. Despite some differences in procedural capacity between CCPs and physicians (including RSI and tube thoracostomy), both report no differences in mortality. Both studies investigate CCP HEMS crews after removal of a physician from the HEMS crew. This might influence the results in favour of CCPs, as general advances in trauma care, including in-hospital management, might have biased the outcomes. Another possibility is that these CCPs worked alongside prehospital physicians for a significant period before then being dispatched on their own. Physician review and feedback of paramedic practice has been shown to improve paramedic decision making. 37 The third study to show superior outcomes for physicians compared to CCPs is the evaluation by Schmidt et al 32 of an American and German HEMS. One major confounding factor is the absence of penetrating trauma in the German patient cohort, while the cause of unexpected deaths in the USA system was 'mostly the result of penetrating trauma to the head'. Together with differences in dispatch times and the lack of description of USA HEMS paramedic competencies, it is very difficult to attribute effects on outcomes to the HEMS staffing alone.

CCP versus non-physician-led care
The applicability of Mabry's study 31 is limited by its setting within military conflict, however, it is the only publication comparing CCP and paramedic care on the same transport platform. The better survival rate in the CCP group can be explained by a number of factors beyond advanced procedural capability of the CCPs. The most obvious of these is the difference in training and experience. The CCPs had an average of 9 years experience and critical care training, whereas the paramedics might have had as little as 1 year of clinical practice prior to deployment. Additionally, CCPs had an extensive clinical governance system with regular peer review and medical director feedback in place. By contrast, supervision of the paramedic group was the responsibility of flight surgeons, often primary care trained, with 'little or no experience in EMS medical direction, trauma or critical care'. These physicians also attended an unspecified proportion of missions in the paramedic group. It should be noted that the CCP group was always a dual crew, whereas the paramedic group was single-crewed for a large proportion of missions. Mitchell's study 29 also shows that CCPs achieve better outcomes than paramedics, however, several other factors might have influenced this. First, 84% of HEMS missions (and 44% of ground transports) in this study were transfers from other facilities where patients would have received treatment not controlled for in this study. Additionally, the impact of helicopter versus ground transport needs to be considered. The actual effect of speed of transport by helicopter remains debateable, 34 but is likely to be more significant in the rural setting of this study. The actual impact of CCP attendance is, therefore, difficult to measure.
Kerr's comparison 25 of CCP HEMS and paramedic ground transport found an absolute reduction in mortality of 8.2% in trauma patients with ISS ≥31 in the HEMS CCP group. While this is encouraging, the statistically significant mortality increase of 1.0% for patients with ISS <31 in the HEMS CCP group needs to be addressed. The authors do not offer an explanation, and while these dichotomous results could be a statistical anomaly, the large sample size of the study makes this less likely. A possible explanation is that the risks of advanced interventions, when applied indiscriminately in non-critical situations, outweigh their benefit. Dispatch of critical care teams by helicopters has been shown to be of no benefit to non-critical patients, 38 and careful consideration to the appropriateness of all prehospital interventions is mandatory.
Finally, Wirtz et al 27 compared nurse/CCP and nurse/nurse HEMS crews: nurses and CCPs had equal competencies and freedom of decision making. Not surprisingly, no difference in outcome was found between the groups, supporting the argument that comprehensive protocols, training and experience are more important determinants of effective prehospital care than an individual's professional background.

Additional critical care skills for CCPs
The best available evidence identified by this review is a welldesigned, prospective, randomised controlled trial by Bernard et al, 22 demonstrating improved neurologic outcome for patients with traumatic brain injury undergoing prehospital RSI by CCPs in Australia. This is remarkable also for the fact that very few studies have ever been able to show a clear benefit from prehospital RSI by any provider. 39 The results are also in stark contrast to the only other prospective study of paramedic RSI, completed in San Diego. 4 The San Diego study matched patients with traumatic brain injury undergoing RSI with historic controls and found an increased mortality in the paramedic RSI group. The reasons for these contradictory results are likely to be found in training and skill maintenance. In the San Diego study, a large number of paramedics received a 1-day training programme and would, on average, undertake RSI once every 2 years. 4 On the other hand, Australian MICA paramedics receive extensive training in critical care and, due to targeted dispatch, regularly attend severely injured patients. 22 Similarly, York et al 24 showed that a nurse/CCP HEMS crew can carry out prehospital chest tube thoracostomy with complication rates that are not different to those occurring in hospital. Like MICA paramedics, this HEMS team was only dispatched to high-acuity trauma, and placed 76 chest tubes over a period of 2 years. Another intervention studied was a prehospital protocol for heart failure, including non-invasive ventilation, carried out by a nurse/ paramedic team on a MCCU in Sweden. 26 The team successfully applied non-invasive ventilation to 91% of patients, but the authors could not demonstrate any benefits in long-term outcomes; a finding which is consistent with other studies. 40 This likely reflects the underlying disease process, which will be difficult to affect within the short time frame of prehospital care.

Strengths and weaknesses
This is the first review to look at the evidence supporting the development of CCPs. While the quality of the evidence identified is variable, it allows an assessment of the anticipated effectiveness of this new professional group, alongside likely conditions for success. However, we may have overlooked some relevant evidence, particularly that published in a non-English language and less accessible sources. There may also be an element of publication bias, with a reluctance to publicise unexpected or adverse outcomes. The nature of the source data makes it impossible to pool individual studies or undertake detailed numerical analysis, so the findings of this review are necessarily narrative in form. Furthermore, CCP and prehospital physician training, practice and skill maintenance varies between EMSs, and this needs to be considered when applying the results of this review.

CONCLUSION
CCPs are a group of paramedics with critical care skills who are dispatched to severely ill and injured patients. As is the case for many aspects of prehospital care, there is currently only limited evidence to support this model. The best available evidence suggests a benefit from prehospital RSI carried out by CCPs in patients with severe traumatic brain injury, 22 but the impact of CCPs on long-term outcomes remains unclear for many conditions, and further high-quality research in this area would be welcome.
The evidence reviewed indicates that CCPs are able to deliver care to critically ill and injured patients that is superior to care delivered by paramedics and nurses without additional training and competencies. Whether CCPs can achieve the same standards as doctors trained in prehospital medicine remains unclear, but seems possible under certain conditions. High-quality training in procedures, up-to-date protocols and access to the relevant critical care equipment and medications are essential for optimal procedural capability. Regular exposure to critically ill and injured patients is vital to maintain skills and decision making after initial training. Finally, a robust clinical governance system with feedback, clinical review and medical oversight will identify and address problems and strengthen clinical and decision-making skills. EMSs in which these conditions are provided, have demonstrated encouraging results for CCPs. 17 22 31 Any EMS considering the introduction of critical care skills, be it through paramedics, physicians, or nurses, should consider whether it can deliver this level of support to its prehospital providers. 13