Does prehospital time affect survival of major trauma patients where there is no prehospital care?

BACKGROUND: Survival after major trauma is considered to be time dependent. Efficient prehospital care with rapid transport is the norm in developed countries, which is not available in many lower middle and low-income countries. The aim of this study was to assess the effect of prehospital time and primary treatment given on survival of major trauma patients in a setting without prehospital care.
MATERIALS AND METHODS: This prospective observational study was carried out in a university hospital in Mumbai, from January to December 2014. The hospital has a trauma service but no organized prehospital care or defined interhospital transfer protocols. All patients with life- and/or limb-threatening injuries were included in the study. Injury time and arrival time were noted and the interval was defined as "prehospital time" for the directly arriving patients and as "time to tertiary care" for those transferred. Primary outcome measure was in-hospital death (or discharge).
RESULTS: Of 1181 patients, 352 were admitted directly from the trauma scene and 829 were transferred from other hospitals. In-hospital mortality was associated with age, mechanism and mode of injury, shock, Glasgow Coma Score <9, Injury Severity Score ≥16, need for intubation, and ventilatory support on arrival; but neither with prehospital time nor with time to tertiary care. Transferred patients had a significantly higher mortality (odds ratio = 1.869, 95% confidence interval = 1.233-2.561, P = 0.005) despite fewer patients with severe injury. Two hundred and ninety-four (35%) of these needed airway intervention while 108 (13%) needed chest tube insertion on arrival to the trauma unit suggesting inadequate care at primary facility.
CONCLUSION: Mortality is not associated with prehospital time but with transfers from primary care; probably due to deficient care. To improve survival after major trauma, enhancement of resources for resuscitation and capacity building of on-duty doctors in primary centers should be a priority in countries with limited resources.

Introduction

Trauma is considered to be a time-dependent disease and early initiation of definitive care during the “golden hour” after injury is considered vital to survival of trauma patients. R. Adams Cowley is believed to have coined the term “golden hour.”[12] Donald Trunkey proposed the concept of trauma system with prehospital care as its integral part.[3] Today, postcrash rescue operations, rapid evacuation by ground, or/and air ambulances equipped with paramedics trained in advanced prehospital life support to an appropriate hospital and tertiary care trauma centers have become the standard of trauma care in the developed countries.[2] In the opinion of Rogers and Rittenhouse, the concept of the golden hour has generated “a billion dollar industry” despite the lack of adequate evidence in its support.[2]

However, such an infrastructure is still lacking in low-middle income countries (LMICs) and low-income countries (LICs). Significant time delay to medical care has been reported in LMICs and LICs as compared to high-income countries (HICs).[45] Higher mortality has been reported in LICs and LMICs as compared to HICs and high middle-income countries.[45] Establishment of prehospital care in developing countries has been shown to reduce mortality.[6]

Most literature published on this subject is from HICs with well-established prehospital care. There is paucity of data assessing impact of prehospital time and transfers to tertiary care hospital in trauma patients in places without any prehospital care. A study from rural Maharashtra, which included all admitted trauma patients (major and minor), has reported higher mortality with longer prehospital delay.[7] The present study was planned in a major metropolitan university hospital in Mumbai, to test if prehospital time lag affects survival of major trauma patients and to study if there is any difference in survival between patients who are directly admitted to tertiary care hospital in comparison to those transferred after primary treatment in other hospitals.

Materials and Methods

The study was designed as a prospective observational study from January 2014 to December 2014 after obtaining the Ethics Committee approval with the basic aim of developing a trauma registry with a minimum data set from which the relevant data were obtained. The data were collected prospectively by a specially appointed data collector. This being purely an observational study, consent waiver was obtained from the Institutional Ethics Committee.

The study setting was a 1500-bedded tertiary care Medical College Hospital in Mumbai, with a trauma service with 14 beds exclusively for major trauma patients (patients with life- or limb-threatening injuries). There was no organized prehospital care or interhospital transfer protocols. Patients were received both directly from the scene of the incident and from other hospitals. Trauma service worked under the leadership of the Department of General Surgery and was manned round the clock by surgery and anesthesia residents who initiated resuscitation of all patients as per ATLS® principles. Definitive care was planned in consultation with the specialties (orthopedic, neurosurgery, etc.) who were available on call as required.

All patients with life- or limb-threatening injury admitted in this trauma unit were included in this study. Patients with incomplete data set were excluded from the study. Sample size was not calculated and convenience sampling method was followed.

Injury-time and time of arrival in the tertiary hospital were noted and the intervening period was defined as “prehospital time” for those who arrived directly in the hospital from the scene and as “time to tertiary care” for those who were transferred from some other facility. Variables recorded were mode of transport, type of admission (direct or transfer), age, gender, mechanism of injury (blunt or penetrating), mode of injury (railway, road traffic accident, fall, and others [assault, industrial, or unknown history]), time of injury, time of arrival, physiological parameters on arrival (systolic blood pressure [SBP] and Glasgow Coma Score [GCS]), anatomical severity of injury (Injury Severity Score [ISS]). ISS was calculated by summing the square of the three highest abbreviated injury scores for injuries to different body regions. Primary outcome measure was in-hospital death or discharge. Secondary outcome measures were resuscitative interventions within 1st h of arrival (intubation, chest tube insertion, and transfusion of blood and blood products) and need for ventilatory support and operative treatment within 24 h of hospital admission.

Descriptive statistics were used to describe the study sample, including mean, standard deviation (SD), median and interquartile range (IQR), total range, and percentages (rounded to the nearest whole number). As the continuous data had skewed distribution or extreme values, it was categorized into subgroups for the purpose of analysis. Age was grouped as pediatric (0–12 years), adults (13–60 years), and geriatric patients (above 60 years) as per the administrative definition followed in the hospital. SBP was dichotomized at 90 mmHg (SBP <90 mmHg being the standard definition of shock in normotensive adults). ISS was dichotomized at 16 (ISS ≥16 being the standard definition of severe injury). Injury-arrival time lag was grouped into four groups of <2, 2–4, 4–8, and >8 h which roughly corresponded to the quartiles. Bivariate analysis was conducted to determine the association of variables with mortality using Chi-square test. Multivariate logistic regression analysis was used to determine if prehospital time correlated with in-hospital mortality in directly admitted patients and whether time to tertiary care was a predictor of mortality among the cohort of transferred patients and all patients. Injury-arrival interval was used as a continuous variable for multivariate logistic regression analysis. A P < 0.05 was considered statistically significant.

Results

One thousand two hundred and eighty-six records of patients admitted to the trauma unit from January 2014 to December 2014 were available. Of those patients, 1209 had complete time data. Further, 28 records had to be excluded for not containing key variables (i.e., SBP, GCS score, or ISS score). Thus, a total of 1181 records were included for the analysis. Three hundred and fifty-two (30%) patient were directly admitted from the site of the incident whereas 829 were transferred from other hospitals for further care.([Figure 1]. While most (91%) transferred patients came by ambulances, 96% of direct patients came either by hiring taxis or were brought in by the police vehicle. Information regarding route taken, traffic delays and reasons for transfer were not captured.

Figure 1

Flow chart showing patients included in the study

Of 1181 patients, 87%(1030) were males. Age ranged from 1-90 years (mean of 29 years, SD 17, median 26, IQR 19). Injury-arrival time interval ranged from 0.2-72 hours (mean 7.8 hours, SD 4, median 4.25, IQR 6.8). Ninety percent of admitted patients were those with blunt trauma, the rest having predominantly low velocity penetrating trauma. Modes of injury were road traffic accident (30%), fall (30%), railway injuries (22%), and others (18%). Twenty-eight percent patients were hypotensive (SBP <90 mmHg) while patients with severe head injury (GCS <9) comprised 25%, moderate head injuries (GCS 9–12) were 14%, and the rest with mild head injuries. ISS score of all patients ranged from 1 to 59 (mean 12.5, SD 7.5, median of 14, IQR 8). In all 38% patients needed intubation within 1 h of arrival to the hospital including those intubated before arrival. Nearly 15% required chest tube insertion within an hour of arrival to the hospital. Blood transfusion could be received only by 24 (2%) patients within 1 h of arrival. In all 30% needed operative interventions while 35% required ventilatory support. Overall mortality was 24%.

Table 1 shows bivariate analysis showing association of the variables with mortality independently. Factors significantly associated with mortality were age (>60 years), mechanism of trauma (blunt), mode of injury (railway), shock (SBP <90 mmHg), GCS <9, ISS ≥16, intubation on arrival, and the need for ventilatory support on arrival. While overall time to tertiary care did not have significant impact on mortality, patients directly admitted to tertiary care hospital had significantly better survival on bivariate analysis.

Table 1

Factors correlating with mortality after major trauma - bivariate analysis

Variables	Subgroup	Total	Deaths	Mortality (%)	P
n		1181	288	24
Age	0-12	206	41	20	<0.001*
	1-60	896	200	22
	>60	79	47	59
Mechanism	Blunt	1064	276	26	<0.001*
	Penetrating	117	12	6
Mode	Railway	268	98	32	<0.001*
	Road traffic	360	105	29
	Fall	359	70	19
	Others	194	15	8
SBP (mmHg)	<90	321	96	30	0.007*
	≥90	860	192	53
GCS	<9	299	184	62	<0.001*
	9-12	161	41	25
	13-15	721	63	9
ISS	<16	899	194	22	<0.001*
	≥16	282	94	33
Intubation within 1 h of arrival (including intubation before arrival)	Yes	448	225	58	<0.001*
	No	733	63	9
Chest tube within 1 h of arrival (including chest tube before arrival)	Yes	172	52	30	NS
	No	1009	236	23
Ventilatory support	Yes	414	245	59	0.007*
	No	767	43	6
Treatment	Conservative	834	194	23	NS
	Operative	347	94	27
Time to tertiary care (h)	<2	312	85	27	NS
	2-4	261	63	24
	4-8	269	65	24
	>8	339	76	22.0
Primary care	None (direct admission)	352	61	17	0.0002*
	Some (transferred)	829	227	27

*Significant. NS: Not significant, SBP: Systolic blood pressure, GCS: Glasgow Coma Score, ISS: Injury Severity Score

Three hundred and fifty-two patients were admitted directly from the scene. The prehospital time ranged from 0.2 to 9.5 h (mean 2.13 h, SD 3.04, median 1.52, IQR 2.08). To assess the effect of prehospital time lag on survival, multivariate logistic regression analysis was carried out of the cohort of directly admitted patients; which is presented in Table 2 and does not reveal any association of mortality with prehospital delay (odds ratio [OR] =1.138, confidence interval [CI] = 0.273–1.113, P = 0.573).

Table 2

Multiple logistic regression analysis for in-hospital mortality of patients directly admitted to tertiary care

Variable	SE	OR (95% CI)	P
Gender:Male	0.211	1.019 (0.301-1.211)	0.901
Age
13-60	Reference
≤12	0.223	0.654 (0.076-1.765)	0.056
>60	0.156	5.532 (3.10-6.531)	<0.001*
Mechanism: Penetrating	0.573	0.017 (0.009-0.072)	0.000*
Mode of injury
Fall	Reference
Railway	0.132	2.424 (1.731-3.121)	0.001*
Road traffic	0.242	1.111 (0.128-1.78)	0.218
Others	0.346	0.765 (0.346-1.21)	0.346
SBP >90 mmHg	0.216	0.482 (0.026-0.654)	0.006*
GCS
13-15	Reference
<9	0.295	2.180 (1.20-2.341)	0.001*
9-12	0.417	0.679 (0.126-1.132)	0.176
ISS ≥16	0.245	1.180 (0.232-1.254)	0.428
Intubation within 1 h of arrival	0.237	2.384 (2.10-3.76)	0.001*
Chest tube within 1 h of arrival	0.216	1.552 (0.058-1.23)	0.088
Ventilatory support	0.122	8.006 (6.34-10.561)	0.000*
Operative intervention	0.212	0.976 (0.921-1.32)	0.911
Prehospital time	0.001	1.138 (0.273-1.113)	0.573

*Significant. CI: Confidence interval, SE: Standard error, OR: Odds ratio, SBP: Systolic blood pressure, GCS: Glasgow Coma Score, ISS: Injury Severity Score

Eight hundred and twenty-nine patients were transferred from other hospitals. The time to tertiary care ranged from 1 to 72 h (mean 7.83 h, SD 4.17, median 5.10, IQR 4.9). Analysis of the cohort of transferred patients also revealed that there was no association of time to tertiary care hospital and mortality (OR = 0.977, CI = 0.371–1.214, P = 0.917) [Table 3].

Table 3

Multiple logistic regression analysis for in-hospital mortality of patients transferred to tertiary care

Variable	SE	OR (95% CI)	P
Gender: Male	0.231	1.034 (0.319-1.46)	0.871
Age (years)
13-60	Reference
≤12	0.293	0.612 (0.136-1.345)	0.059
>60	0.153	8.434 (6.10-11.561)	<0.001*
Mechanism: Penetrating	0.512	0.028 (0.009-0.412)	0.000*
Mode of injury
Fall	Reference
Railway	0.433	1.145 (1.031-3.121)	0.578
Road traffic	0.249	0.938 (0.118-1.218)	0.860
Others	0.376	1.138 (0.566-1.311)	0.631
SBP >90 mmHg	0.259	0.039 (0.064-0.414)	0.029*
GCS
13-15	Reference
<9	0.325	2.501 (1.267-2.951)	0.001*
9-12	0.498	0.515 (0.326-1.672)	0.217
ISS ≥16	0.205	2.118 (1.932-3.356)	0.003*
Intubation within 1 h of arrival	0.250	2.374 (2.45-4.32)	0.001*
Chest tube within 1 h of arrival	0.336	1.536 (0.218-1.73)	0.094
Ventilatory support	0.236	7.915 (5.32-12.341)	0.000*
Operative intervention	0.292	0.977 (0.921-1.231)	0.917
Time to tertiary care	0.211	0.977 (0.371-1.214)	0.917

*Significant. CI: Confidence interval, SE: Standard error, OR: Odds ratio, SBP: Systolic blood pressure, GCS: Glasgow Coma Score, ISS: Injury Severity Score

The result was similar (OR = 1.00, CI = 0.453–1.613, P = 0.473) when the data including all the patients – both the directly admitted and transferred – were analyzed using multivariate logistic regression analysis [Table 4]. Transferred patients, however, had nearly two times higher risk of mortality as compared to the patients who were directly admitted (OR = 1.869, CI = 1.233–2. 561, P = 0.005).

Table 4

Multiple logistic regression analysis for in-hospital mortality of all patients admitted to tertiary care

Variable	SE	OR (95% CI)	P
Gender: Male	0.291	1.029 (0.712-1.701)	0.921
Age
13-60	Reference
≤12	0.210	0.792 (0.086-1.77)	0.067
>60	0.180	7.544 (4.60-9.531)	<0.001*
Mechanism: Penetrating	0.533	0.028 (0.009-0.562)	0.000*
Mode of injury
Fall	Reference
Railway	0.264	1.163 (0.021-1.221)	0.573
Road traffic	0.242	1.169 (0.228-1.908)	0.518
Others	0.316	0.942 (0.236-1.71)	0.866
SBP >90 mmHg	0.210	0.032 (0.011-0.87)	0.027*
GCS
13-15	Reference
<9	0.290	2.29 (1.80-2.871)	0.000*
9-12	0.487	0.664 (0.146-1.672)	0.154
ISS ≥16	0.208	1.131 (0.342-1.451)	0.553
Intubation within 1 h of arrival	0.257	2.385 (2.080-4.896)	0.001*
Chest tube within 1 h of arrival	0.256	1.572 (0.021-1.73)	0.077
Ventilatory support	0.232	7.962 (6.74-12.761)	0.000*
Operative intervention	0.219	0.962 (0.941-1.67)	0.861
Injury-hospital arrival interval	0.001	1.000 (0.453-1.613)	0.473
Transferred	0.220	1.869 (1.233-2.561)	0.005*

*Significant. CI: Confidence interval, SE: Standard error, OR: Odds ratio, SBP: Systolic blood pressure, GCS: Glasgow Coma Score, ISS: Injury Severity Score

Table 5 depicts comparison of directly admitted and transferred patients. Significantly higher numbers of directly admitted patients were in shock, had ISS of 16 or above. Despite more severe injury, they had better survival. Transferred patients also needed airway and chest interventions as much as the directly admitted patients.

Table 5

Comparison of directly admitted and transferred patients - Bivariate analysis

Variables	Subgroup	Total	Direct (%)	Transferred (%)	P
n		1181	352	829
SBP (mmHg)	<90	321	121 (30)	200 (24)	0.002*
	≥90	860	231	629
GCS	<9	299	80 (23)	219 (26)	NS
	9-12	161	49	112
	13-15	721	223	498
ISS	<16	899	207	692	<0.00001*
	≥16	282	145 (41)	137 (17)
Intubation within 1 h of arrival	Yes	406	112 (32)	294 (35)	NS
	No	733	240	493
	Already intubated	42		42
Chest tube within 1 h of arrival	Yes	165	57 (16)	108 (13)	NS
	No	1009	295	714
	Transferred with the tube	7		07
Ventilatory support	Yes	414	252 (71)	162 (19)	<0.00001*
	No	767	100	667

*Significant. NS: Not significant, SBP: Systolic blood pressure, GCS: Glasgow Coma Score, ISS: Injury Severity Score

Discussion

The results indicate that survival after major trauma is a function of injury severity, particularly the physiological derangements, injury mechanism, and patient factors, such as age, but not of injury-arrival time lag.

The concept of rapid transport of the injured has probably developed from the wartime data, where survival benefits of reduced time lag from injury to surgery have been well documented.[3] However, that data cannot be generalized as most of these patients were those with isolated systemic penetrating trauma which is contrary to civilian practice.

The authors’ data which mainly include blunt civilian injuries did not confirm the survival advantage of shorter out of hospital time in trauma patients. Published literature on this subject is mostly from HICs with well-established prehospital care. The study by Sampalis et al. in 1993 reported that total prehospital time over 60 min was associated with a significant increase in the odds of mortality.[8] The 1999 study by the same group found that reduced prehospital time was associated with reduced odds of dying, when outcomes were controlled for the severity of injury and age of the patient.[9] However, in 2003, Lerner et al. while evaluating the impact of total out of hospital time over mortality stated that it had no influence over mortality.[10] One of the most exhaustive research on time-to-definitive care in trauma was a 2010 prospective cohort study by Newgard et al. in North America.[11] These investigators have not been able to confirm any relationship between emergency medical service (EMS) intervals and in-hospital mortality among injured patients with physiologic abnormalities despite extensive analysis across several subgroups, including injury type, age, and mode of transport. This group further studied physiologically unstable trauma patients (SBP ≤90 mmHg, GCS ≤12, respiratory rate <10, >29 and those needing advanced airway) in 2015 and found no correlation between in-hospital mortality and EMS time intervals (activation, response, on-scene, transportation, and total).[12] In a systematic review of twenty-one articles, Harmsen et al. concluded that for hemodynamically stable undifferentiated trauma patients, increased on-scene-time and total prehospital time do not increase odds of mortality. However, swift transport is beneficial for patients suffering neurotrauma and those patients with penetrating injury who are hemodynamically unstable.[13]

Undoubtedly, prehospital care needs to develop in LMICs and LICs.[5] In a systematic review, prehospital trauma system has been shown to reduce mortality, particularly in middle-income countries as reported by Henry and Reingold.[6] However, Neilsen et al. have reported lack of funding, leadership, and legislation as the major barriers to its development in LICs and LMICs.[5]

The results also indicate survival advantage to those presenting directly to the tertiary care hospital as compared to those transferred from other hospitals. Direct admission to trauma unit bypassing emergency care has been shown to improve survival in Korean conflict.[3] This may be explained as most war-injured patients have penetrating trauma and do require early surgical intervention. Sampalis et al. in 1997 reported a significant reduction in mortality and morbidity on direct transport to level I trauma center.[14] Their cohort predominantly consisted of patients with road traffic injuries and falls, i.e., blunt injury patients. Cheddie et al., in an appraisal of tertiary trauma center in South Africa, also reported a significant decrease in mortality in patients admitted directly to tertiary trauma center.[15] However, their hospital had 33.7% patients with penetrating trauma, much higher than just 10% in the present study. In an extensive systematic review of 36 observational studies, Hill et al. did not find any significant association between transfer status and in-house mortality although the populations were heterogenous.[16]

The transferred patients had significantly lower number of patients with severe injury [Table 5] yet had higher mortality. An American study has also reported higher mortality and morbidity among transferred patients.[14] Higher mortality has been reported among patients getting treated in level III trauma centers in comparison to those in multidisciplinary level I trauma center in Australia, due to delayed decisions and errors.[17] Further analysis of the present data is also suggestive of deficiency in the primary care of transferred patients. Had primary care been appropriate, the patients would have been transferred after managing airway, breathing, and after controlling obvious external bleeding. Table 5 shows that 35% of transferred patients had to be intubated and in 13%, chest tube had to be inserted within 1 h of arrival. Of 829 transferred patients, intubation and chest tube insertion had been carried out at the primary facility only in 42 and 7 patients, respectively. Although present data have not evaluated infrastructural facilities in referring hospitals, a study from South India has reported inadequate resources in district hospitals to manage trauma patients.[18] There could also be difficulties during transportation. Researchers from Puducherry in India have studied prehospital data of patients who died due to trauma. They have reported loss of valuable time due to lack of interfacility transfer protocols in transferred patients.[19]

The present study has some limitations. In this study, the information about time of injury was provided mostly by the accompanying persons and could not be authenticated due to the lack of prehospital care. Furthermore, transferring hospitals varied from primary care facilities to tertiary care private hospitals. Although a majority of patients were transferred for further management, a few were transferred after definitive care for nursing care, ventilatory care, or delayed definitive care for financial reasons. The study has also not captured the data of available resources and deficiencies in the referring hospitals and problems related to transport.

Conclusion

Survival after major trauma is associated with severity of injury, rather than time taken to reach the hospital and those who are directly taken to tertiary care trauma center have significantly better chance of survival than those transferred from other hospitals, probably because of deficient initial care in referring hospitals. This implies that although safe transport in an appropriate vehicle is desirable, rushing to the trauma center without appropriate initial care may actually be detrimental to patient outcome.

Further study is necessary to evaluate the reasons for deficient care in referring hospitals so that appropriate corrective measures can be taken. In general, primary care facilities need to be strengthened by providing resources including human resources. Implementation of the essential trauma care guidelines published by the WHO and principles of initial assessment and safe transfer taught in ATLS® or similar courses could be possible solutions.

Financial support and sponsorship

The study was supported by Swedish National Board of Health and Welfare and the Laerdal foundation for emergency medicine.

Conflicts of interest

There are no conflicts of interest.