This study examined the 30-day mortality rate in Indian urban trauma centers and estimated it to be 12.4% of all trauma admissions. This Indian rate was higher than the in-hospital mortality rates reported by HIC trauma centers for similarly injured patients [8, 9]. However, the current 12.4% mortality rate was an improvement over that previously reported by Indian trauma registries of 21.4% (2015) from similar urban trauma centers [6]. The 43% reduction in 30-day mortality is dramatic, as compared to the recent earlier studies. This reduction is more significant than most other trends (for example, reduction in maternal or infant mortality). There have been improvements in communications, development of infrastructure, and equipment. But there has been no significant trauma systems implementation or intervention (neither pre-hospital nor in-hospital) that can be credited for this observed improvement in mortality. Therefore, the authors would like to attribute this reduction to the differing case-mix between the previous studies and the current one. Our study sites were Indian urban referral trauma centers at University hospitals and are not representative of the broader situation across India [6, 12]
The majority of deaths are in the first week, but after the first 24 h, interventions need to be designed to address this subgroup of fatalities. The delayed deaths may be attributed to an inadequate pre-hospital resuscitation during the transfer from the injury site to the hospital, with no pre-hospital fluid or blood resuscitation during a hospital transfer. The low incidence of patients in hypovolemic shock (and barely recordable BP) on hospital arrival (3.5%) may suggest that many of the severely compromised patients who would have died within 24-h of in-hospital stay may have died in the pre-hospital phase. For physiological parameters (SBP, HR, RR, SpO2), statistically significant associations existed. However, this significant association may not be clinically relevant and such finding may be due to the high power (i.e., large sample size) of the study.
A low GCS (< 9) on arrival was seen to be associated with 53% mortality in our study. In many trauma studies, a low on-arrival GCS was considered as a strong predictor of trauma mortality [13,14,15]. Our data confirmed this finding and the overall mortality progressively decreased with an increasing on-arrival GCS. It reduced from 19% in the moderate GCS category to 3% in the mild GCS category. This finding of the GCS and mortality association in this large dataset validates the appropriate use of GCS for triage on-admission in Indian trauma patients. Increasing age is associated with a higher probability of 30-day mortality and this is consistent with findings worldwide [15].
The process of care delays tend to be relatively pronounced across LMICs, including India [6]. The ‘second delay’ or pre-hospital transit time was the median time between injury and arrival (the delay in reaching care, also called the second delay) was slightly longer for survivors than non-survivors. This may mean that although there is no formal system of pre-hospital triage, injury victims with severe conditions likely to die need to be sent directly to the trauma center by the first responder, who could be a bystander or the police [16].
The ‘third’ delay or the delay in initiating management is an important delay peculiar to the LMICs and has not been researched adequately. In our study, this was the median time from arrival at hospital to first measurement of vital signs, and this served as a proxy measure for the third-delay. The third delay was significantly shorter (p < 0.001) for survivors than non-survivors. Relative proportions of survival in each of the groups were compared based upon time to first vital sign recording. We found that delays greater than 15 min, as a proxy for the third delay, was seen in non-survivors. While it is tempting to attribute death to the longer third delay in non-survivors, this particular variable was missing in 26.3% of cases. However, this is an opportunity for improvement, as the patient had arrived at the trauma receiving center, but did not have their vitals checked immediately, resulting in a delayed triage. This is often attributed to the backlog of patients in the receiving area, a lack of pre-hospital notification, other on-going procedures in a human resource constrained environment and varying trauma care protocols [16].
This third delay could be improved by immediate triaging on arrival and pre-hospital notification and trauma team call-out protocols, which have been piloted in a separate arm of this AITSC research project. The variability in the process of care across different participating institutions is visible in the graphs. However, we did not compare between participating institutions as they had differing financial and manpower resources.
While the study’s strengths were that geographically diverse large-Indian cities were included in a registry with a relatively low proportion of missing values, this study had many limitations. The study design included only high-volume urban referral trauma tertiary-care centers, and therefore, the findings are externally valid only in the Indian urban setting. There is a referral bias as more severe cases are referred to the study hospitals, which leads to relatively higher mortality in tertiary care institutions [17]. Further, patients who were unable to reach a tertiary care hospital or died en route were not included in the AITSC registry or this study. Therefore, we are unable to comment on the pre-hospital care processes and outcomes in this paper. Perhaps some of the higher socioeconomic group patients and stable patients may have availed treatment in private institutions, and there is a bias toward the lower socioeconomic group trauma patients, who come to public university hospitals. This contributes to the pedestrian (road or rail) trauma victim pool in our registry, as compared to car-occupants.
Also, mortality after discharge (even within 30 days) could not be captured in this study, as there was no protocol of home-based follow-up in this study. Capturing time from documented notes is often inaccurate and injury time is often a matter of conjecture. Also, resuscitative treatment can often begin before actual documentation, especially in an unstable patient. In India, admission time varies with the administrative formalities for issuing an admission case file.
The implications of this study are that a standard measure is used to measure mortality across Indian trauma centers, allowing comparison across trauma centers. On-admission physiological vital signs are adequate triage tools for prioritizing patients who are likely to die within 24 h. The best practices and reduced time delays in the best performing participating sites can be replicated across other institutions to improve survival in trauma victims.
Besides pre-hospital notification, the AITSC project also studied improved trauma reception and resuscitation, trauma quality improvement programs, and post-trauma discharge rehabilitation [9]. The findings of these arms once published will augment the knowledge base of this research paper. Mortality remains a gross measure of outcome, and further research into trauma morbidity would be important.