Secondary Logo

Journal Logo

AAST 2020 PODIUM PAPERS

Regionalization of trauma care by operative experience: Does the volume of emergent laparotomy matter?

Tang, Andrew MD; Chehab, Mohamad MD; Ditillo, Michael DO; Asmar, Samer MD; Khurrum, Muhammad MD; Douglas, Molly MD; Bible, Letitia MD; Kulvatunyou, Narong MD; Joseph, Bellal MD, FACS

Author Information
Journal of Trauma and Acute Care Surgery: January 2021 - Volume 90 - Issue 1 - p 11-20
doi: 10.1097/TA.0000000000002911
  • Free
  • Editor's Choice

Abstract

The relationship between experience and performance on the individual level is an intuitive one. At the institutional level, this relationship becomes more intricate, with the overall outcome being a function of input from multiple providers.1 Nonetheless, the effect of a larger number of patients undergoing a similar surgical procedure at a given hospital on patient outcomes has been explored in multiple surgical disciplines.2–10 The so-called volume-outcome relationship is an important consideration, more so in the field of trauma, where preoperative patient optimization is limited by the extent of injuries incurred.11 This makes efforts to improve trauma patient outcomes, whether be it on the trauma service level, trauma center level, or even regional trauma care level, essential.

The American College of Surgeons Committee on Trauma recognizes the annual trauma volume as a hard criterion for appropriate trauma center verification level assignment.12 However, the annual trauma volume as an important determinant of patient outcomes has been consistently subject to debate.13–18 Recently, more interest has been raised in differentiated pathways of care or high-risk trauma patients. These pathways are directed by the receiving center’s case-specific volume, such as the volume of traumatic brain injury patients or geriatric trauma patients.19,20

One commonly performed procedure in the trauma setting is laparotomy following either blunt or penetrating injury. Emergent laparotomy is a highly morbid procedure and is associated with significant mortality.21–23 Defining important determinants influencing trauma patient outcomes following emergent laparotomy is essential in directing efficient resource allocation and enhancing overall quality of care.24 Therefore, the purpose of our study was to examine the relationship between a trauma center’s injury-specific laparotomy volume and outcomes in blunt and penetrating abdominal trauma patients and to determine whether this derived volume is a suitable proxy for outcomes.

PATIENTS AND METHODS

Study Design and Population

We performed a 1-year (2017) retrospective analysis of the American College of Surgeons (ACS) Trauma Quality Improvement Program (TQIP) database. More than 700 trauma centers across the United States contribute to this database, providing more than 300 patient- and center-related variables. The TQIP database also has a deidentified unique facility key that allows analysis of individual or grouped center outcomes at a national level. The University of Arizona Institutional Review Board granted this study exemption from approval because the TQIP database contains only deidentified data.

Inclusion and Exclusion Criteria

We identified all adult (age, ≥18 years) trauma patients that presented following blunt or penetrating injury and underwent emergent (within 24 hours of admission) laparotomy as the first intervention to hemorrhage control. We excluded transfer patients and patients admitted to centers with a missing facility key.

Patient Stratification

Patients were initially stratified into two groups based on their type of injury: blunt and penetrating. Within each injury type group, patients were stratified into three subgroups based on the admitting center’s laparotomy volume: low volume (LV), medium volume (MV), and high volume (HV). Volume was defined as the injury-specific (blunt vs. penetrating) annual number of emergent laparotomies performed by the center for hemorrhage control. Low volume was defined as ≤12 cases per year (one or less cases per month), MV as 13 to 24 cases per year (one to two cases per month), and HV as ≥25 cases per year (more than two cases per month).

Outcomes

Our primary outcome measure was overall in-hospital mortality. Our secondary outcome measures were 24-hour mortality, time to laparotomy, computed tomography (CT) of the abdomen/pelvis before laparotomy, focused assessment with sonography for trauma (FAST) before laparotomy, transfusion requirements, major complications, and hospital length of stay (LOS).

Data Points

For each patient, we analyzed the following data points: demographics (age and sex), body mass index, comorbidities and Charlson Comorbidity Index (CCI), emergency department vital signs (systolic blood pressure, heart rate, and the Glasgow Coma Scale [GCS]), prehospital cardiac arrest, injury characteristics (mechanism of injury, body region–specific Abbreviated Injury Scale [AIS], and Injury Severity Score [ISS]), and the use of the resuscitative endovascular balloon occlusion of the aorta. We also analyzed data pertaining to the ACS trauma center verification level and procedure volume.

Statistical Analysis

We performed descriptive statistics for the baseline characteristics of the study sample. Continuous normally distributed variables were reported as a mean and SD. Continuous skewed variables were reported as a median and interquartile range. Categorical variables were reported as counts and proportions. To analyze the differences between the three subgroups, we used the Pearson χ2 for categorical variables, the analysis of variance test for continuous normally distributed variables, and the Mann-Whitney U test for continuous skewed variables.

To explore the predictors of in-hospital mortality in our patient cohort, we performed multivariate logistic regression analysis. The following variables were included in the model: age, sex, CCI, systolic blood pressure, GCS, prehospital cardiac arrest, ISS, ACS trauma center verification level, and injury-specific center laparotomy volume. Because the TQIP database collects data from more than 700 hospitals, we performed a hierarchical mixed-effects logistic regression model with a random hospital effect. Centers were identified using their unique facility key. This approach takes into account the hierarchical structure of the data among patients from the same center. The regression model was then assessed using the Hosmer-Lemeshow goodness-of-fit test, and the corresponding area under the receiver operating curve was represented.

We considered a p value of less than 0.05 (p < 0.05) as statistically significant. All statistical analyses were carried out using the Statistical Package for the Social Sciences (SPSS, version 26; SPSS, Inc., Armonk, NY).

RESULTS

A total of 822,571 trauma patients were originally identified, of which 8,588 underwent emergent laparotomy for either blunt (4,936; 57.5%) or penetrating injuries (3,652; 42.5%) (Fig. 1),

Figure 1
Figure 1:
Patient flow diagram.

After stratifying patients based on the admitting center’s injury-specific laparotomy volume, patients with blunt injuries were comparable in terms of demographics and comorbidities. However, patients in the HV group had more severe head injuries characterized by a lower median GCS and a higher median head-AIS. In addition, HV patients had higher median chest-AIS and overall ISS (Table 1).

TABLE 1 - Baseline Characteristics of Patients With Blunt Injuries Requiring Laparotomy
Variables LV
(n = 1,648)
MV
(n = 1,547)
HV
(n = 1,741)
p
Demographics
 Age, n (%) 0.068
  18–44 y 838 (50.8) 792 (51.2) 914 (52.5)
  45–64 y 547 (33.2) 500 (32.3) 557 (32.0)
  ≥65 y 263 (16.0) 255 (16.5) 270 (15.5)
 Male, n (%) 1,107 (67.2) 1,033 (66.8) 1,148 (65.9) 0.739
BMI, mean ± SD, kg/m2 28.3 ± 7.0 28.0 ± 7.1 28.5 ± 6.7 0.664
Comorbidities
 Hypertension, n (%) 300 (18.2) 252 (16.3) 301 (17.3) 0.360
 Diabetes mellitus, n (%) 109 (6.6) 129 (8.3) 131 (7.5) 0.179
 Congestive heart failure,  n (%) 19 (1.2) 29 (1.9) 2.5 (1.4) 0.236
 Smoking, n (%) 294 (17.8) 259 (16.7) 335 (19.2) 0.173
 CCI, median [IQR] 0 [0–2] 0 [0–2] 0 [0–2] 0.900
Vital signs
 SBP, mean ± SD, mm Hg 105 ± 34 105 ± 35 106 ± 35 0.289
 HR, mean ± SD, bpm 104 ± 31 104 ± 31 106 ± 31 0.121
 GCS, median [IQR] 14 [3–15] 14 [3–15] 13 [3–15] 0.016*
 Prehospital cardiac arrest,  n (%) 115 (7.0) 105 (6.8) 110 (6.3) 0.730
Injury characteristics
 Head-AIS, median [IQR] 0 [0–3] 0 [0–3] 2 [0–3] 0.032*
 Chest-AIS, median [IQR] 3 [1–3] 3 [1–3] 3 [2–3] 0.019*
 Abdomen-AIS,  median [IQR] 3 [2–4] 3 [2–4] 3 [2–4] 0.109
 Extremity-AIS,  median [IQR] 2 [0–3] 2 [0–3] 2 [0–3] 0.288
 ISS, n (%) <0.001*
  ≤16 231 (14.0) 213 (13.8) 180 (10.3)
  17–36 920 (55.8) 816 (52.7) 943 (54.2)
  ≥37 497 (30.2) 518 (33.5) 618 (35.5)
REBOA 18 (1.1) 21 (1.4) 67 (3.8) <0.001*
*Statistically significant.
BMI, body mass index; IQR, interquartile range; SBP, systolic blood pressure; HR, heart rate; bpm, beats per minute; REBOA, resuscitative endovascular balloon occlusion of the aorta.

In the penetrating group, more HV patients were male and belonged to the younger age category. In terms of injury characteristics, HV patients had a higher median chest- and extremity-AIS, and a higher overall ISS (Table 2).

TABLE 2 - Baseline Characteristics of Patients With Penetrating Injuries Requiring Laparotomy
Variables LV
(n = 1,359)
MV
(n = 887)
HV
(n = 1,406)
p
Demographics
 Age, n (%) 0.031*
  18–44 y 1,054 (77.6) 699 (78.8) 1,128 (80.2)
  45–64 y 250 (18.4) 165 (18.6) 249 (17.7)
  ≥65 y 55 (4.0) 23 (2.6) 29 (2.1)
 Male, n (%) 1,174 (86.4) 797 (89.9) 1,274 (90.6) 0.001*
BMI, mean ± SD, kg/m2 26.7 ± 5.6 27.1 ± 6.1 26.9 ± 5.9 0.414
Comorbidities
 Hypertension, n (%) 138 (10.2) 77 (8.7) 138 (9.8) 0.498
 Diabetes mellitus,  n (%) 60 (4.4) 32 (3.6) 56 (4.0) 0.629
 Congestive heart  failure, n (%) 9 (0.7) 2 (0.2) 4 (0.3) 0.183
 Smoking, n (%) 329 (24.2) 224 (25.3) 365 (26.0) 0.567
 CCI, median (IQR) 0 (0–2) 0 (0–2) 0 (0–2) 0.675
Vital signs
 SBP, mean ± SD,  mm Hg 111 ± 33 107 ± 34 110 ± 34 0.054
 HR, mean ± SD, bpm 103 ± 29 102 ± 30 101 ± 29 0.376
 GCS, median (IQR) 15 (12–15) 15 (12–15) 15 (13–15) 0.336
 Prehospital cardiac  arrest, n (%) 43 (3.2) 38 (4.3) 39 (2.8) 0.135
Injury characteristics
 Head-AIS, median  (IQR) 0 (0–0) 0 (0–0) 0 (0–0) 0.891
 Chest-AIS, median  (IQR) 0 (0–3) 0 (0–3) 1 (0–3) 0.001*
 Abdomen-AIS,  median (IQR) 3 (2–4) 3 (3–4) 3 (3–4) 0.058
 Extremity-AIS,  median (IQR) 0 (0–1) 0 (0–2) 0 (0–2) <0.001*
 ISS, n (%) 0.006*
  ≤16 503 (37.0) 327 (36.9) 477 (33.9)
  17–36 750 (55.2) 480 (54.1) 763 (54.3)
  ≥37 106 (7.8) 80 (9.0) 166 (11.8)
REBOA 10 (0.7) 9 (1.0) 28 (2.0) 0.010*
*Statistically significant.
BMI, body mass index; IQR, interquartile range; SBP, systolic blood pressure; HR, heart rate; bpm, beats per minute; REBOA, resuscitative endovascular balloon occlusion of the aorta.

Upon examining the distribution of centers based on the ACS verification level, the vast majority of ACS level I trauma centers were either HV or MV centers (82% for blunt and 76% for penetrating injuries), while the majority of ACS level II trauma centers were LV (60% for blunt and 71% for penetrating injuries). All ACS level III trauma centers were LV centers (100% for blunt and 100% for penetrating injuries) (Fig. 2).

Figure 2
Figure 2:
Distribution of centers based on ACS verification level and laparotomy volume for blunt and penetrating injuries.

Overall, there were 2,060 (24.0%) mortalities, with a higher mortality rate in the blunt group (1,551; 31.4%) compared with the penetrating group (509; 14.0%).

On univariate analysis of outcomes, in-hospital and 24-hour mortality rates for blunt injury patients were lower in HV (29.2% and 13.4%, respectively) centers compared with MV (32.4% and 16.5%) and LV (32.9% and 15.8%) centers (p = 0.041 and p = 0.037). Time to laparotomy was significantly shorter in HV (72 minutes) centers compared with MV (81 mins) and LV (94 minutes) centers (p < 0.001). Findings were similar for penetrating injury patients, where in-hospital and 24-hour mortality rates were lower in HV (12.1% and 7.3% respectively) centers compared with MV (14.3% and 9.5%) and LV (15.6% and 9.8%) centers (p = 0.027 and p = 0.042). Time to laparotomy in those patients was also significantly shorter in HV (35 minutes) centers compared with MV (46 minutes) and LV (51 minutes) centers (p < 0.001) (Table 3).

TABLE 3 - Univariate Analysis of Outcomes
Outcome LV MV HV p
Blunt injuries (n = 4,936) 1,648 (33.4) 1,547 (31.3) 1,741 (35.3)
 In-hospital mortality, n (%) 542 (32.9) 501 (32.4) 508 (29.2) 0.041*
 24-h Mortality, n (%) 260 (15.8) 255 (16.5) 234 (13.4) 0.037*
 Time to laparotomy, median (IQR), min 94 (56–158) 81 (49–145) 72 (41–144) <0.001*
 Abdomen/pelvis CT, n (%) 916 (55.6) 808 (52.2) 870 (50.0) 0.005*
 FAST examination, n (%) 1,035 (62.8) 983 (63.5) 1,208 (73.5) <0.001*
 Transfusion requirements
  PRBC, U/4 h, median (IQR) 5 (2–10) 5 (3–11) 6 (3–11) 0.217
  PRBC, U/24 h, median (IQR) 6 (3–13) 7 (4–14) 7 (4–14) 0.689
  Plasma, U/4 h, median (IQR) 3 (0–7) 3 (1–8) 4 (1–8) 0.021*
  Plasma, U/24 h, median (IQR) 4 (1–9) 4 (2–10) 4 (2–10) 0.147
  Platelets, U/4 h, median (IQR) 1 (0–2) 1 (0–2) 1 (0–2) 0.633
  Platelets, U/24 h, median (IQR) 1 (0–3) 1 (0–2) 1 (0–2) 0.364
 Major complications, n (%)
  Cardiac 213 (12.9) 164 (10.6) 214 (12.3) 0.114
  VTE 113 (6.9) 145 (9.4) 118 (6.8) 0.007*
  Renal 123 (7.5) 97 (6.3) 121 (7.0) 0.412
  Pulmonary 135 (8.2) 117 (7.6) 143 (8.2) 0.744
  Sepsis 54 (3.3) 50 (3.2) 61 (3.5) 0.896
 Hospital LOS, median (IQR), d 10 (4–20) 11 (4–21) 11 (5–21) 0.098
Penetrating injuries (n = 3,652) 1,359 (37.2) 887 (24.3) 1,406 (38.5)
 In-hospital mortality, n (%) 212 (15.6) 127 (14.3) 170 (12.1) 0.027*
 24-h Mortality, n (%) 133 (9.8) 84 (9.5) 102 (7.3) 0.042*
 Time to laparotomy, median (IQR), min 51 (38–69) 46 (33–63) 35 (24–52) <0.001*
 Abdomen/pelvis CT, n (%) 337 (24.8) 176 (19.8) 257 (18.3) <0.001*
 FAST examination, n (%) 560 (41.2) 326 (36.8) 535 (38.1) 0.075
 Transfusion requirements
  PRBC, U/4 h, median (IQR) 4 (2–9) 4 (2–10) 5 (2–10) 0.089
  PRBC, U/24 h, median (IQR) 5 (2–10) 5 (2–12) 5 (2–12) 0.079
  Plasma, U/4 h, median (IQR) 2 (0–6) 2 (0–7) 3 (0–7) 0.111
  Plasma, U/24 h, median (IQR) 3 (0–6) 3 (0–8) 4 (1–9) <0.001*
  Platelets, U/4 h, median (IQR) 0 (0–1) 0 (0–1) 0 (0–1) 0.719
  Platelets, U/24 h, median (IQR) 1 (0–2) 0 (0–2) 0 (0–2) 0.498
 Major complications, n (%)
  Cardiac 94 (6.9) 52 (5.9) 89 (6.3) 0.597
  VTE 98 (7.2) 58 (6.5) 105 (7.5) 0.697
  Renal 74 (5.4) 55 (6.2) 51 (3.6) 0.012*
  Pulmonary 59 (4.3) 38 (4.3) 73 (5.2) 0.475
  Sepsis 43 (3.2) 30 (3.4) 38 (2.7) 0.617
 Hospital LOS, median (IQR), d 9 (5–17) 10 (6–17) 10 (5–18) 0.112
*Statistically significant.
IQR, interquartile range; PRBC, packed red blood cell; VTE, venous thromboembolism.

Blunt injury patients at HV centers were less likely to undergo abdomen/pelvis CT before laparotomy (50.0% vs. 52.2% vs. 55.6%; p = 0.005) but more likely to undergo a FAST examination (73.5% vs. 63.5% vs. 62.8%; p < 0.001) compared with patients at MV and LV centers. Similarly, penetrating injury patients at HV centers were less likely to undergo abdomen/pelvis CT before laparotomy (18.3% vs. 19.8% vs. 24.8%; p < 0.001) compared with patients at MV and LV centers; however, rates of FAST examination were not different (38.1% vs. 36.8% vs. 41.2%; p = 0.075) between the three groups (Table 3).

Overall, no significant differences in transfusion requirements, major complications, and hospital LOS were noted between the three volume groups in both blunt and penetrating injury patients (Table 3).

On multivariate analysis of outcomes, admission to HV centers was associated with lower odds of in-hospital mortality for both blunt (odds ratio, 0.74 [0.59–0.93]; p = 0.011) and penetrating (0.86 [0.77–0.96]; p = 0.023) injury patients. Predictors of in-hospital mortality were found to be increasing age, higher CCI, prehospital cardiac arrest, and increasing ISS. American College of Surgeons trauma center verification level was not associated with in-hospital mortality. (Table 4) (Fig. 3)

TABLE 4 - Multivariate Analysis of Outcomes: Predictors of In-hospital Mortality
Variables Blunt Injuries* Penetrating Injuries**
aOR (CI) p aOR (CI) p
Age
 18–44 y Ref. Ref.
 45–64 y 1.71 (1.42–2.06) <0.001† 1.10 (0.80–1.51) 0.554
 ≥65 y 4.86 (3.83–6.18) <0.001† 4.59 (2.63–8.01) <0.001†
Male 1.14 (0.96–1.36) 0.144 1.08 (0.74–1.58) 0.693
CCI (every 1-U increase) 1.42 (1.28–1.57) <0.001† 1.36 (1.08–1.70) 0.008†
SBP (every 10 mm Hg increase) 0.97 (0.95–0.99) 0.005† 0.92 (0.89–0.95) <0.001†
GCS (every 1-U increase) 0.86 (0.84–0.87) <0.001† 0.86 (0.84–0.88) <0.001†
Prehospital cardiac arrest 4.70 (3.33–6.62) <0.001† 7.37 (4.14–13.13) <0.001†
ISS
 ≤16 Ref. Ref.
 17–36 1.48 (1.10–2.00) 0.010† 2.52 (1.89–3.42) <0.001†
 ≥37 3.10 (2.26–4.24) <0.001† 6.87 (4.66–10.11) <0.001†
ACS center level
 Level I Ref. Ref.
 Level II 0.92 (0.73–1.14) 0.427 1.10 (0.79–1.53) 0.575
 Level III 1.30 (0.55–3.10) 0.549 1.27 (0.93–1.73) 0.129
Center volume
 Low (≤12 cases) Ref. Ref.
 Medium (13–24 cases) 0.84 (0.67–1.05) 0.119 0.99 (0.92–1.07) 0.208
 High (≥25 cases) 0.74 (0.59–0.93) 0.011† 0.86 (0.77–0.96) 0.023†
*Hosmer-Lemeshow goodness-of-fit test: p = 0.341; AUROC, 0.810 (0.796–0.825).
**Hosmer-Lemeshow goodness-of-fit test: p = 0.443; AUROC, 0.802 (0.787–0.816).
†Statistically significant.
aOR, adjusted odds ratio; AUROC, area under the receiver operating curve; CI, confidence interval; Ref., reference; SBP, systolic blood pressure.

Figure 3
Figure 3:
Receiver operating characteristic curves for predictive models of in-hospital mortality following blunt and penetrating injuries.

DISCUSSION

The findings of our study indicate that a trauma center’s injury-specific laparotomy volume is a suitable proxy for patient outcomes. We found that admission to HV trauma centers for emergent laparotomy after either blunt or penetrating injury was associated with improved outcomes. This was evident as improved in-hospital and 24-hour mortality rates, with shorter times to hemorrhage control. We also found no differences in transfusion requirements, rates of major complications, and hospital LOS between LV, MV, and HV trauma centers.

Our study is unique in examining the relationship between a trauma center’s injury-specific laparotomy volume and outcomes in blunt and penetrating trauma patients requiring emergent laparotomy. Our patient cohort is that of severely injured patients with a diverse set of injuries. These patients often require complex care from a multidisciplinary team. While our study was not designed to measure the efficacy of the component parts of a successful trauma system, one possible explanation for our finding is that the experience derived from repetitively performing similar operative interventions enhances outcomes on repeat exposure. The term institutional memory, addressing the combined human and associated infrastructure, has been coined by Tepas et al.19 in their study examining the impact of volume on outcomes of traumatic brain injury (TBI) patients. In line with the findings of our study, the authors reported decreased mortality rates among TBI patients admitted to centers with a high volume of TBI patients. The authors also found that such patients were more likely to be discharged home or to rehabilitation instead of the less preferable and more costly skilled nursing facility.19 In a similar study design, Matsushima et al.20 examined the volume-outcome relationship in the geriatric trauma population. The authors found that trauma centers with lower volumes of geriatric trauma care were associated with higher rates of in-hospital mortality, major complications, and failure to rescue and thereby concluded that a differentiated pathway of care for specific high-risk groups can improve outcomes.20 In our study, we found no differences in transfusion requirements, rates of major complications, and hospital LOS between the different groups, despite the differences in mortality. A number of factors may have contributed to this finding. First, patients admitted to HV centers had significantly worse injuries given the higher overall ISS and particularly worse head injuries as evidenced by the lower GCS scores and higher head-AIS scores in blunt injury patients. Such injury patterns have been shown to be associated with increased transfusion requirements,25 major complications,26 and hospital LOS.27 Second, given the significantly higher mortality rates in patients admitted to LV and MV trauma centers, there is a potential for survival bias that might have skewed the results of the secondary outcomes.

Studies that examined the relationship between overall trauma center volume and outcomes have exhibited conflicting results. Some investigators have illustrated an inverse relationship between mortality and both hospital volume and per-surgeon volume of severely injured trauma patients, in line with the findings of our study.28–32 Other investigators have demonstrated that increased volume does not bestow a survival benefit and is in fact associated with higher adverse outcomes.13–18 One study revealed that two inflection points exist for the volume-outcome relationship, whereby after a certain threshold, overloading the trauma center with minimally injured patients may hinder the overall effectiveness of the system.1 The heterogeneity in the results from these studies may be attributed to the differing statistical analyses, source populations, databases, and geographic differences, in addition to institutional and provider variations. However, the most important distinction is perhaps the definition of volume used, which in all of these studies was based on the overall trauma patient volume. Recurrent exposure to similar operative cases affects patient care at many levels, from more efficient initial evaluations, to judicious use of targeted diagnostic tests, and timely availability of operating rooms. Although this recurrent exposure is possible when considering overall trauma center volume, it is most amplified and thereby discernible when adopting a procedure-specific volume definition. This was evident in our study, where time to laparotomy was significantly shorter in HV trauma centers. One possible explanation for this is the lower utility of CT scan before laparotomy in both blunt and penetrating injury patients admitted to HV trauma centers, with more frequent use of FAST examination in cases of blunt injury. This is in line with multiple studies that have evaluated the role of sonography in predicting the need for laparotomy in such patients, with the decision to proceed with CT imaging justified in hemodynamically stable patients with indeterminate results.33–36 In addition, in their retrospective analysis of patients from the Prospective Observational Multicenter Major Trauma Transfusion, Schreiber et al.37 found that in patients with a positive FAST examination requiring laparotomy, increasing time to operation was associated with decreased survival, with much of this delay being accounted for by the time needed to perform a CT scan.

In nontrauma cases, studies have consistently shown that concentrating a large number of procedures at a single institution lowers mortality rates and improves outcomes.2–10 Moreover, these studies have shown that the beneficial effects of this relationship are most pronounced in cases of complex procedures.3,8,9 However, a number of factors may object extrapolating the beneficial effects of high patient volumes in the elective care setting to the trauma care setting. Since the timing of injuries is never predictable, rescheduling a procedure in case of suboptimal circumstances is not possible. In addition, the care for the trauma patient encompasses a wide array of diagnostic and interventional procedures that are time sensitive. Therefore, although the skill of the individual surgeon is important, outcomes in such trauma patients often rely on a well-designed and an efficiently functioning system of care.24 Research on health care team effectiveness shows that perceived team effectiveness is improved as a result of repeated experiences of similar case scenarios between different clinical personnel involved in the team decision making, ultimately resulting in improvement in patient outcomes and organizational effectiveness.38,39 Similarly, when assessing the importance of individual surgical experience compared with trauma system infrastructure, Haut et al.40 concluded that the system effect of a structured trauma program outweighs any potential benefits of individual surgeon experience in the care of trauma patients. This further ascertains the paramount role of case-specific high-volume experience on patient outcomes.

In light of the findings of our study and the existing body of literature, we believe that the utilization of novel and targeted ways to assess trauma center performance is warranted. Further exploration of center-specific operative volume and their associations with risk-adjusted outcomes, rather than overall volume per se, is merited. Ultimately, this may guide policy makers to use center-specific operative volume as a quality assessment metric to improve patient outcomes, through feedback and payment incentives, which will promote institutional quality improvement efforts.41 This article adds to the growing literature that high volume is associated with better outcomes for surgical patients, which in turn supports regionalizing patient care. Alternatively, while some studies suggest that referral of patients to high volume hospitals has the potential to reduce mortality, additional studies are needed to determine the feasibility of selective referral and to examine the consequences of such initiatives.42

This study is not without limitations. First, given the retrospective nature of our database, the observed reduction in mortality in HV trauma centers can only be interpreted as an association and not necessarily a causation. Second, there was no randomization of patients as to which operative-volume center they were admitted to, which can lead to residual confounding bias. Third, we do not have data on the exact cause of death. However, since the aim of our study was to assess patient outcomes based on the admitting center volume, the exact cause of death would have been valuable to explore but not necessary for our analysis. Fourth, identifying rates of negative laparotomy and rates of missed laparotomy was not possible, for such diagnosis codes are not available in the International Classification of Diseases, 10th Revision, coding system. Overall, this study was derived from the ACS-TQIP database, a well-established database in the field of trauma research. It also has the strength of incorporating a large, multi-institutional sample of trauma patients in the United States, making it nationally representable.

CONCLUSIONS

Our study demonstrates that HV trauma centers have more favorable patient outcomes with respect to mortality following emergent laparotomy for both blunt and penetrating injury. We believe that the findings of our study ascertain the validity of using case-specific hospital volumes as a proxy for outcomes. These results will help focus the discussion on how to allocate trauma patients to appropriate receiving centers.

AUTHORSHIP

A.T., M.C., Michael Ditillo, S.A., and B.J. designed this study. M.C., M.K., Michael Ditillo, L.B., and N.K. searched the literature. A.T., S.A., M.C., N.K., and L.B. collected the data. M.C., L.B., Molly Douglas, S.A., and B.J. analyzed the data. All authors participated in data interpretation and article preparation.

DISCLOSURE

The authors declare no conflicts of interest.

REFERENCES

1. Tepas JJ, Patel JC, DiScala C, Wears RL, Veldenz HC. Relationship of trauma patient volume to outcome experience: can a relationship be defined? J Trauma Acute Care Surg. 1998;44(5):827–831.
2. Birkmeyer JD, Finlayson SR, Tosteson AN, Sharp SM, Warshaw AL, Fisher ES. Effect of hospital volume on in-hospital mortality with pancreaticoduodenectomy. Surgery. 1999;125(3):250–256.
3. Begg CB, Cramer LD, Hoskins WJ, Brennan MF. Impact of hospital volume on operative mortality for major cancer surgery. JAMA. 1998;280(20):1747–1751.
4. Hannan EL, Kilburn H Jr., Bernard H, O’Donnell JF, Lukacik G, Shields EP. Coronary artery bypass surgery: the relationship between inhospital mortality rate and surgical volume after controlling for clinical risk factors. Med Care. 1991;29:1094–1107.
5. Hannan EL, Popp AJ, Tranmer B, Fuestel P, Waldman J, Shah D. Relationship between provider volume and mortality for carotid endarterectomies in New York state. Stroke. 1998;29(11):2292–2297.
6. Wen SW, Simunovic M, Williams JI, Johnston KW, Naylor CD. Hospital volume, calendar age, and short term outcomes in patients undergoing repair of abdominal aortic aneurysms: the Ontario experience, 1988-92. J Epidemiol Community Health. 1996;50(2):207–213.
7. Solomon RA, Mayer SA, Tarmey JJ. Relationship between the volume of craniotomies for cerebral aneurysm performed at New York state hospitals and in-hospital mortality. Stroke. 1996;27(1):13–17.
8. Luft HS, Bunker JP, Enthoven AC. Should operations be regionalized? The empirical relation between surgical volume and mortality. N Engl J Med. 1979;301(25):1364–1369.
9. Birkmeyer JD, Siewers AE, Finlayson EV, Stukel TA, Lucas FL, Batista I, Welch HG, Wennberg DE. Hospital volume and surgical mortality in the United States. N Engl J Med. 2002;346(15):1128–1137.
10. Dimick JB, Stanley JC, Axelrod DA, Kazmers A, Henke PK, Jacobs LA, Wakefield TW, Greenfield LJ, Upchurch GR Jr. Variation in death rate after abdominal aortic aneurysmectomy in the United States: impact of hospital volume, gender, and age. Ann Surg. 2002;235(4):579–585.
11. Halaszynski TM, Juda R, Silverman DG. Optimizing postoperative outcomes with efficient preoperative assessment and management. Crit Care Med. 2004;32(4):S76–S86.
12. American College of Surgeons. Verification Review Consultation Frequently Asked Questions 633 N Saint Clair Street Chicago, IL 60611-3295: American College of Surgeons 2020. Available at: https://www.facs.org/qualityprograms/trauma/tqp/centerprograms/vrc/faq/standards#:~:text=CD%202%2D4%3A%20A%20Level,a%20Level%20I%20trauma%20center. Accessed June 15, 2020.
13. Richardson JD, Schmieg R, Boaz P, Spain DA, Wohltmann C, Wilson MA, Carrillo EH, Miller FB, Fulton RL. Impact of trauma attending surgeon case volume on outcome: is more better? J Trauma Acute Care Surg. 1998;44(2):266–272.
14. Waddell TK, Kalman PG, Goodman SJ, Girotti MJ. Is outcome worse in a small volume Canadian trauma Centre? J Trauma Acute Care Surg. 1991;31(7):958–961.
15. Cooper A, Hannan EL, Bessey PQ, Farrell LS, Cayten CG, Mottley L. An examination of the volume-mortality relationship for New York state trauma centers. J Trauma Acute Care Surg. 2000;48(1):16.
16. Norwood S, Myers MB. Outcomes following injury in a predominantly rural-population—based trauma center. Arch Surg. 1994;129(8):800–805.
17. Allen DM, Hicks G, Bota GW. Outcomes after severe trauma at a northern Canadian regional trauma Centre. Can J Surg. 1998;41(1):53.
18. Margulies DR, Cryer HG, McArthur DL, Lee SS, Bongard FS, Fleming AW. Patient volume per surgeon does not predict survival in adult level I trauma centers. J Trauma Acute Care Surg. 2001;50(4):597–603.
19. Tepas JJ III, Pracht EE, Orban BL, Flint LM. High-volume trauma centers have better outcomes treating traumatic brain injury. J Trauma Acute Care Surg. 2013;74(1):143–148.
20. Matsushima K, Schaefer EW, Won EJ, Armen SB, Indeck MC, Soybel DI. Positive and negative volume-outcome relationships in the geriatric trauma population. JAMA Surg. 2014;149(4):319–326.
21. Duchesne JC, Kimonis K, Marr AB, Rennie KV, Wahl G, Wells JE, Islam TM, Meade P, Stuke L, Barbeau JM. Damage control resuscitation in combination with damage control laparotomy: a survival advantage. J Trauma Acute Care Surg. 2010;69(1):46–52.
22. Joseph B, Zangbar B, Pandit V, Vercruysse G, Aziz H, Kulvatunyou N, Wynne J, O’Keeffe T, Tang A, Friese RS. The conjoint effect of reduced crystalloid administration and decreased damage-control laparotomy use in the development of abdominal compartment syndrome. J Trauma Acute Care Surg. 2014;76(2):457–461.
23. Joseph B, Zangbar B, Pandit V, Kulvatunyou N, Haider A, O’Keeffe T, Khalil M, Tang A, Vercruysse G, Gries L. Mortality after trauma laparotomy in geriatric patients. J Surg Res. 2014;190(2):662–666.
24. London JA, Battistella FD. Is there a relationship between trauma center volume and mortality? J Trauma Acute Care Surg. 2003;54(1):16–25.
25. Eastridge BJ, Malone D, Holcomb JB. Early predictors of transfusion and mortality after injury: a review of the data-based literature. J Trauma Acute Care Surg. 2006;60(6):S20–S25.
26. Mondello S, Cantrell A, Italiano D, Fodale V, Mondello P, Ang D. Complications of trauma patients admitted to the ICU in level I academic trauma centers in the United States. Biomed Res Int. 2014;2014:1–7.
27. Tardif PA, Moore L, Boutin A, Dufresne P, Omar M, Bourgeois G, Bonaventure PL, Kuimi BL, Turgeon AF. Hospital length of stay following admission for traumatic brain injury in a Canadian integrated trauma system: a retrospective multicenter cohort study. Injury. 2017;48(1):94–100.
28. Konvolinka CW, Copes WS, Sacco WJ. Institution and per-surgeon volume versus survival outcome in Pennsylvania’s trauma centers. Am J Surg. 1995;170(4):333–340.
29. Nathens AB, Jurkovich GJ, Maier RV, Grossman DC, MacKenzie EJ, Moore M, Rivara FP. Relationship between trauma center volume and outcomes. JAMA. 2001;285(9):1164–1171.
30. Smith RF, Frateschi L, Sloan EP, Campbell L, Krieg R, Edwards LC, Barrett JA. The impact of volume on outcome in seriously injured trauma patients: two years’ experience of the Chicago Trauma System. J Trauma Acute Care Surg. 1990;30(9):1066–1075.
31. Pasquale MD, Peitzman AB, Bednarski J, Wasser TE. Outcome analysis of Pennsylvania trauma centers: factors predictive of nonsurvival in seriously injured patients. J Trauma Acute Care Surg. 2001;50(3):465–474.
32. Bell TM, Boustany KC, Jenkins PC, Zarzaur BL. The relationship between trauma center volume and in-hospital outcomes. J Surg Res. 2015;196(2):350–357.
33. Peytel E, Menegaux F, Cluzel P, Langeron O, Coriat P, Riou B. Initial imaging assessment of severe blunt trauma. Intensive Care Med. 2001;27(11):1756–1761.
34. Lee BC, Ormsby EL, McGahan JP, Melendres GM, Richards JR. The utility of sonography for the triage of blunt abdominal trauma patients to exploratory laparotomy. AJR Am J Roentgenol. 2007;188(2):415–421.
35. McKenney MG, Martin L, Lentz K, Lopez C, Sleeman D, Aristide G, Kirton O, Nunez D, Najjar R, Namias N. 1,000 consecutive ultrasounds for blunt abdominal trauma. J Trauma Acute Care Surg. 1996;40(4):607–612.
36. Udobi KF, Rodriguez A, Chiu WC, Scalea TM. Role of ultrasonography in penetrating abdominal trauma: a prospective clinical study. J Trauma Acute Care Surg. 2001;50(3):475–479.
37. Barbosa RR, Rowell SE, Fox EE, Holcomb JB, Bulger EM, Phelan HA, Alarcon LH, Myers JG, Brasel KJ, Muskat P. Increasing time to operation is associated with decreased survival in patients with a positive FAST exam requiring emergent laparotomy. J Trauma Acute Care Surg. 2013;75(1 0 1):S48–S52.
38. Lemieux-Charles L, McGuire WL. What do we know about health care team effectiveness? A review of the literature. Med Care Res Rev. 2006;63(3):263–300.
39. Gagliardi AR, Nathens AB. Exploring the characteristics of high-performing hospitals that influence trauma triage and transfer. J Trauma Acute Care Surg. 2015;78(2):300–305.
40. Haut ER, Chang DC, Hayanga AJ, Efron DT, Haider AH, Cornwell EE. Surgeon-and system-based influences on trauma mortality. Arch Surg. 2009;144(8):759–764.
41. Hanna K, Haddadin Z, Sakran J, Zeeshan M, Asmar S, Kulvatunyou N, Tang A, Northcutt A, Joseph B. Barriers to improving health care value in emergency general surgery: a nationwide analysis. J Trauma Acute Care Surg. 2020;89(2):289–300.
42. Dudley RA, Johansen KL, Brand R, Rennie DJ, Milstein A. Selective referral to high-volume hospitals: estimating potentially avoidable deaths. JAMA. 2000;283(9):1159–1166.

DISCUSSION

BRIAN J. EASTRIDGE, M.D. (San Antonio, Texas): Thank you, Dr. Winchell. Good morning, members and guests. First, I’d like to thank the AAST for the privilege of discussing this paper. I would also like to thank Dr. Tang, Dr. Joseph and colleagues for an insightful and well-written manuscript that was prepared well in advance for review.

The premise of the work was to elicit the relationship between trauma center laparotomy volume and outcomes in blunt/penetrating trauma patients.

Data from the TQIP data repository identified 8,500 patients who underwent emergency laparotomy during the one-year study period. High volume centers demonstrated decreased time to the OR and improved in-hospital mortality for both blunt and penetrating trauma patients.

The authors conclude that the case-specific hospital volumes may serve as a valid proxy metric for outcomes to support the regionalization of trauma care.

To align these data with it and the conclusions I have several questions for the authors.

The trauma centers were stratified based on their blunt and penetrating emergent laparotomy volumes with higher being greater than or equal to 25; medium, 13 to 24; and low, less than 12.

What was your rationale to utilize these specific case volume thresholds for operative experience to assess the volume outcome relationships in your analysis?

Question Number 2. Your study inclusion criteria were all adult trauma patients that presented following blunt and penetrating injury that underwent emergency laparotomy as the first intervention for hemorrhage control.

How did you identify or qualify these laparotomies done for hemorrhage control within the data repository? And as a related question, since your definition of emergent laparotomy included those laparotomies up to 24 hours, were there any high outliers for time associated with any of the particular volume strata?

Specifically related to the concept of laparotomies done for hemorrhage control, were you able to determine if the deaths in your study were caused by hemorrhage, other primary traumatic pathology, or the latent effects of injury?

And, finally, my fourth question is on analysis admission of trauma patients requiring emergent laparotomy to high-volume blunt and high-volume penetrating centers was independently associated with in-hospital mortality.

Likewise, these high-volume blunt and penetrating trauma centers had shorter time for admission to laparotomy when compared to the medium- and low-volume centers.

Your data demonstrates a subtle but statistically significant trauma center effect. Your title “Regionalization of Trauma Care by Operative Experience: Does Volume of Emergent Laparotomy Matter” implies that your analysis may have sort of greater systemic value at the trauma system level.

Did you or could you consider integrating time from injury to the OR in patients requiring emergency laparotomy as this may be a more valid metric to assess the trauma system and potentially its implications, your data’s implications, in the regionalization of trauma care.

I think this paper has promise to add to the body of literature that increased case volume is associated with improved outcomes and, thus, further substantiate the value of regionalization of trauma care. Once again I would like to thank the Program Committee and the AAST for the opportunity to discuss this fine paper.

ANDREW TANG, M.D. (Tucson, ARIZONA): Thank you, Dr. Winchell, and thank you, Dr. Eastridge, for the privilege.

In answer to your question, the first one was the rationale for the way we stratified the three categories. When we looked at how to go about doing this we wanted to create parameters that were clinically meaningful and, at the same time, easily relatable to trauma surgeons.

And so these numbers essentially break down to less than one case per month, one to two cases per month, or more than two cases per month by any trauma center.

It just so happens that this categorization resulted in a somewhat homogenous distribution of patients amongst the three evaluated groups. So that’s how we came about with our stratification algorithm.

Your second question, Dr. Eastridge, was how did we identify and qualify the laparotomies done for hemorrhage control within the data set. TQIP happens to contain a specific variable titled “surgery for hemorrhage control.”

And it is defined as the first type of surgery for hemorrhage control within the first 24 hours of emergency department/hospital arrival. We used this variable to identify our cohort of patients.

Your follow-up question to that was very interesting, were there any – and how did the high outliers in terms of time to laparotomy affect our outcome.

We identified that 95 percent of laparotomies were actually performed within the 4-hour window after patient arrival, and so there was no statistical difference between the three volume groups.

Your next question is were we able to determine the cause of death, whether it was related to hemorrhage or due to some other pathologies. Unfortunately, the cause of death is not granular enough or specific enough in the TQIP data base. We do know from experience and generalized data that the most common causes of death within 24 hours following a traumatic event are hemorrhage or traumatic brain injury. Our study did show a lower mortality rate in the high-volume trauma centers.

When we looked at our data set the median head AIS was one in our blunt subset and zero in our penetrating subset. So that leaves us to safely assume that the vast majority of patients who passed away within the first 24 hours were secondary to hemorrhage.

Dr. Eastridge, your last question was very interesting. You talked about would the time of injury to laparotomy be a more meaningful way to assess a trauma system’s performance. We totally agree with you.

In fact, using this particular way of assessing, again, the time of injury to laparotomy, would be a more precise measure of the regional trauma system performance rather than an individual trauma center’s performance.

Unfortunately, the time of injury is not something that’s well captured in the TQIP data base. In fact when we looked at our data 30 percent of prehospital data was missing. And so that lends itself to some challenges with analysis.

But, nonetheless, I do suspect that this can be an excellent collaborative effort between the various trauma centers within a particular regional trauma system where the prehospital data, such as time of injury, can be more specifically and accurately captured such that it will allow us to perform a more accurate and meaningful analysis of exactly the point you are making, time of injury to laparotomy.

Thank you so much.

HASAN B. ALAM, M.D. (Chicago, Illinois): How did you pick the cut-off numbers for the low, moderate, and high volumes for trauma laparotomy?

Laparotomy outcomes may reflect more of an individual skill set than hospital performance. Have you looked at the volumes of the individual surgeons?

PATRICK REILLY, M.D. (Philadelphia, Pennsylvania): Do you have PATI score data, or maybe NISS data, for the penetrating trauma cohort rather than just AIS/ISS data?

DAVID H. LIVINGSTON, M.D. (Newark, New Jersey): Did you do the analysis for those patients with shock, e.g., blood pressure less than 90, or those requiring or than four units of blood or MTP? It is these patients who die of hemorrhage.

ANDREW TANG, M.D. (Tucson, ARIZONA): That’s a great question. We did not perform that particular sub-analysis but I do think that that would be a very, very interesting and even more granular and specific way of assessing performance because, after all, those are the most critical subset of patients. This is something we can look into.

MATTHEW MARTIN, M.D. (San Diego, California): Alternative hypothesis – centers that have a lower threshold for laparotomy and operate on many who don’t need it have better outcomes. Did you look at the disease pathology and need for surgery, as well as negative laparotomy rates?

SAMIR FAKHRY, M.D. (Nashville, Tennessee): Very nice study and presentation. Do you have any data on the volumes of blood transfusion for the laparotomy patients and if so did they differ by center volume?

D’ANDREA K. JOSEPH, M.D. (Roslyn, New York): Thank you for an excellent presentation. Your results are not unexpected and supports an argument for decreasing the number of trauma centers in as yet defined radius. However, how do you reconcile this with the College’s current requirements for verification? Is the COT incorrect? Is it possible that your data could be refined further to look at shock/alternative intervention, angio, etc.?

ROBERT J. WINCHELL, M.D. (New York, New York): And so, in the last 30 seconds here, I take the prerogative of the moderator to tell you that my look at your data says that high and low volume are surrogates for Level I designation and did you do a subgroup analysis within high-, medium- and low-volume Level I centers?

ANDREW TANG, M.D. (Tucson, ARIZONA): We did, in fact, look at this specific question. Our data showed that a trauma center’s laparotomy volume did not necessarily correlate with their ACS trauma center designation. That is because there are high volume centers that are ACS Level II, and likewise, Level I centers that have low laparotomy volumes. But in general, more ACS level Is tend to be centers with higher laparotomy volumes. Thank you, Dr. Winchell.

Keywords:

Laparotomy; volume; blunt; penetrating; trauma

Copyright © 2020 American Association for the Surgery of Trauma.