Massive Transfusion Protocol Reactivation as a Novel Marker of Physician Team Under-triage After Injury

Abstract

Background:

Large trauma centers have protocols for the assessment of injury and triaging of care with attempts to over-triage to ensure adequate care for all patients. We noted that a significant number of patients undergo a second massive transfusion protocol (MTP) activation in the first 24-hours of care and conducted a retrospective cohort study of patients involved over a 3-year period.

Methods:

Transfusion service records of MTP activations 2019-2021 were linked to Trauma Registry records and divided into cohorts receiving a single vs a reactivation of the MTP. Time of activation and amounts of blood products issued were linked to demographic, injury severity, and outcome data. Categorical and continuous data were compared between cohorts with chi-squared, Fisher’s, and Wilcoxan tests as appropriate, and multivariable regression models were used to seek interactions (p < 0.05).

Results:

MTP activation was recorded for 1,884 acute trauma patients over our three-year study period, 142 of whom (7.5%) had reactivation. Factors associated with reactivation included older age (46 vs 40 yrs), higher injury severity score (ISS, 27 vs 22), leg injuries, and presentation during morning shift change (5-7 AM, 3.3% vs 7.7%). Patients undergoing MTP reactivation used more RBCs (5U vs 2U) and had more ICU days (3 vs 2).

Conclusions:

Older patients and those presenting during shift change are at risk for failure to recognize their complex injury patterns and under-triage for trauma care. The fidelity and granularity of transfusion service records can provide unique opportunities for quality assessment and improvement in trauma care.

Introduction:

Prompt and accurate identification and triage of injuries are essential for effective trauma care. Benchmarks for triage quality have been studied, and the consequences of over- and under-triage documented, but this work is largely focused on ensuring that pre-hospital providers transport patients to facilities with resources commensurate with their injuries and that personnel and resources are mobilized effectively.(1-10) Triage accuracy once patients reach advanced care is less well studied and represents an important gap in our understanding. Specifically, we have no established benchmarks to identify patients whose injury severity may have been underappreciated by their physician teams. Thus, we have no valid ways to collect data on under-triage, to assess those data, or to develop strategies to prevent or minimize under-triage.

To address this knowledge gap, we leveraged the granular nature of blood bank data from a large, US regional, academic Level I trauma center with a well-developed and supported massive transfusion protocol (MTP). Using these data, we identified specific patient and provider factors associated with reactivation of a massive transfusion protocol within 24-hours after initial activation stand-down (defined as the cooler containing emergency-release blood products being returned to the blood bank at the instruction of the bedside physician team). Our hypothesis in this work was that specific risk factors would be identifiable and that reactivation patients would have worse outcomes than those requiring a single MTP activation.

Methods:

Institutional Review and Approval

This retrospective observational cohort study was approved by our Institutional Review Board (IRB Center No.00006777) and follows Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for observational studies (STROBE checklist available as Supplemental Digital Content [SDC] Table 1). The requirement for written informed consent was waived by our IRB as this study used only retrospective clinical data collected during routine practice.

Study design and setting

From blood bank data, we identified all patients admitted to our Level I trauma center from January 2019 through December 2021 who required activation of the institutional massive transfusion protocol (MTP). Data from patients thus identified were then combined with institutional trauma registry data; non-trauma-related activations were excluded, along with duplicate and incomplete records (Figure 1). Timestamps for MTP activations were collated and a subset of patients who required two or more activations within a 24-hour period were identified as the “reactivation” cohort. Importantly, initial MTP activation occurs automatically (i.e., without the need for an order from a clinician) for patients that meet criteria as “Full Trauma Activations” (Criteria outlined in SDC Table 2). The blood bank was activated automatically at time = 0 in 92% of our cases. Correspondingly, the decision to stand-down the blood bank, rather than the decision to activate it in the first place, represents the first documented physician-level decision that provides insight into their impression regarding the clinical trajectory of most trauma patients who required MTP activation in our system. After MTP has been activated, the same physician-trauma team (led by a trauma surgery attending and emergency medicine attending) is responsible for a patient’s resuscitation at least until blood bank stand down (i.e., no hand-offs occur during an MTP activation). Additionally, MTP “reactivation” events were defined as the emergency release blood cooler having been physically returned to the blood bank at the direction of the bedside physician trauma team, serving as the basis for the assumption that this same team may have underappreciated the severity of the patient’s condition, given the subsequent request for re-activation within 24 hours of the standdown.

Figure 1: CONSORT Diagram for Cohort Creation.

CONSORT Diagram for Massive Transfusion Audit Database and Institutional Trauma Registry Collation

Reactivation patients were then compared to those requiring only a single MTP activation (no matter how many blood products the latter might require over the course of their massive transfusion event). Variables of interest included: age, arrival status (from the scene of injury versus transfer), injury severity score (ISS), pattern of injury indicated by regional anatomic injury scores (AISs), arrival day and time, admission International Normalization Ratio of the prothrombin time, and admission blood alcohol level (BAL) and urine toxicology.

Data Presentation & Statistical Analysis

Continuous data are presented as medians with interquartile ranges (IQR); categorical data are presented as proportions. Unadjusted comparisons of risk factors for reactivation and outcome as blood product utilization and in-hospital mortality were performed using Wilcoxon rank sum, Chi-square, and Fisher’s exact tests as appropriate for continuous or categorical data. Pre-defined candidate risk factors for reactivation included age ≥65 years, injury severity score (ISS), regional anatomic injury scores (AIS), from-scene arrival, arrival during trauma team sign-out time (0500-0659 hours), admission shock index (SI, pulse/systolic blood pressure), INR ≥1.5(11), and intoxication at admission (abnormal BAL or urine toxicology). Arrival time was compared as the proportion of patients requiring reactivation arriving in a two-hour window surrounding trauma team hand off versus the proportion of those requiring reactivation in the remaining twenty-two hours of the day. This two-hour time window was selected to capture both Trauma Surgery hand-off (0530h) and Emergency Medicine Trauma Team hand-off (0600h). Clinical outcomes of interest included: in-patient mortality, blood product utilization (including whole blood, packed red blood cells, plasma, and apheresis platelets in plasma), intensive care unit and total length of stay.

Input variables likely to lead to undertriage for hemorrhage and so lead to differences between single activation and reactivation patients were sought in multivariable logistic regression analysis. Given the small sample size of 142 reactivation patients and multiple potential variables, AIS were restricted to lower extremities where a potential difference between trauma teams and orthopedic consultants was likely and discriminant probability was set at 0.05. All data analyses were performed using R Studio (Version 1.4.1717, Boston, MA).

Results:

General and demographic

Overall, MTP activation was recorded for 1,884 acute trauma patients over our three-year study period, 142 of whom (7.5%) met reactivation criteria (Table 1). “Full Trauma Activation” status was recorded for 92% of patients, and this percentage was identical for reactivation patients. Amongst reactivation patients with complete time stamp data, the median time between initial activation and reactivation was 104 minutes [IQR: 60-159]. Reactivation patients were more severely injured than single activation patients (ISS 27 [IQR:17-38] vs 22 [IQR:10-32]p<0.01) and more likely to have lower extremity injury (p < 0.001) but showed no difference in in-patient mortality (19% vs. 19%; p=0.9). Compared to patients requiring a single activation, reactivation patients had higher blood product utilization in the first 24 hours (2 vs 5 units, respectively; p <0.01). Reactivation patients also had longer ICU lengths of stay (2 [IQR 1-6] vs 3 [IQR 2-8] days; p<0.01) and total length of stay (7 [IQR2-8] vs. 10 [IQR5-25] days, p<0.001). However, after adjustment for ISS and age ≥65 years, associations between reactivation and increased lengths of stay were attenuated (ICU days 0.75 [95% CI −1.1, 2.6] p = 0.4; total days 5.5 days [95% CI 1.5, 9.4]; p<0.01).

Table 1:

Cohort Demographics, Injury Characteristics, and Outcomes

Demographics & Injury Characteristics	No Reactivation N=1,742 1	Reactivation N=142 1	p-value 2
Gender			>0.9
Female	413 (24%)	34 (24%)
Male	1,329 (76%)	108 (76%)
Median Age, years, (IQR3)	40 (26, 58)	43 (29, 66)	0.06
Race			0.1
Asian	97 (5.6%)	10 (7.0%)
Black	248 (14%)	20 (14%)
Native American	58 (3.3%)	0 (0%)
Pacific Islander	31 (1.8%)	1 (0.7%)
White	1,230 (71%)	108 (76%)
Not Documented	78 (4.5%)	3 (2.1%)
Primary Injury Type			0.4
Penetrating	530 (30%)	35 (25%)
Blunt	1,195 (69%)	107 (75%)
Burn	10 (0.6%)	0 (0%)
Other	7 (0.4%)	0 (0%)
Hospital Transfer			0.11
No	1,244 (71%)	113 (80%)
Unknown	1 (<0.1%)	0 (0%)
Yes	497 (29%)	29 (20%)
Injury Severity Score (IQR3)	22 (10, 32)	27 (17, 38)	<0.01
Shock Index on Admission	0.89 (0.69, 1.05)	0.93 (0.66, 1.06)	0.18
Outcomes
Length of Stay (days)	7 (2, 8)	10 (5, 25)	<0.01
ICU Length of Stay (days)	2 (1, 6)	3 (2, 8)	<0.01
Ventilator Days	2 (0, 3)	2 (1, 5)	<0.01
Inpatient Mortality	331 (19%)	27 (19%)	>0.9
Time to death (hours)	25 (2, 125)	42 (1, 134)	0.03
Cause of Death4
Hemorrhage	65 (19%)	2 (7%)	-
TBI/High Spine Injury	98 (30%)	13 (48%)	-
Organ Failure	64 (19%)	3 (11%)	-
Poly-/catastrophic Trauma	63 (19%)	5 (19%)	-
Unknown or Not Recorded	41 (12%)	3 (11%)	-

¹n (%); Median (IQR)

²p-value: estimate that probability results are by chance: Pearson's Chi-squared test, Fisher's exact test.

³IQR: Inter-Quartile Range

⁴Cause of death: descriptive summary from trauma registry entries

Risk Factors for MTP Reactivation (Under-triage)

The relationships between reactivation and the pre-selected potential risk factors—age ≥65 years, from-scene arrival, arrival during trauma team sign out, ISS, AIS-Lower Extremity, Shock Index, INR ≥1.5, and intoxication—are shown in Table 2. Reactivation was only associated with older age (≥65 years, 12%; <65 years, 6.6%, p<0.01), non-transfer status, ISS, AIS-LE, and arrival during trauma team sign-out (16%) rather than any other time (7.2%, p<0.01) (Figure 2). Elevated admission Shock Index, INR, BAL, abnormal urine toxicology, and day of the week of arrival were not associated with reactivation (SDC Tables 3 and 4 and SDC Figure 1).

Table 2.

Candidate Risk Factors for Under Triage Associated with MTP Reactivation in Logistic Regression Analysis

Characteristic	Odds Ratio (95% Confidence Interval)	p-value
Geriatric (Age ≥ 65 Years)	1.86 (1.25, 2.72)	0.002
Transfer From Outside Hospital	0.64 (0.41, 0.97)	0.039
Injury Severity Score	1.03 (1.02, 1.04)	<0.001
Lower Extremity Anatomic Injury Score	1.22 (1.12, 1.34)	<0.001
Arrival During Trauma Hand-Off Window1	2.46 (1.20, 4.63)	0.008
Initial ED2 Shock Index (Pulse/Systolic BP)	1.36 (0.85,2.08)	0.2
Initial INR3 ≥ 1.5	1.2 (0.75, 1.85)	0.4
Positive EtOH4 (>0 BAL5)	0.74 (0.47, 1.11)	0.2
Positive Urine Tox	0.89 (0.61, 1.29)	0.6

¹Trauma Hand-Off Window: 5 - 6:59 AM compared to other 2-hour time blocks

²ED: Emergency Department

³INR: International Normalization Ratio of the prothrombin time

⁴EtOH: Ethanol

⁵BAL: Blood Alcohol Level

Figure 2: Frequency of Patient Requiring MTP Activation and Reactivation by ED Arrival Time.

Upper pane depicts total frequency of patients who required any massive transfusion protocol activation by hour arrived in the emergency department on a 24-hour scale. Lower pane depicts the frequency of patients who required MTP reactivation by hour arrived in the emergency department on a 24-hour scale. Red shading depicts the two-hour window (0500-0659) around trauma team hand-off.

In-hospital outcomes

Inpatient mortality was 19% in both single activation and re-activation cohorts (p>0.9). Median time to death in hours was somewhat longer for reactivation patients (42 hours [IQR 1,134] vs 25 hours [IQR 2,125], p 0.03), though the range was narrower for reactivation patients (9 – 432 hours vs 0 – 3307 hours for non-reactivation) and deaths among non-reactivation patients included all 61 deaths that occurred within the first hour of care at our center. Cause of death data is depicted among the in-hospital outcome summaries in Table 1.

Discussion:

Using blood-bank MTP reactivation and trauma registry data, our study provides novel insight into aspects of the acute in-hospital triage of seriously injured patients in advanced civilian trauma care. Focusing on the decision point regarding the need to continue or “stand down” the immediate beside access to blood products for resuscitation represented by MTP activation, we identify a particular patient group—those aged ≥65 years —those with lower extremity injuries--and a particular time of day in the acute care setting—early morning trauma-team hand-off —more likely to have their initial MTP activation cancelled, despite their ultimately being determined to be worse injured than those with a single MTP activation. In our cohort, these patients went on to use more blood products and to require increased ICU and ventilator days.

Acute assessment of severe trauma is a dynamic and difficult art at best, and assessment of blood product use in resuscitation—retrospective or prospective—has been confounded by the effects of survival bias from the beginning.(12) With that acknowledged, our findings corroborate prior work in both forward care and in-hospital settings demonstrating an underappreciation of injury severity among geriatric populations.(13-17) Femoral fractures are common in this population and often associated with slow venous bleeding that can be missed in the bulk of the thigh. Our findings also highlight a potential systems vulnerability during hand-off periods and processes among trauma providers that may contribute to an initial underappreciation of injury severity. To our knowledge, the concept of triage accuracy has not been explicitly studied in the context of trauma center-based, physician led teams nor have specific metrics for triage been proposed that provide guidance regarding how best to balance diagnostic vigilance with responsible resource utilization. However, our work suggests that standing down the blood bank from an active MTP is a reasonable decision point for the trauma team to pause and reach consensus regarding an individual patient’s status, including the presence of risk factors for under-triage, and design the next steps of their care plan accordingly.

As a research tool, retrospective evaluation of MTP stand-down/reactivation events has several advantages. These events provide an objective marker of under-triage, free from the potential biases introduced by clinician adjudication.(18) They avoid the Hawthorne effect, in that the trauma teams are behaving normally without the influence of knowingly being monitored which may influence assessment of under-triage using recorded trauma resuscitations.(19) They are also documented by data routinely collected and stored as a part of rigorous, federally-regulated and time-catalogued, blood-banking procedures independent of trauma-team bedside resuscitation recording.

Our study has several important limitations. Our use of retrospective data means that granular detail of each patient’s trajectory and physician team decision making were not available raising the possibility that unmeasured confounding influenced our conclusions. Additionally, MTP activation criteria and stand down decisions are a function of our institutional practice and may limit generalizability of our results and reproducibility of our methods by other centers. Finally, despite multi-year data from a high-volume trauma center, our number of reactivations was relatively low (n=142) potentially leading to a lack of power in detecting risk factors or outcomes associations with MTP reactivation.

Summary and conclusions:

Our findings support emerging evidence that physician-led trauma teams may underappreciate injury severity in geriatric populations, can miss slowly developing blood loss, and that hand-offs between trauma teams represent a systems level vulnerability with respect to patient care. For injured patients with an MTP in progress, the decision to de-escalate resources in the form of standing-down the blood bank offers a natural reassessment point to reach consensus amongst members of the trauma team regarding patient trajectory, risk factors for missed/underappreciated injury, and next-step planning. The fidelity and granularity of transfusion service records can provide unique opportunities for quality assessment and improvement in trauma care.

Supplementary Material

1

SDC Table 1. STROBE Checklist of items that should be included in reports of cohort studies

SDC Table 2. Institutional Full Trauma Activation Criteria

SDC Table 3. Alcohol Intoxication and Massive Transfusion Protocol Reactivation

SDC Table 4. INR Derangement and Massive Transfusion Protocol Reactivation

SDC Figure 1. Proportion of Reactivations by Day of the Week Bar chart depicting the proportion of patients requiring MTP reactivation by day of the week. There was a slight increase in proportion of patients requiring reactivation on Friday compared to any other day of the week, however, this increase (9.4% vs 7.2%) was not statistically significant (p=0.2).

Social Media Summary, Tags, and Author Handles:

Examining massive transfusion protocol (MTP) reactivation within 24-hours of an MTP stand-down order as a novel marker of physician under-triage identified age≥65 and ED arrival during trauma team hand-off and lower extremity injury as risk factors for under-triage.

@WeykampMike @TraumaBryce @MonicaVavilala @HMCTraumaT32