The trauma registry is a locally maintained database that includes information on all trauma patients, including demographic and injury data, and is an essential component of a Level 1 trauma center. When accurate, the information from this database can promote efforts in performance improvement, quality assurance, injury prevention, and trauma research. In addition, the registry is used for risk-adjusted benchmarking of trauma centers, as well as in support of continued accreditation and national designation as trauma centers (Calland et al., 2012 ; Moore & Clark, 2008). The data that are used for benchmarking reports heavily rely on consistent and accurate data in order to create these reports (Calland et al., 2012). There are many elements that can hinder accurate data. Patients who die on arrival, die in the emergency department (ED), or die within 48 hr often do not have accurate injury severity scores (ISSs). The models used for risk adjustment in trauma centers often use a patient's physiology on arrival to the ED, ISS, and age to predict mortality in severely injured patients (Calland et al., 2012).
A trauma patient is considered to be any patient with a trauma team activation and/or any admitted patient with a diagnosis of traumatic injury. One essential field in this database is the ISS. The ISS of a patient tells how injured the patient is. This information is collected by each individual institution's trauma registrars, who undergo training on proper and accurate coding of injuries based on International Classification of Diseases (ICD) codes (Moore & Clark, 2008). The data are then sent to the center's regional and national registries. In trauma-related deaths, the Trauma Quality Improvement Program (TQIP) database divides these patients into those who died within 15 min of their arrival and those who died after 15 min of their arrival. Most admitted trauma patients, particularly those who survive beyond 48 hr of their injury, undergo a thorough evaluation and radiological examination to determine their injuries. For these patients, the ISS recorded in the trauma registry is accurate. Some patients do not have the opportunity to undergo such intensive testing to uncover their injuries. This is especially true for patients who expire on arrival, in the ED, or within 48 hr, as they are usually not subjected to comprehensive testing. Without diagnostic examination, many of these patients are without accurate diagnosis. Their ISS, therefore, does not accurately reflect how injured they are. Autopsies can identify unrecognized injuries and may provide an improved ISS for the mortally injured patient. To combat this deficiency, the Trauma Service developed a plan to obtain autopsy reports on all mortalities, although this study only looked at the autopsy reports of those patients who expired on arrival, in the ED, or within 48 hr.
Obtaining data from autopsy reports, which would include injuries not diagnosed before death, could help increase the patient's original ISS, which only reflected their external injuries discovered on physical examination (Marx, Simon, Jumbelic, Sposato, & Nieman, 2004). A study that evaluated trauma patients who died within 1–2 hr of arrival to the ED found that an average of 1.2 additional injuries per patient were found with the help of the autopsy report (Ong et al., 2002 ; Sharma, Gupta, Harish, & Singh, 2005). Having this additional information would improve the accuracy of the trauma registry and allow for more credible benchmarking, injury prevention, and trauma research endeavors.
It is our belief that obtaining autopsy reports for these patients can help better evaluate their injuries, which has previously been advocated by many centers, thereby allowing for a more accurate ISS (Calland et al., 2012). We sought to determine whether and how autopsy data improve the accuracy of the trauma registry data for mortally injured patients. We believe that the addition of the autopsy data will increase the accuracy of trauma registries in centers that obtain these reports. This retrospective study, conducted at an American College of Surgeons (ACS)-verified Level 1 trauma center, demonstrates the importance of obtaining autopsy reports and its benefits to trauma registry accuracy.
MATERIALS AND METHODS
A retrospective study was conducted to look at trauma deaths between January 2014 and August 2017 at our institution. Our research department submitted an institutional review board request through the institution's research center, which was approved before gathering of data from the trauma registry. Of the 186 trauma-related deaths in this time frame, we analyzed the ISS of patients who died within 48 hr of the trauma-related injury using the information from the trauma evaluation that was performed. These patients were further divided into patients who died within and after 15 min of their injury, just as they are in the trauma registry. From the patients who had autopsies, their ISS was reevaluated using the additional data provided by the examination.
The average ISS of cases prior to autopsy report was analyzed. Once the autopsy reports were obtained, the trauma registrars updated each patient's diagnosis section of the trauma registry, thereby increasing the ISS. The overall average ISS after autopsy was compared with the overall average ISS of those without autopsy. The difference between the ISS prior to autopsy and the ISS after autopsy was also assessed. The average difference between patients' ISS before and after autopsy was analyzed by a statistician using a t test to calculate statistical significance at p < .05. In addition, a comparison was made between the change in ISS in those who died within 15 min and after 15 min of arrival to the ED. This difference was also analyzed using a t test to calculate statistical significance at p < .05. The same analysis was also performed comparing the change in ISS between gender, race, and ethnicity.
The population of interest in this study comprised all patients seen in the NewYork-Presbyterian/Queens ED after sustaining a traumatic injury and who expired on arrival, while in the ED, or within 48 hr of arrival. The time frame for this study was between January 2014 and August 2017. A total of 35 cases were identified as being eligible for this sampling. The data for this study were collected solely from the Trauma Service's trauma registry.
Of the 35 cases identified (Table 1), the majority were male (63%). The average age of subjects was 61.8 years. The youngest patient was 19 years, whereas the oldest patient was 94 years. The majority of patients had full trauma team activations (89%). The majority of the patient population was white (51%; 72% of which were Hispanic or Latino), and 31% were Asian. The average length of stay in the ED was 45.7 min. Twenty of the 35 (57%) identified cases had autopsies performed. One patient's cause of death was ruled to be nontraumatic and therefore was excluded.
From the 19 cases that met our inclusion criteria, we compared the change in ISS of the patients before and after autopsy (Table 2). The average ISS of these cases before including the additional data from autopsy was 13, whereas the average ISS of these same cases after inclusion of the autopsy data was 49, conferring an overall significant increase in ISS (p < .001).
In addition, of the 19 cases, only one patient presented dead on arrival. The remaining 18 were divided into those who died within the first 15 min of arrival to the ED (Table 3) and into those who died thereafter (Table 4). The average ISS before autopsy in those who died before and after 15 min was 5 and 20, respectively (Figure 1). In comparison, the average ISS after autopsy in those who died before and after 15 min was 57 and 35, respectively. This equals a significant average change in ISS of 53 and 12, respectively (p ≤ .001). However, in comparing the change in ISS between genders (Figure 2) and ethnicity (Figure 3), the change was not significant. In addition, our patient population was too small to be able to compare the change in ISS between races (Figure 4).
As noted in the literature, early traumatic death often results in underestimated ISS. This is apparent in our comparison of the ISS in patients who died in the ED before and after autopsy. Having a significant increase in ISS with the addition of information provided by autopsy shows that there are many injuries not diagnosed immediately on arrival to the ED. The increase in ISS shows that the data submitted to the trauma registry are incomplete without the information provided by the autopsy report. This seems to be particularly true in patients who died in the early time frame following their trauma (on arrival or within 15 min of arrival). When analyzing the significant difference in the ISS of the patients who died within the first 15 min of arrival into the ED with those who died after the first 15 min, it is clear that those who die early had more missed injuries at the time of their death than those who were alive for a longer period of time. This is likely due to the fact that the trauma team did not have the time to perform the necessary tests, including imaging, before the death of the patient. The data from these autopsy reports allow institutions to include additional injuries sustained during trauma into the trauma registry that would otherwise have been unknown to the trauma service, further validating the information in the institution's trauma registry.
Accurate data from the trauma registry are the cornerstone of an ACS-verified Level 1 trauma center. When data are accurate, they can be used to facilitate performance improvement, injury prevention, and research endeavors. One data element that is an essential component to the Trauma Service is ISS. When the ISS is accurate, it can reflect how injured the patient is. Some of our hospital's most critically injured patients are those who expire on arrival, while in the ED, or within 48 hr of arrival. Many of those patients do not undergo radiological examination to determine their internal injuries. As a result, their ISSs are not an accurate reflection of their actual injuries. For example, one patient included in our data set presented to the ED with only a few superficial lacerations, which led to our trauma registrars to assign the patient a low ISS, which is equivalent to only the minor injuries that were noted on secondary survey. However, on autopsy, the patient was found to have an aortic transection, which increased the patient's ISS to reflect an unsurvivable injury. The recalculated ISS for such patients better explains their early death and inability for the trauma department to resuscitate them during their short admission. Once patients expire, there is an opportunity for autopsies. Although some patients' families do not agree to this due to religious or cultural reasons, our study revealed a benefit for those patients who have autopsies performed.
Although there has been a downward trend in the performance of autopsies in the last 10 years, recent literature has noted the continued importance of the autopsy as a final medical examination (Van Den Tweel & Wittekind, 2015). Despite the advances in diagnostic technology, there are still significant discrepancies between clinical and anatomical findings to encourage the continued use of the autopsy after a patient's death. In addition to diagnostic technology for the most part taking the place of autopsies, the financial burden on the hospital is another reason for the general decline in the performance of autopsies. A full autopsy can cost several thousand dollars, which is generally not reimbursed by insurance companies. However, in a study conducted at a tertiary trauma center, after autopsy data had been added to the medical record, 28% of the records investigated had changes to their Diagnosis Related Groups, which allowed for increase in funding to the hospital (Cheng, Gilchrist, Robinson, & Paul, 2009). Another study in Florida noted an 85% error rate in medical coding that had led to significant loss of funding to the hospital system (Marshall & Adema, 2005). Therefore, although autopsies can be costly to the health care system, it is these data that are used to determine the funding and reimbursement of the hospital. At our hospital, after every patient death, it is up to the medical practitioner to call the Medical Examiner's office and discuss each case with the medical examiner. Our medical examiner is required to accept every case that is related to a trauma for further investigation. Our trauma registrars then actively seek to obtain the autopsy reports once they have been completed. The information from the autopsy reports is then used to modify the data sent to the registry as appropriate.
Our study demonstrated the significant increase in ISS and improvement in trauma registry accuracy that can be accomplished with autopsy reports. The average ISS of patients prior to autopsy was 13. After obtaining autopsy reports for the eligible patient population, the average ISS increased to 49. An ISS of 13 does not reflect a severely injured population of patients. It would appear that these patients did not have injuries severe enough to warrant early mortality. After autopsies were received and analyzed, the average ISS of 49 reflected a more critically injured population that likely expired as a result of these injuries. This study helped improve the accuracy of the data in our trauma registry, which will assist in the development of performance improvement efforts, injury prevention ventures, and trauma research efforts.
Our study is limited by a small number of patients (n = 19).
Our study identified the importance of obtaining autopsies in trauma patients. The data from autopsy reports adjust the ISS of these patients to include missed injuries that may have been the cause of death, improving the accuracy of the database. Although autopsies can cause a significant financial burden to the health care system, in the long term, these fiscal implications are small compared with the benefits of the additional information provided by these autopsies. The data generated by coders are used by various agencies such as the Centers for Disease Control and Prevention to generate epidemiological patterns of diseases, which aid in the creation of screening and prevention programs. Similarly, trauma centers and national trauma organizations can use these data to educate the community about patterns of injury seen within particular regions. This will also allow for the creation of prevention programs. However, to make effective prevention programs, it is crucial to analyze patterns of injury that come from an accurate database. Having more accurate registry data for trauma deaths may also further guide the development of performance improvement and trauma research. In addition, it will allow for accurate benchmarking of trauma centers, as it is the trauma registry that national quality improvement programs use to assess the performance of each individual trauma center. Finally, these data are used for billing purposes and are used to determine a hospital's reimbursement for the hospital stay. Without the autopsy data, the ICD-10 codes would not reflect the severity of injury and, in turn, would not justify the resources required for the care of the patient, including trauma team and operating room resources, as well as the use of an intensive care unit bed. As a result, the hospital would have a smaller reimbursement than it would if the patient's list of injuries accurately reflected his or her mortally injured state on arrival. Therefore, although a tedious process, having a strong working relationship between the Medical Examiner's office and the trauma center can help simplify the process of obtaining these autopsy reports so that the registrars can appropriately code for the injuries sustained that were not found antemortem.
- Mortally injured patients often expire prior to a complete evaluation for injuries, leading to missed injuries that do not get reported to the trauma registry.
- Autopsy reports provide additional information that better identifies the severity of injury in mortally injured patients.
- Autopsy data should be obtained by trauma centers to improve the accuracy of their trauma registry.
The authors thank Dr. Pierre Saldinger, Dr. Jason Sample, Mary Ellen Zimmermann, and Sarah Stankiewicz for their continued support of trauma research at their institution.
Calland J. F., Nathens A. B., Young J. S., Neal M. L., Goble S., Abelson J., Hemmila M. R. (2012). The effect of dead-on-arrival and emergency department death classification on risk-adjusted performance in the American College of Surgeons Trauma Quality Improvement Program. Journal of Trauma and Acute Care Surgery, 73(5), 1086–1092. doi:10.1097/ta.0b013e31826fc7a0
Cheng P., Gilchrist A., Robinson K. M., Paul L. (2009). The risk and consequences of clinical miscoding due to inadequate medical documentation: A case study of the impact on health services funding. Health Information Management Journal, 38(1), 35–46. doi:10.1177/183335830903800105
Marshall J., Adema D. (2005). Reinventing radiology reimbursement. Radiology Management, 27(2), 36–46.
Marx W. H., Simon H. M., Jumbelic M., Sposato E., Nieman G. (2004). Severity of injury is underestimated in the absence of autopsy
verification. The Journal of Trauma: Injury, Infection, and Critical Care, 57(1), 46–50. doi:10.1097/01.ta.0000135502.18989.61
Moore L., Clark D. E. (2008). The value of trauma registries. Injury, 39(6), 686–695. doi:10.1016/j.injury.2008.02.023
Ong A. W., Cohn S. M., Cohn K. A., Jaramillo D. H., Parbhu R., Mckenney M. G., Bell M. D. (2002). Unexpected findings in trauma patients dying in the intensive care unit: Results of 153 consecutive autopsies. Journal of the American College of Surgeons, 194(4), 401–406. doi:10.1016/s1072-7515(02)01123-7
Sharma B., Gupta M., Harish D., Singh V. P. (2005). Missed diagnoses in trauma patients vis-à-vis significance of autopsy
. Injury, 36(8), 976–983. doi:10.1016/j.injury.2004.09.025
Van Den Tweel J. G., Wittekind C. (2015). The medical autopsy
as quality assurance tool in clinical medicine: Dreams and realities. Virchows Archiv, 468(1), 75–81. doi:10.1007/s00428-015-1833-5