- Email: [email protected]
Trauma Recidivism and Mortality Following Violent Injuries in Young Adults
j o u r n a l o f s u r g i c a l r e s e a r c h m a y 2 0 1 9 ( 2 3 7 ) 1 4 0 e1 4 7
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.JournalofSurgicalResearch.com
Association for Academic Surgery
Trauma Recidivism and Mortality Following Violent Injuries in Young Adults Angela M. Kao, MD,a Kathryn A. Schlosser, MD,a Michael R. Arnold, MD,a Kevin R. Kasten, MD,a Paul D. Colavita, MD,a Bradley R. Davis, MD,a Ronald F. Sing, DO,b and B. Todd Heniford, MDa,* a
Division of Gastrointestinal and Minimally Invasive Surgery, Department of Surgery, Carolinas Medical Center, Charlotte, North Carolina b Division of Trauma and Acute Care Surgery, Department of Surgery, Carolinas Medical Center, Charlotte, North Carolina
article info
abstract
Article history:
Background: Trauma recidivism accounts for approximately 44% of emergency department
Received 2 March 2018
admissions and remains a significant health burden with this patient cohort carrying
Received in revised form
higher rates of morbidity and mortality.
16 August 2018
Methods: A level 1 trauma center registry was queried for patients aged 18-25 y presented
Accepted 4 September 2018
between 2009 and 2015. Patients with nonaccidental gunshot wounds, stab wounds, or blunt assault-related injuries were categorized as violent injuries. Primary outcomes included mortality and recidivism, which were defined as patients with two unrelated
Keywords:
traumas during the study period. Hospital records and the Social Security Death Index
Violence
were used to aid in outcomes.
Intentional injury
Results: A total of 6484 patients presented with 1215 (18.7%) sustaining violent injuries
Firearm-related injury
(87.4% male, median age 22.2 y). Mechanism of violent injuries included 64.4% gunshot
Penetrating trauma
wound, 21.1% stab, and 14.8% blunt assault. Compared with nonviolent injuries, violent
Homicide
injury patients had increased risk of mortality (9.3% versus 2.1%, P < 0.0001). Out-of-
Young adult
hospital mortality was 2.6% (versus 0.5% nonviolent, P < 0.0005), with an average time to
Teenage
death being 6.4 mo from initial injury. Recidivism was 24.9% with mean time to second
Gunshot wound
violent injury at 31.9 21.0 mo; 14.9% had two trauma readmissions, and 8.0% had 3.
Stab wound
Ninety percent of subsequent injuries occurred within 5 y, with 19.1% in the first year.
Assault
Conclusions: The burden of injury after violent trauma extends past discharge as patients have significantly higher mortality rates following hospital release. Over one-quarter present with a second unrelated trauma or death. Improved medical, psychological, and social collaborative treatment of these high-risk patients is needed to interrupt the cycle of violent injury. Published by Elsevier Inc.
Accepted for oral presentation at the 13th Annual Academic Surgical Congress, January 30, 2018, Jacksonville, Florida. * Corresponding author. Carolinas Medical Center, 1025 Morehead Medical Drive, Suite 300, Charlotte, NC 28204. Tel.: þ1 704 355 3168; fax: þ1 704 355 3168. E-mail address: [email protected] (B.T. Heniford). 0022-4804/$ e see front matter Published by Elsevier Inc. https://doi.org/10.1016/j.jss.2018.09.006
kao et al recidivism after violent injuries
Introduction Violent trauma accounts for over 57,000 deaths and 1.4 million emergency department (ED) visits annually, including over 700,000 admissions in patients aged between 10 and 24 y.1,2 An estimated 44% of trauma admissions occur in individuals with a prior history of injury-related admissions, a phenomenon known as trauma recidivism.3-5 Violent penetrating injuries requiring hospitalization and surgery incur a significant cost, and recurrent injuries increase this cost burden. In 2010, estimated medical costs for injuries related to interpersonal violence totaled $8.5 billion, with 40% of costs absorbed by hospitals and government payers.6 Studies have also shown victims of firearm assault are at increased risk of violencerelated crimes and mortality after hospitalization, with firearm violence accounting for over 36,000 deaths and 81,000 nonfatal injuries annually.3-5,7 Firearm-related injuries account for 91% of homicides and remain the second leading cause of death in young adults.1,2 Identifying and interrupting cycles of violence in high-risk cohorts, and thus preventing future injury or death, is an excellent target for potential hospital-based intervention. Despite the increased risk of repeat victimization, the longterm outcomes of young adults sustaining violence-related trauma have not been well studied. Although firearm-related injuries remain at the center of national conversation, they account for only 9.2% of all violent assaults among individuals aged 18-24 y.8 Reported rates of violent injury recidivism vary widely between 10% and 45%, with most studies looking only at firearm injuries or patients of all ages.3-5,9 Past studies have been limited in their ability to accurately capture patients after discharge, as this cohort is often lost to follow-up or present to different hospitals and outpatient providers, which are not captured by databases reviewing inpatient and/or ED admission.10,11 Furthermore, most fatal firearm injuries occur at the scene or before ED arrival, limiting the ability of surveillance databases to accurately capture mortality rates outside a medical facility. Using the Social Security Death Index (SSDI) database and review of patient records, the purpose of our study was to evaluate young patients who present after a violent trauma to determine their long-term outcomes, as well as define patterns of injury. Our comprehensive study assessed patient outcomes utilizing the SSDI to accurately capture out-ofhospital deaths and review of subsequent hospitalizations following the initial trauma injury.
Methods Study population A retrospective cohort study was performed at the Carolinas Medical Center, an accredited level 1 trauma center. The Carolinas Medical Center is the only designated level 1 trauma center in Charlotte, North Carolina, which serves a catchment area of over 2.5 million people. The F.H. Sammy Ross Trauma Center receives over 119,000 ED visits and 5000 trauma activations annually, including over 1600 intensive care unit
141
admissions each year. The trauma center is also affiliated within an extensive hospital network and serves as the regional hub of the Metrolina Trauma Advisory Committee trauma system, increasing the likelihood of capturing subsequent injuries. Approval of this study was obtained from the Carolinas HealthCare System Institutional Review Board, which waived the need for informed consent. All trauma patients aged between 18 and 25 y who presented to the trauma center between 2009 and 2015 were reviewed. Using injury type and injury E-code, those with nonaccidental gunshot wounds (GSWs), stab wounds, or blunt assault-related injuries were classified as having sustained a violence-related injury. Individual patient records were then manually reviewed to confirm violent injuries were not unintentional (e.g., work-related). Trauma recidivism/recidivists were defined as patients with a history of prior unrelated trauma visit(s) or presenting with new unrelated traumatic injuries during the analyzed follow-up period. Patients who presented with a previous injury before the study period were also categorized as having a “history of prior trauma”. Patients under the age of 18 y were excluded from our analysis. Patient demographics, injury type, clinical presentation, hospital course, inpatient mortality, and discharge disposition were recorded for all patients including any subsequent admissions.
Study design After data query, trauma patients were separated into cohorts based on their presenting mechanism of injury: violent versus nonviolent. Patients were compared for demographics, initial ED presentation, discharge disposition, and mortality. A subgroup analysis of the violent trauma cohort was performed based on individual trauma recidivism. Patients identified as recidivists were compared with those without recurrent trauma visits to evaluate for potential risk factors for recidivism and mortality. The primary outcomes were recurrent trauma and allcause mortality. Secondary outcomes included time to recurrence and mechanism of subsequent injuries. Recurrence was identified when patients presented after discharge with at least one unrelated subsequent injury. Mortality was recorded using both hospital follow-up and SSDI Master File database, a database of death records created from the United States Social Security Administration Master File Extract. The SSDI Master File database contains over 85 million death records and is the most comprehensive research tool to verify deaths. In firearm and other violent injuries where majority of fatalities occur at the scene of injury and are underestimated by inpatient or ED databases, the SSDI captured death information in patient mortalities occurring outside the medical system. Patients were crossreferenced against the SSDI using available patient identifiers including name, date of birth, and social security number to identify patients with out-of-hospital mortalities. Demographic variables analyzed included age, sex, race/ ethnicity, insurance status, and hospital charges. Injuryrelated variables included mechanism of injury for index trauma and subsequent presentations.
142
j o u r n a l o f s u r g i c a l r e s e a r c h m a y 2 0 1 9 ( 2 3 7 ) 1 4 0 e1 4 7
Statistical analysis
Results
discharged from the hospital; 2.7% died while in the ED and 4.0% died after inpatient admission. The mean time to death was 5.2 14.6 mo after initial injury. Mortality was significantly higher following GSW (11.9%) compared to blunt assault (5.6%, P < 0.001) and stab wounds (3.9%, P < 0.001) (Table 2). In patients who survived until discharge, violent trauma was also associated with greater out-of-hospital mortality (2.6% versus 0.5%, P < 0.0005). When isolating deaths that occurred following discharge only, mortality rate was higher following stab wounds (3.5%, n ¼ 9) and assault (3.4%, n ¼ 6) compared to GSW (2.2%, n ¼ 17). All out-ofhospital deaths occurred after 30 d from index trauma with mean time to death 6.4 mo (range 1.6-71 mo). Firearm homicide was confirmed as the cause of death in 62.5% (n ¼ 20) of mortalities after discharge. Other causes of death after discharge included motor vehicle accidents (n ¼ 4) and infections (n ¼ 2); cause of death was undetermined in 6 patients. Of the violent trauma patients who had delayed out-ofhospital mortality, 89.7% sustained a GSW as their index trauma (P < 0.0001). Twenty-five percent were recidivists with at least one recurrent violent trauma, although this did not reach statistical significance.
Patient characteristics/demographics
Recidivism after violent injury
A total of 6484 patients aged between 18 and 25 y presented to our level 1 trauma center between 2009 and 2015. Of these, 1215 (18.7%) patients sustained a violence-related injury. Distribution of violent injuries included 64.4% GSWs, 21.1% stab wounds, and 14.8% blunt assaults. Demographic and patient characteristics of patients sustaining violent and nonviolent injuries are shown in Table 1. Compared to nonviolent traumas, patients in the violent injury cohort were more likely to be male (87.4% versus 67.8%, P < 0.0001) and African-American (64.9% versus 29.1%, P < 0.0001). Violently injured patients were also more likely to have a history of prior trauma (64.2% versus 50.0%, P < 0.03) and live in an area below the median income (19.1% versus 15.4%, P < 0.001). Among violent injury patients, 2.7% died in ED following resuscitative efforts, 73.6% were admitted to the hospital, while 23.7% were treated and sent home from the ED or left against medical advice. Violent traumas were more likely to require emergent surgical intervention (21.9% versus 7.9%, P < 0.0001) and have increased length of stay (median 1.1, IQR 1.0-4.0 d versus median 1.0, IQR 1.0-3.0 d, P < 0.0001) compared to non-violent traumas. Positive blood alcohol levels (42.9% versus 37.9%, P < 0.0001) and urine drug screen tests (41.0% versus 39.4%, P < 0.0001) were more common in violent trauma than non-violent trauma. Overall mean follow-up time after violent trauma was 28.3 mo.
Of the 1134 (93.3%) patients who sustained a violent trauma and survived until hospital discharge, 27.7% (n ¼ 314) had either recurrent injury or fatality outside the hospital. Patients who presented to the hospital with another injury (n ¼ 282; 24.9%) unrelated to their index trauma were identified as recidivists (Table 3). Recurrence was more common among black individuals (73.1% versus 20.2% white, 1.8% Hispanic, 4.6% others, P < 0.01). Age, gender, length of stay, and injury severity score did not correlate with trauma recidivism. Recidivism was also more common in patients living in an area below median income. Ninety percent of subsequent injuries occurred within 5 y, with 19.1% (n ¼ 54) of recidivism occurring in first 12 mo after the index trauma. Mean time to second violent injury was 31.9 21.0 mo. Among recidivist patients, 63.5% (n ¼ 179) had a GSW during their index trauma (versus 21.6% stab wound and 14.9% assault, P < 0.0001). Of recidivists with an index firearm-related injury, 20.1% (n ¼ 36) had a second firearm-related injury, and 22.3% (n ¼ 40) presented with a blunt assault injury (Table 2). Multiple traumas were also common among recidivists; 14.9% (n ¼ 42) presented with two unrelated traumas during the study period and 8.9% (n ¼ 25) had three or more trauma visits. The most common violent mechanism of injury in subsequent traumas was blunt assault (26.6%), followed by GSW (15.2%) and stab wound (6.7%). Thirty-seven percent (37.7%) of recidivists presented with non-violent injuries, most commonly motor vehicle accidents, motorcycle collisions or pedestrian injuries. Interestingly, 13.8% (n ¼ 39) of recidivist admissions were related to suicidal attempt or ideation, highlighting the potential benefit of psychologic screening in this cohort.
Descriptive statistics including medians, interquartile ranges, counts, and percentages were calculated for the overall patient cohort. Univariate logistic regression analysis was performed by initial mechanism of injury (violent versus nonviolent), trauma recidivism, and mortality. Categorical variables were assessed using chi-square test or Fisher’s exact test, and continuous variables were reported as median with interquartile range (IQR) and compared using standard t tests. For ordinal or interval data (which were not normally distributed), Wilcoxon rank-sum and KruskaleWallis tests were used. Multivariate analysis was performed to control for significant variables on univariate analysis with P-value less than 0.05 and confounding clinical factors including: age, race, injury severity score, and mechanism of injury. Statistical Analysis Software, version 9.3 (SAS Institute, Inc, Cary, NC) was used for all analyses with P < 0.05 considered statistically significant.
Mortality Patients who sustained violent trauma had increased risk of overall mortality when compared to patients in the same age cohort who sustained non-violent trauma (9.3% versus 2.1%, P < 0.0001). In contrast to those with non-violent injuries, violent trauma patients also had higher rates of in-hospital mortality (6.7% versus 1.6%, P < 0.0001). Among patients who presented after violent trauma, 93.3% survived to be
Multivariate analysis After controlling for confounding variables (Table 4), increased injury severity score was independently associated with
kao et al recidivism after violent injuries
143
Table 1 e Patient characteristics of violent versus nonviolent traumas. Variable
Violent trauma (n ¼ 1215)
Nonviolent trauma (n ¼ 5269)
P-value 0.02
Age (y)
22.2 (20.4-23.9)
21.9 (20.0-23.9)
Male, n (%)
1062 (87.4)
3574 (67.8)
<0.0001
White
278 (23.0)
2824 (54.2)
<0.0001
Black
785 (64.9)
1517 (29.1)
26 (2.2)
213 (4.1)
Race, n (%)
Hispanic Asian/Pacific Islander
5 (0.4)
66 (1.3)
Native American
4 (0.2)
12 (0.2)
109 (9.0)
572 (11.0)
History of prior trauma, n (%)
Other
780 (64.2)
2634 (50.0)
0.03
Below median income*, n (%)
1232 (19.1)
811 (15.4)
0.001
Insurance type, n (%) Private
155 (12.8)
924 (17.6)
Medicaid/Medicare
448 (37.0)
1911 (36.4)
Uninsured/self pay
557 (46.0)
1851 (35.3)
50 (4.2)
561 (10.7)
145 (12.1)
4624 (95.1)
1022 (85.4)
111 (2.3)
<0.0001
Government Other Injury type, n (%) Blunt Penetrating
5.0 (1.0-12.0)
<0.0001 0.002
ISS
5.0 (1.0-10.0)
Positive blood alcohol level, n (%)
521 (42.9)
1997 (37.9)
<0.0001
Positive urine drug screen, n (%)
498 (41.0)
2076 (39.4)
<0.0001
Admit to floor
452 (37.2)
1982 (40.9)
<0.0001
Admit to ICU
176 (14.5)
868 (17.9)
Discharged home/left AMA
288 (23.7)
1589 (32.8)
ED disposition, n (%)
ED death Operating room Length of stay (d) Hospital charges Overall mortality, n (%)
33 (2.7)
24 (0.5)
266 (21.9)
381 (7.9)
1.1 (1.0-4.0)
1.0 (1.0-3.0)
$23,712 ($12,763-$50,723) 113 (9.3)
$20,462 ($11,403-$41,470)
0.0002 0.0003
109 (2.1)
<0.0001
Inpatient death, n (%)
81 (6.7)
85 (1.6)
<0.0001
Out-of-hospital death, n (%)
32 (2.6)
24 (0.5)
0.0002
Continuous variables reported as median (IQR). Bold values are reached statistical significance where P < 0.05. ISS ¼ Injury Severity Score; ICU ¼ intensive care unit; AMA ¼ against medical advice; ED ¼ emergency department; IQR ¼ interquartile range. * Based on median income of patient zipcode.
mortality after violent trauma (odds ratio [OR] 1.1; 95% confidence interval [CI] 1.067-1.155). Similarly, patients with longer intensive care unit length of stay also were more likely to have higher mortality after violent trauma (OR 1.4; 95% CI 1.1681.788). When evaluating mortality in all trauma patients, violent traumas had increased risk of mortality compared with nonviolent traumas (OR 3.9; 95% CI 2.315-6.418). On multivariate analysis (Table 5), patients discharged to a psychiatric/mental health facility on discharge were 3.2 times more likely to have a repeat trauma admission (OR 3.2; 95% CI 1.248-8.192). Patients living in a zipcode area below median income were 1.8 times more likely to return with a second traumatic injury (OR 1.8; 95% CI 1.221-2.532).
Burden of payment Patients who sustained a violent trauma had significantly higher median ED/hospital charges compared with those with nonviolent traumas ($23,712, IQR $12,763-$50,723 versus $20,462, IQR $11,403-$41,470, P < 0.0001) with overall hospital charges in violent traumas totaling over $53.3 million dollars. No difference was noted in the index hospital charge of patients with recurrent injuries versus those presenting with single trauma admission ($21,892, IQR $11,805-$51,230 versus $24,405, IQR $12,877-$51,507, P ¼ 0.20). The average hospital charge of subsequent injuries was $57,031 $78.062. A higher proportion of violent trauma patients were uninsured/self pay (46.0% versus
144
j o u r n a l o f s u r g i c a l r e s e a r c h m a y 2 0 1 9 ( 2 3 7 ) 1 4 0 e1 4 7
Table 2 e Outcomes by mechanism of violent injury. Variable
All violent trauma (n ¼ 1134)
Assault (n ¼ 173)
Gunshot wound (n ¼ 706)
Stab wound (n ¼ 255)
P-value
Trauma recidivism*, n (%)
282 (24.9)
42 (24.2)
179 (25.3)
61 (23.9)
<0.0001
Violent recidivism*, n (%)
137 (12.1)
20 (11.6)
86 (12.2)
31 (12.2)
0.16
1
215 (76.2)
31 (73.8)
145 (81.0)
39 (65.0)
0.004
2
42 (14.9)
5 (11.9)
24 (13.4)
12 (20.0)
3þ
25 (8.9)
6 (14.3)
10 (5.6)
9 (15.0)
Number of recurrent traumas, n (%)
Recurrent trauma mechanism, n (%) GSW
43 (15.2)
2 (4.8)
36 (20.1)
5 (8.2)
0.007
Stab
19 (6.7)
2 (4.8)
10 (5.6)
7 (11.5)
0.39
Assault
75 (26.6)
16 (38.1)
40 (22.3)
19 (31.1)
0.1
Suicidal ideation/attempt
39 (13.8)
14 (33.3)
16 (8.9)
9 (14.8)
0.16
106 (37.7)
8 (19.0)
77 (43.1)
21 (34.4)
0.16
93 (11.9)
10 (3.9)
<0.0001
Nonviolent injury (e.g. MVC) Overall mortality, n (%)
113 (9.3)
10 (5.6)
Inpatient death, n (%)
81 (6.7)
4 (2.3)
76 (9.7)
1 (0.4)
<0.0001
Out-of-hospital death, n (%)
32 (2.6)
6 (3.4)
17 (2.2)
9 (3.5)
<0.0001
Bold values are reached statistical significance where P < 0.05. MVC ¼ Motor Vehicle Collision. * Recidivism calculated from patients who survived to discharge (n ¼ 1134).
35.3%, P < 0.0001) or had primary funding through Medicaid (37.0% versus 36.4%, P < 0.0001) compared with patients with nonviolent traumas. On multivariate analysis, patients with Medicaid were independently associated with higher mortality (OR 2.1; 95% CI 1.079-4.227) as well as readmission after violent trauma (OR 1.6; 95% CI 1.007-2.737).
Discussion In evaluating long-term outcomes after trauma, we demonstrated that young patients who sustained a violence-related injury have significantly higher rates of mortality and recidivism, with 27.7% of patients having a trauma-associated fatality outside the hospital or presenting to the hospital for a trauma-related admission. Previous trauma recidivism studies evaluating patients of all ages have shown higher rates of recidivism in younger patients and increased mortality in older patients, largely confounded by age-related comorbidities contributing to disease-related deaths.12 By evaluating recidivism in young patients with few comorbidities, subsequent injuries and mortalities are more likely to be related to traumatic injuries or other preventable causes. Although firearm homicides account for the second most common cause of death in young adults, 80% of violent traumas are nonfatal, and victims who survive their initial injuries may be predisposed to a cycle of violence.1,2 Previously identified risk factors associated with violent trauma include unemployment, lack of medical insurance, minority race, drug or substance abuse, and low income.3,9,11 Our study population represents an age group previously identified as high-risk for violence-related trauma and captured any assault-related injury to evaluate for mechanism of injury as a potential risk factor. Our findings demonstrated trauma recidivism and mortality due to a second trauma occurred
most commonly in patients surviving a GSW as their initial injury, as compared with stab wounds, blunt assault, and nonviolent trauma. The reason for a GSW patient being more likely to present with a subsequent traumatic injury has not been studied. It may be due to the fact that gun possession or being involved with persons who carry guns demonstrates a conscious predisposition to violence as compared with having suffered a knife wound or having been assaulted. By using information from the SSDI, we could more accurately capture deaths that occurred outside a medical facility and found a dramatically higher mortality rate of 9.3% among the violent injury cohort, compared with a mortality rate of 2.1% in those suffering nonviolent injuries. Use of the social security death database enhanced the robustness of follow-up by providing mortality information on deaths occurring outside the hospital, including scene of subsequent fatal injuries. A total of 113 deaths were identified by the SSDI, of which 32 (28.3%) occurred following discharge and increasing our inpatient death count obtained from medical records alone by 40%. In young, otherwise healthy patients who are unlikely to die from comorbid disease, an out-of-hospital death rate of 2.6% following a violence-induced injury supports the continued risk for early death shown in other studies.7,13 The findings in the present study expand on the recent but limited research on trauma recidivism in heterogeneous populations by focusing on the subgroup of young adults (18-25 y) with high rates of violent injury and significant potential for years of life lost from death and disability, therefore, demonstrated an ideal targeted group for hospitalbased intervention. Qualitative studies have shown that survivors of GSWs have substantial morbidity following discharge, including short- and long-term disabilities and with many reporting decreased quality of life.13,14 Subjectively, GSW victims reported overall diminished physical and mental health at 8 mo
145
kao et al recidivism after violent injuries
Table 3 e Violent trauma patient characteristics, recidivists versus nonrecidivists*. Variable Age (y)
Nonrecidivists (n ¼ 852) 22.2 (20.5-23.9)
Recidivistsy (n ¼ 282) 22.0(20.1-24.0)
Table 4 e Multivariate analysis of violent trauma deaths. Variable P-value 0.4
Race White
22.8%
20.2%
Black
62.7%
73.1%
Hispanic
2.40%
1.8%
Asian/Pacific Islander
0.7%
0.0%
American Indian
0.4%
0.4%
Other Male
11.0%
4.6%
87.6%
85.8%
0.01
13.7%
10.4%
Medicaid/ Medicare
35.9%
42.3%
Uninsured/ self pay
45.6%
43.9%
4.80%
61.9%
reference
Black
0.7
0.364-1.203
0.83
Other
0.5
0.184-1.349
0.29
Medicaid/Medicare
-
-
reference
-
-
2.1
1.179-4.227
0.02
Uninsured/self pay
1.1
0.496-2.551
0.16
Other
0.3
0.048-2.057
0.20
Index trauma 0.44
0.23
Assault
Stab wound
reference
-
1.2
0.528-2.731
0.76
1.2
0.452-3.171
ISS
1.1
1.067-1.155
<0.0001
0.82
ICU LOS (d)
1.4
1.168-1.788
<0.0001
ED disposition
2.90%
63.5%
0.98
White
Private
Index trauma Gunshot wound
0.887-1.130
P-value
Race
Gunshot wound
Private
1.0
95% CI
Insurance type
Insurance type
Other
Age (y)
Odds ratio
0.94
Home
reference
Floor
1.8
0.601-5.087
-
0.39
-
ICU
4.5
1.417-14.358
0.005
Or
2.8
0.906-8.888
0.29
Bold values are reached statistical significance where P < 0.05. ISS ¼ Injury Severity Score; LOS ¼ length of stay; ICU ¼ intensive care unit; ED, emergency department; CI ¼ confidence interval.
Stab wound
22.8%
Assault
15.4%
21.6% 14.9%
ISS
4.0 (1.0-10.0)
4.0 (1.0-9.0)
0.14
LOS (d)
1.6 (1.0-4.4)
1.6 (1.0-4.3)
0.68
ICU LOS (d)
0.0 (0.0-2.0)
0.0 (0.0-2.0)
0.25
Admit to floor
40.4%
48.7%
0.64
Admit to ICU
13.4%
9.9%
Operating room
22.2%
22.3%
Discharged from ED/AMA
24.0%
19.1%
ED disposition
Bold values are reached statistical significance where P < 0.05. ISS ¼ Injury Severity Score; LOS ¼ length of stay; ICU ¼ intensive care unit; ED, emergency department; AMA ¼ against medical advice. * Data reflects index trauma. y Recidivism calculated from patients who survived to discharge (n ¼ 1134).
following their injury, with increased pain and decreased functionality.14 These findings may be magnified due to the young age of the patients studied herein. These disabilities, their impact on the patients’ quality of life, and the financial burden associated with them, are significantly increased by the patients’ expected length of life. In addition, these individuals are frequently minorities from low-income, urban areas, as reinforced by the present study, which can accentuate the chronic disabilities or symptoms that limit their quality of life or ability to work. Similar studies in adolescents have shown that early exposure to violence accompanied by
declines in physical and psychological function can predispose as well as perpetuate a cycle of trauma and injury.15-18 As well, present study findings demonstrate 13.8% of patients present later with a suicide attempt or ideation, which highlights the psychological issues surrounding violent trauma and the potential gain from psychologic screening and evaluation in this patient cohort at the initial injury. The public health implications of our study support the need for multidisciplinary, point of care interventions, particularly for high-risk populations such as the one targeted in our patient population. Joint efforts between hospitals and law enforcement should attempt to intercept the cycle of violence in young individuals before repeat violence-related hospitalizations or arrests. Previous bans, such as the 1996 Dickey Amendment, greatly limited federal funding and research dedicated to firearm and violence prevention; this was until the recent media coverage of mass shooting events prompted lifting of the ban in 2012 and increased advocacy for gun control at the state and federal levels.19 The number of publications focused on firearm violence decreased 64% between 1998 and 2012 despite steady firearm-related injury rates.20 One study by Resnick et al. studied the impact of restrictive firearm legislation between states and found that despite decreases in overall firearm fatality and suicide rates, current firearm legislation has made negligible impact on the rates of homicide and African-American firearm-related fatalities. This is an alarming statistic when considering 41.7% of firearm deaths occur in these population subsets.2,19 The present study targets a patient population with demonstrated history of interpersonal
146
j o u r n a l o f s u r g i c a l r e s e a r c h m a y 2 0 1 9 ( 2 3 7 ) 1 4 0 e1 4 7
Table 5 e Multivariate analysis of violent trauma recidivists. Variable Age (y)
Odds ratio
95% CI
1.0
0.925-1.057
P-value 0.73
Race White
reference
-
Black
1.1
0.773-1.620
0.001
-
Other
0.4
0.221-0.735
0.0004
Insurance type Private Medicaid/Medicare
reference
-
1.6
1.007-2.737
0.02
Uninsured/self pay
1.2
0.733-1.950
0.33
Other
0.7
0.255-1.635
0.14
Index trauma Stab wound Gunshot wound Assault
reference
-
1.1
0.775-1.627
0.89
1.3
0798-2.166
0.32
Below median income
1.8
1.221-2.532
0.002
ISS
1.0
0.954-1.011
0.21
LOS (d)
1.0
0.975-1.038
0.71
reference
-
Jail
0.8
0.399-1.700
Nursing facility
0.4
0.106-1.540
0.07
Psych
3.2
1.248-8.192
0.004
Discharge disposition Home
0.53
Bold values are reached statistical significance where P < 0.05. ISS ¼ Injury Severity Score; LOS ¼ length of stay.
violence and is particularly high-risk for future injuries and self-harm. While legislative efforts have focused on firearmrelated injuries due to their lethality and role in mass murder violence, individuals with any history of prior assault, stab, or firearm injury should herald a warning for potential intervention before discharge, including the 25% of patients discharged home from the ED after violent trauma. In addition, the health cost burden is significantly increased in patients with violent trauma, who, on average, had hospital charges of $44,094 per patient. With 83% of this patient cohort uninsured/self pay or funded through Medicaid, the government and hospitals absorb a significant cost burden. According to 2015 estimates by Centers for Disease Control and Prevention, nonfatal hospitalizations from firearms, stab wounds, and blunt assault totaled over $48 billion in medical and workloss costs.2,21 One recent study reported estimates of costs for firearm hospitalizations alone totaled $735 million, an astounding figure when considering the disproportionate number of patients within this cohort who are uninsured or with a primary expected government payer such as Medicaid.6 The economic burden following violent injuries is increased exponentially by recidivist injuries, which incur costs associated with additional hospitalizations. In patients who survived a violent trauma, our study demonstrated average hospital charges in subsequent injuries of $57,031 per patient. Interventions aimed at preventing future injuries in highrisk patients such as our study population, where 1 in 4 will present with trauma recidivism or fatality, have enormous
potential in reducing shared health-care costs as well as the burden of injury in young adults. Hospital-based violence intervention programs (HVIPs) are based on the idea of a golden window of opportunity where victims of trauma are more likely to be responsive to interventions and resources that promote behavioral change. HVIPs have been shown to dramatically reduce the risk of reinjury, with one study citing a fourfold reduction after implementation of changes to screen patients for reinjury risk and provide resources for mentorship, mental health services, and risk factor modifications.22-24 In addition, cost-effectiveness analysis of HVIPs have demonstrated clear benefits in reducing patient and hospital costs, largely due to reduced recidivism rates.25,26 There are several limitations present in this study. In this retrospective analysis of administrative data, additional patients’ injuries may have been listed as unintentional, and it is possible that a patients’ identity may be incorrectly registered. Indeed, if so, it would only lead to an underestimation of the true number of violence-related deaths. Another limitation of our study was use of a single level 1 trauma center to accrue patients; thus, we may not have captured patients who presented to another facility with recurrent injury; again, this would lead to an underestimation of recidivism and increase the 25% recurrent trauma rates and 14% suicide attempt or ideation. However, as the only level 1 trauma center in the metropolitan area and Western North Carolina, any major and many minor traumatic injuries seen at outlying trauma centers/EDs would have been likely transferred for evaluation. Importantly, our trauma center is part of one of the nation’s largest hospital systems, which includes more than 40 surrounding hospitals. It is linked by a single electronic medical record, and any medical evaluation to an outpatient provider or tertiary care center would be captured by the study patient review process. The use of the SSDI captured mortality in a number of patients where death occurred outside of a medical facility, which enhanced the robust follow-up of our study.
Conclusion In young trauma patients sustaining a violent injury, the true burden of injury extends past discharge. These patients have a significantly higher rate of morbidity and mortality, both related to their trauma and potential future injuries. The psychological implications and consequences, including the remarkably high rate of suicide attempts and consideration of such cannot be ignored. These all carry significant implications on the health-care system to identify at-risk patients. Nearly one-quarter of patients who survive a violent injury will present with a second unrelated trauma, with 20% of injuries occurring within the first year. Improved medical, psychological, and social collaborative treatment of these high-risk patients is needed to interrupt the cycle of violent injury.
Acknowledgments Author contributions: All persons who meet authorship criteria are listed as authors, and all authors certify that they
kao et al recidivism after violent injuries
have participated sufficiently in the work to take public responsibility for the content, including participation in the concept, design, analysis, writing, or revision of the manuscript. Furthermore, each author certifies that this material or similar material has not been and will not be submitted to or published in any other publication before its appearance in Journal of Surgical Research. AM Kao, BT Heniford, KR Kasten, and RF Sing contributed to conception and design of study. AM Kao, KA Schlosser, and MR Arnold contributed to acquisition of data. AM Kao, MR Arnold, BT Heniford, and RF Sing contributed to analysis and/ or interpretation of data. AM Kao contributed to drafting of the manuscript. KA Schlosser, MR Arnold, KR Kasten, BD Davis, PD Colavita, RF Sing, and BT Heniford contributed to revision of the manuscript critically for important intellectual content. Sources of funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Disclosure The authors have no relevant financial and personal relationships that could inappropriately influence this work or its conclusions.
references
1. Lyons BH, Fowler KA, Jack SPD, et al. Surveillance for violent deaths d national violent death reporting system, 17 states, 2013. MMWR Surveill Summ. 2016;65:1e42. 2. Prevention UC for DC and. Web-Based Injury statistics query and reporting system. n.d. Available at: http://www.cdc.gov/ injury/wisqars/index.html. Accessed September 12, 2017. 3. Brooke BS, Efron DT, Chang DC, et al. Patterns and outcomes among penetrating trauma recidivists: it only gets worse. J Trauma. 2006;61:16e19. discussion 20. 4. Sims DW, Bivins BA, Obeid FN, et al. Urban trauma: a chronic recurrent disease. J Trauma. 1989;29:940e946. discussion 946-947. 5. Morrissey TB, Byrd CR, Deitch EA. The incidence of recurrent penetrating trauma in an urban trauma center. J Trauma. 1991;31:1536e1538. 6. Spitzer SA, Staudenmayer KL, Tennakoon L, et al. Costs and financial burden of initial hospitalizations for firearm injuries in the United States, 2006-2014. Am J Public Health. 2017;107:770e774. 7. Rowhani-Rahbar A, Zatzick D, Wang J, et al. Firearm-related hospitalization and risk for subsequent violent injury, death, or crime Perpetration. Ann Intern Med. 2015;162:492. 8. Statistics B of J. NCVS Victimization Analysis Tool. Washington, DC; n.d. Available at: www.bjs.gov/index.cfm? ty¼nvat. Accessed November 12, 2017.
147
9. Cooper C, Eslinger D, Nash D, et al. Repeat victims of violence: report of a large concurrent case-control study. Arch Surg. 2000;135:837e843. 10. Rattan R, Parreco J, Namias N, et al. Hidden costs of hospitalization after firearm injury: national analysis of different hospital readmission. Ann Surg. 2017;3:493. 11. Kaufman E, Rising K, Wiebe DJ, et al. Recurrent violent injury: magnitude, risk factors, and opportunities for intervention from a statewide analysis. Am J Emerg Med. 2016;34:1823e1830. 12. Strong BL, Greene CR, Smith GS. Trauma recidivism Predicts long-term mortality Missed opportunities for prevention ( retrospective cohort study ). Ann Surg. 2017;265:847e853. 13. Fahimi J, Larimer E, Hamud-Ahmed W, et al. Long-term mortality of patients surviving firearm violence. Inj Prev. 2016;22:129e134. 14. Greenspan AI, Kellermann AL. Physical and psychological outcomes 8 months after serious gunshot injury. J Trauma. 2002;53:709e716. 15. DiScala C, Sege R. Outcomes in children and young adults who are hospitalized for firearms-related injuries. Pediatrics. 2004;113:1306e1312. 16. Herbert A, Gilbert R, Gonza´lez-Izquierdo A, et al. 10-y risks of death and emergency Re-admission in adolescents Hospitalised with violent, drug- or alcohol-related, or selfInflicted injury: a population-based cohort study. PLoS Med. 2015;12:e1001931. 17. Carter PM, Walton MA, Roehler DR, et al. Firearm violence among high-risk emergency department youth after an assault injury. Pediatrics. 2015;135:805e815. 18. Cunningham RM, Carter PM, Ranney M, et al. Violent reinjury and mortality among youth seeking emergency department care for assault-related injury: a 2-year prospective cohort study. JAMA Pediatr. 2015;169:63e70. 19. Resnick S, Smith RN, Beard JH, et al. Firearm deaths in America. Ann Surg. 2017;266:1. 20. Alcorn T. Trends in research publications about gun violence in the United States, 1960 to 2014. JAMA Intern Med. 2017;177:124. 21. Fowler KA, Dahlberg LL, Haileyesus T, Annest JL. Firearm injuries in the United States. Prev Med (baltim). 2015;79:5e14. 22. Smith R, Dobbins S, Evans A, et al. Hospital-based violence intervention: risk reduction resources that are essential for success. J Trauma Acute Care Surg. 2013;74:976e982. 23. Dicker RA, Jaeger S, Knudson MM, et al. Where do we go from here? interim analysis to forge ahead in violence prevention. J Trauma - Inj Infect Crit Care. 2009;67:1169e1175. 24. Kramer EJ, Dodington J, Hunt A, et al. Violent reinjury risk assessment instrument (VRRAI) for hospital-based violence intervention programs. J Surg Res. 2017;217:177e186.e2. 25. Juillard C, Smith R, Anaya N, et al. Saving lives and saving money: hospital-based violence intervention is cost-effective. J Trauma Acute Care Surg. 2015;78:252e258. 26. Chong VE, Smith R, Garcia A, et al. Hospital-centered violence intervention programs: a cost-effectiveness analysis. Am J Surg. 2015;209:597e603.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.JournalofSurgicalResearch.com
Association for Academic Surgery
Trauma Recidivism and Mortality Following Violent Injuries in Young Adults Angela M. Kao, MD,a Kathryn A. Schlosser, MD,a Michael R. Arnold, MD,a Kevin R. Kasten, MD,a Paul D. Colavita, MD,a Bradley R. Davis, MD,a Ronald F. Sing, DO,b and B. Todd Heniford, MDa,* a
Division of Gastrointestinal and Minimally Invasive Surgery, Department of Surgery, Carolinas Medical Center, Charlotte, North Carolina b Division of Trauma and Acute Care Surgery, Department of Surgery, Carolinas Medical Center, Charlotte, North Carolina
article info
abstract
Article history:
Background: Trauma recidivism accounts for approximately 44% of emergency department
Received 2 March 2018
admissions and remains a significant health burden with this patient cohort carrying
Received in revised form
higher rates of morbidity and mortality.
16 August 2018
Methods: A level 1 trauma center registry was queried for patients aged 18-25 y presented
Accepted 4 September 2018
between 2009 and 2015. Patients with nonaccidental gunshot wounds, stab wounds, or blunt assault-related injuries were categorized as violent injuries. Primary outcomes included mortality and recidivism, which were defined as patients with two unrelated
Keywords:
traumas during the study period. Hospital records and the Social Security Death Index
Violence
were used to aid in outcomes.
Intentional injury
Results: A total of 6484 patients presented with 1215 (18.7%) sustaining violent injuries
Firearm-related injury
(87.4% male, median age 22.2 y). Mechanism of violent injuries included 64.4% gunshot
Penetrating trauma
wound, 21.1% stab, and 14.8% blunt assault. Compared with nonviolent injuries, violent
Homicide
injury patients had increased risk of mortality (9.3% versus 2.1%, P < 0.0001). Out-of-
Young adult
hospital mortality was 2.6% (versus 0.5% nonviolent, P < 0.0005), with an average time to
Teenage
death being 6.4 mo from initial injury. Recidivism was 24.9% with mean time to second
Gunshot wound
violent injury at 31.9 21.0 mo; 14.9% had two trauma readmissions, and 8.0% had 3.
Stab wound
Ninety percent of subsequent injuries occurred within 5 y, with 19.1% in the first year.
Assault
Conclusions: The burden of injury after violent trauma extends past discharge as patients have significantly higher mortality rates following hospital release. Over one-quarter present with a second unrelated trauma or death. Improved medical, psychological, and social collaborative treatment of these high-risk patients is needed to interrupt the cycle of violent injury. Published by Elsevier Inc.
Accepted for oral presentation at the 13th Annual Academic Surgical Congress, January 30, 2018, Jacksonville, Florida. * Corresponding author. Carolinas Medical Center, 1025 Morehead Medical Drive, Suite 300, Charlotte, NC 28204. Tel.: þ1 704 355 3168; fax: þ1 704 355 3168. E-mail address: [email protected] (B.T. Heniford). 0022-4804/$ e see front matter Published by Elsevier Inc. https://doi.org/10.1016/j.jss.2018.09.006
kao et al recidivism after violent injuries
Introduction Violent trauma accounts for over 57,000 deaths and 1.4 million emergency department (ED) visits annually, including over 700,000 admissions in patients aged between 10 and 24 y.1,2 An estimated 44% of trauma admissions occur in individuals with a prior history of injury-related admissions, a phenomenon known as trauma recidivism.3-5 Violent penetrating injuries requiring hospitalization and surgery incur a significant cost, and recurrent injuries increase this cost burden. In 2010, estimated medical costs for injuries related to interpersonal violence totaled $8.5 billion, with 40% of costs absorbed by hospitals and government payers.6 Studies have also shown victims of firearm assault are at increased risk of violencerelated crimes and mortality after hospitalization, with firearm violence accounting for over 36,000 deaths and 81,000 nonfatal injuries annually.3-5,7 Firearm-related injuries account for 91% of homicides and remain the second leading cause of death in young adults.1,2 Identifying and interrupting cycles of violence in high-risk cohorts, and thus preventing future injury or death, is an excellent target for potential hospital-based intervention. Despite the increased risk of repeat victimization, the longterm outcomes of young adults sustaining violence-related trauma have not been well studied. Although firearm-related injuries remain at the center of national conversation, they account for only 9.2% of all violent assaults among individuals aged 18-24 y.8 Reported rates of violent injury recidivism vary widely between 10% and 45%, with most studies looking only at firearm injuries or patients of all ages.3-5,9 Past studies have been limited in their ability to accurately capture patients after discharge, as this cohort is often lost to follow-up or present to different hospitals and outpatient providers, which are not captured by databases reviewing inpatient and/or ED admission.10,11 Furthermore, most fatal firearm injuries occur at the scene or before ED arrival, limiting the ability of surveillance databases to accurately capture mortality rates outside a medical facility. Using the Social Security Death Index (SSDI) database and review of patient records, the purpose of our study was to evaluate young patients who present after a violent trauma to determine their long-term outcomes, as well as define patterns of injury. Our comprehensive study assessed patient outcomes utilizing the SSDI to accurately capture out-ofhospital deaths and review of subsequent hospitalizations following the initial trauma injury.
Methods Study population A retrospective cohort study was performed at the Carolinas Medical Center, an accredited level 1 trauma center. The Carolinas Medical Center is the only designated level 1 trauma center in Charlotte, North Carolina, which serves a catchment area of over 2.5 million people. The F.H. Sammy Ross Trauma Center receives over 119,000 ED visits and 5000 trauma activations annually, including over 1600 intensive care unit
141
admissions each year. The trauma center is also affiliated within an extensive hospital network and serves as the regional hub of the Metrolina Trauma Advisory Committee trauma system, increasing the likelihood of capturing subsequent injuries. Approval of this study was obtained from the Carolinas HealthCare System Institutional Review Board, which waived the need for informed consent. All trauma patients aged between 18 and 25 y who presented to the trauma center between 2009 and 2015 were reviewed. Using injury type and injury E-code, those with nonaccidental gunshot wounds (GSWs), stab wounds, or blunt assault-related injuries were classified as having sustained a violence-related injury. Individual patient records were then manually reviewed to confirm violent injuries were not unintentional (e.g., work-related). Trauma recidivism/recidivists were defined as patients with a history of prior unrelated trauma visit(s) or presenting with new unrelated traumatic injuries during the analyzed follow-up period. Patients who presented with a previous injury before the study period were also categorized as having a “history of prior trauma”. Patients under the age of 18 y were excluded from our analysis. Patient demographics, injury type, clinical presentation, hospital course, inpatient mortality, and discharge disposition were recorded for all patients including any subsequent admissions.
Study design After data query, trauma patients were separated into cohorts based on their presenting mechanism of injury: violent versus nonviolent. Patients were compared for demographics, initial ED presentation, discharge disposition, and mortality. A subgroup analysis of the violent trauma cohort was performed based on individual trauma recidivism. Patients identified as recidivists were compared with those without recurrent trauma visits to evaluate for potential risk factors for recidivism and mortality. The primary outcomes were recurrent trauma and allcause mortality. Secondary outcomes included time to recurrence and mechanism of subsequent injuries. Recurrence was identified when patients presented after discharge with at least one unrelated subsequent injury. Mortality was recorded using both hospital follow-up and SSDI Master File database, a database of death records created from the United States Social Security Administration Master File Extract. The SSDI Master File database contains over 85 million death records and is the most comprehensive research tool to verify deaths. In firearm and other violent injuries where majority of fatalities occur at the scene of injury and are underestimated by inpatient or ED databases, the SSDI captured death information in patient mortalities occurring outside the medical system. Patients were crossreferenced against the SSDI using available patient identifiers including name, date of birth, and social security number to identify patients with out-of-hospital mortalities. Demographic variables analyzed included age, sex, race/ ethnicity, insurance status, and hospital charges. Injuryrelated variables included mechanism of injury for index trauma and subsequent presentations.
142
j o u r n a l o f s u r g i c a l r e s e a r c h m a y 2 0 1 9 ( 2 3 7 ) 1 4 0 e1 4 7
Statistical analysis
Results
discharged from the hospital; 2.7% died while in the ED and 4.0% died after inpatient admission. The mean time to death was 5.2 14.6 mo after initial injury. Mortality was significantly higher following GSW (11.9%) compared to blunt assault (5.6%, P < 0.001) and stab wounds (3.9%, P < 0.001) (Table 2). In patients who survived until discharge, violent trauma was also associated with greater out-of-hospital mortality (2.6% versus 0.5%, P < 0.0005). When isolating deaths that occurred following discharge only, mortality rate was higher following stab wounds (3.5%, n ¼ 9) and assault (3.4%, n ¼ 6) compared to GSW (2.2%, n ¼ 17). All out-ofhospital deaths occurred after 30 d from index trauma with mean time to death 6.4 mo (range 1.6-71 mo). Firearm homicide was confirmed as the cause of death in 62.5% (n ¼ 20) of mortalities after discharge. Other causes of death after discharge included motor vehicle accidents (n ¼ 4) and infections (n ¼ 2); cause of death was undetermined in 6 patients. Of the violent trauma patients who had delayed out-ofhospital mortality, 89.7% sustained a GSW as their index trauma (P < 0.0001). Twenty-five percent were recidivists with at least one recurrent violent trauma, although this did not reach statistical significance.
Patient characteristics/demographics
Recidivism after violent injury
A total of 6484 patients aged between 18 and 25 y presented to our level 1 trauma center between 2009 and 2015. Of these, 1215 (18.7%) patients sustained a violence-related injury. Distribution of violent injuries included 64.4% GSWs, 21.1% stab wounds, and 14.8% blunt assaults. Demographic and patient characteristics of patients sustaining violent and nonviolent injuries are shown in Table 1. Compared to nonviolent traumas, patients in the violent injury cohort were more likely to be male (87.4% versus 67.8%, P < 0.0001) and African-American (64.9% versus 29.1%, P < 0.0001). Violently injured patients were also more likely to have a history of prior trauma (64.2% versus 50.0%, P < 0.03) and live in an area below the median income (19.1% versus 15.4%, P < 0.001). Among violent injury patients, 2.7% died in ED following resuscitative efforts, 73.6% were admitted to the hospital, while 23.7% were treated and sent home from the ED or left against medical advice. Violent traumas were more likely to require emergent surgical intervention (21.9% versus 7.9%, P < 0.0001) and have increased length of stay (median 1.1, IQR 1.0-4.0 d versus median 1.0, IQR 1.0-3.0 d, P < 0.0001) compared to non-violent traumas. Positive blood alcohol levels (42.9% versus 37.9%, P < 0.0001) and urine drug screen tests (41.0% versus 39.4%, P < 0.0001) were more common in violent trauma than non-violent trauma. Overall mean follow-up time after violent trauma was 28.3 mo.
Of the 1134 (93.3%) patients who sustained a violent trauma and survived until hospital discharge, 27.7% (n ¼ 314) had either recurrent injury or fatality outside the hospital. Patients who presented to the hospital with another injury (n ¼ 282; 24.9%) unrelated to their index trauma were identified as recidivists (Table 3). Recurrence was more common among black individuals (73.1% versus 20.2% white, 1.8% Hispanic, 4.6% others, P < 0.01). Age, gender, length of stay, and injury severity score did not correlate with trauma recidivism. Recidivism was also more common in patients living in an area below median income. Ninety percent of subsequent injuries occurred within 5 y, with 19.1% (n ¼ 54) of recidivism occurring in first 12 mo after the index trauma. Mean time to second violent injury was 31.9 21.0 mo. Among recidivist patients, 63.5% (n ¼ 179) had a GSW during their index trauma (versus 21.6% stab wound and 14.9% assault, P < 0.0001). Of recidivists with an index firearm-related injury, 20.1% (n ¼ 36) had a second firearm-related injury, and 22.3% (n ¼ 40) presented with a blunt assault injury (Table 2). Multiple traumas were also common among recidivists; 14.9% (n ¼ 42) presented with two unrelated traumas during the study period and 8.9% (n ¼ 25) had three or more trauma visits. The most common violent mechanism of injury in subsequent traumas was blunt assault (26.6%), followed by GSW (15.2%) and stab wound (6.7%). Thirty-seven percent (37.7%) of recidivists presented with non-violent injuries, most commonly motor vehicle accidents, motorcycle collisions or pedestrian injuries. Interestingly, 13.8% (n ¼ 39) of recidivist admissions were related to suicidal attempt or ideation, highlighting the potential benefit of psychologic screening in this cohort.
Descriptive statistics including medians, interquartile ranges, counts, and percentages were calculated for the overall patient cohort. Univariate logistic regression analysis was performed by initial mechanism of injury (violent versus nonviolent), trauma recidivism, and mortality. Categorical variables were assessed using chi-square test or Fisher’s exact test, and continuous variables were reported as median with interquartile range (IQR) and compared using standard t tests. For ordinal or interval data (which were not normally distributed), Wilcoxon rank-sum and KruskaleWallis tests were used. Multivariate analysis was performed to control for significant variables on univariate analysis with P-value less than 0.05 and confounding clinical factors including: age, race, injury severity score, and mechanism of injury. Statistical Analysis Software, version 9.3 (SAS Institute, Inc, Cary, NC) was used for all analyses with P < 0.05 considered statistically significant.
Mortality Patients who sustained violent trauma had increased risk of overall mortality when compared to patients in the same age cohort who sustained non-violent trauma (9.3% versus 2.1%, P < 0.0001). In contrast to those with non-violent injuries, violent trauma patients also had higher rates of in-hospital mortality (6.7% versus 1.6%, P < 0.0001). Among patients who presented after violent trauma, 93.3% survived to be
Multivariate analysis After controlling for confounding variables (Table 4), increased injury severity score was independently associated with
kao et al recidivism after violent injuries
143
Table 1 e Patient characteristics of violent versus nonviolent traumas. Variable
Violent trauma (n ¼ 1215)
Nonviolent trauma (n ¼ 5269)
P-value 0.02
Age (y)
22.2 (20.4-23.9)
21.9 (20.0-23.9)
Male, n (%)
1062 (87.4)
3574 (67.8)
<0.0001
White
278 (23.0)
2824 (54.2)
<0.0001
Black
785 (64.9)
1517 (29.1)
26 (2.2)
213 (4.1)
Race, n (%)
Hispanic Asian/Pacific Islander
5 (0.4)
66 (1.3)
Native American
4 (0.2)
12 (0.2)
109 (9.0)
572 (11.0)
History of prior trauma, n (%)
Other
780 (64.2)
2634 (50.0)
0.03
Below median income*, n (%)
1232 (19.1)
811 (15.4)
0.001
Insurance type, n (%) Private
155 (12.8)
924 (17.6)
Medicaid/Medicare
448 (37.0)
1911 (36.4)
Uninsured/self pay
557 (46.0)
1851 (35.3)
50 (4.2)
561 (10.7)
145 (12.1)
4624 (95.1)
1022 (85.4)
111 (2.3)
<0.0001
Government Other Injury type, n (%) Blunt Penetrating
5.0 (1.0-12.0)
<0.0001 0.002
ISS
5.0 (1.0-10.0)
Positive blood alcohol level, n (%)
521 (42.9)
1997 (37.9)
<0.0001
Positive urine drug screen, n (%)
498 (41.0)
2076 (39.4)
<0.0001
Admit to floor
452 (37.2)
1982 (40.9)
<0.0001
Admit to ICU
176 (14.5)
868 (17.9)
Discharged home/left AMA
288 (23.7)
1589 (32.8)
ED disposition, n (%)
ED death Operating room Length of stay (d) Hospital charges Overall mortality, n (%)
33 (2.7)
24 (0.5)
266 (21.9)
381 (7.9)
1.1 (1.0-4.0)
1.0 (1.0-3.0)
$23,712 ($12,763-$50,723) 113 (9.3)
$20,462 ($11,403-$41,470)
0.0002 0.0003
109 (2.1)
<0.0001
Inpatient death, n (%)
81 (6.7)
85 (1.6)
<0.0001
Out-of-hospital death, n (%)
32 (2.6)
24 (0.5)
0.0002
Continuous variables reported as median (IQR). Bold values are reached statistical significance where P < 0.05. ISS ¼ Injury Severity Score; ICU ¼ intensive care unit; AMA ¼ against medical advice; ED ¼ emergency department; IQR ¼ interquartile range. * Based on median income of patient zipcode.
mortality after violent trauma (odds ratio [OR] 1.1; 95% confidence interval [CI] 1.067-1.155). Similarly, patients with longer intensive care unit length of stay also were more likely to have higher mortality after violent trauma (OR 1.4; 95% CI 1.1681.788). When evaluating mortality in all trauma patients, violent traumas had increased risk of mortality compared with nonviolent traumas (OR 3.9; 95% CI 2.315-6.418). On multivariate analysis (Table 5), patients discharged to a psychiatric/mental health facility on discharge were 3.2 times more likely to have a repeat trauma admission (OR 3.2; 95% CI 1.248-8.192). Patients living in a zipcode area below median income were 1.8 times more likely to return with a second traumatic injury (OR 1.8; 95% CI 1.221-2.532).
Burden of payment Patients who sustained a violent trauma had significantly higher median ED/hospital charges compared with those with nonviolent traumas ($23,712, IQR $12,763-$50,723 versus $20,462, IQR $11,403-$41,470, P < 0.0001) with overall hospital charges in violent traumas totaling over $53.3 million dollars. No difference was noted in the index hospital charge of patients with recurrent injuries versus those presenting with single trauma admission ($21,892, IQR $11,805-$51,230 versus $24,405, IQR $12,877-$51,507, P ¼ 0.20). The average hospital charge of subsequent injuries was $57,031 $78.062. A higher proportion of violent trauma patients were uninsured/self pay (46.0% versus
144
j o u r n a l o f s u r g i c a l r e s e a r c h m a y 2 0 1 9 ( 2 3 7 ) 1 4 0 e1 4 7
Table 2 e Outcomes by mechanism of violent injury. Variable
All violent trauma (n ¼ 1134)
Assault (n ¼ 173)
Gunshot wound (n ¼ 706)
Stab wound (n ¼ 255)
P-value
Trauma recidivism*, n (%)
282 (24.9)
42 (24.2)
179 (25.3)
61 (23.9)
<0.0001
Violent recidivism*, n (%)
137 (12.1)
20 (11.6)
86 (12.2)
31 (12.2)
0.16
1
215 (76.2)
31 (73.8)
145 (81.0)
39 (65.0)
0.004
2
42 (14.9)
5 (11.9)
24 (13.4)
12 (20.0)
3þ
25 (8.9)
6 (14.3)
10 (5.6)
9 (15.0)
Number of recurrent traumas, n (%)
Recurrent trauma mechanism, n (%) GSW
43 (15.2)
2 (4.8)
36 (20.1)
5 (8.2)
0.007
Stab
19 (6.7)
2 (4.8)
10 (5.6)
7 (11.5)
0.39
Assault
75 (26.6)
16 (38.1)
40 (22.3)
19 (31.1)
0.1
Suicidal ideation/attempt
39 (13.8)
14 (33.3)
16 (8.9)
9 (14.8)
0.16
106 (37.7)
8 (19.0)
77 (43.1)
21 (34.4)
0.16
93 (11.9)
10 (3.9)
<0.0001
Nonviolent injury (e.g. MVC) Overall mortality, n (%)
113 (9.3)
10 (5.6)
Inpatient death, n (%)
81 (6.7)
4 (2.3)
76 (9.7)
1 (0.4)
<0.0001
Out-of-hospital death, n (%)
32 (2.6)
6 (3.4)
17 (2.2)
9 (3.5)
<0.0001
Bold values are reached statistical significance where P < 0.05. MVC ¼ Motor Vehicle Collision. * Recidivism calculated from patients who survived to discharge (n ¼ 1134).
35.3%, P < 0.0001) or had primary funding through Medicaid (37.0% versus 36.4%, P < 0.0001) compared with patients with nonviolent traumas. On multivariate analysis, patients with Medicaid were independently associated with higher mortality (OR 2.1; 95% CI 1.079-4.227) as well as readmission after violent trauma (OR 1.6; 95% CI 1.007-2.737).
Discussion In evaluating long-term outcomes after trauma, we demonstrated that young patients who sustained a violence-related injury have significantly higher rates of mortality and recidivism, with 27.7% of patients having a trauma-associated fatality outside the hospital or presenting to the hospital for a trauma-related admission. Previous trauma recidivism studies evaluating patients of all ages have shown higher rates of recidivism in younger patients and increased mortality in older patients, largely confounded by age-related comorbidities contributing to disease-related deaths.12 By evaluating recidivism in young patients with few comorbidities, subsequent injuries and mortalities are more likely to be related to traumatic injuries or other preventable causes. Although firearm homicides account for the second most common cause of death in young adults, 80% of violent traumas are nonfatal, and victims who survive their initial injuries may be predisposed to a cycle of violence.1,2 Previously identified risk factors associated with violent trauma include unemployment, lack of medical insurance, minority race, drug or substance abuse, and low income.3,9,11 Our study population represents an age group previously identified as high-risk for violence-related trauma and captured any assault-related injury to evaluate for mechanism of injury as a potential risk factor. Our findings demonstrated trauma recidivism and mortality due to a second trauma occurred
most commonly in patients surviving a GSW as their initial injury, as compared with stab wounds, blunt assault, and nonviolent trauma. The reason for a GSW patient being more likely to present with a subsequent traumatic injury has not been studied. It may be due to the fact that gun possession or being involved with persons who carry guns demonstrates a conscious predisposition to violence as compared with having suffered a knife wound or having been assaulted. By using information from the SSDI, we could more accurately capture deaths that occurred outside a medical facility and found a dramatically higher mortality rate of 9.3% among the violent injury cohort, compared with a mortality rate of 2.1% in those suffering nonviolent injuries. Use of the social security death database enhanced the robustness of follow-up by providing mortality information on deaths occurring outside the hospital, including scene of subsequent fatal injuries. A total of 113 deaths were identified by the SSDI, of which 32 (28.3%) occurred following discharge and increasing our inpatient death count obtained from medical records alone by 40%. In young, otherwise healthy patients who are unlikely to die from comorbid disease, an out-of-hospital death rate of 2.6% following a violence-induced injury supports the continued risk for early death shown in other studies.7,13 The findings in the present study expand on the recent but limited research on trauma recidivism in heterogeneous populations by focusing on the subgroup of young adults (18-25 y) with high rates of violent injury and significant potential for years of life lost from death and disability, therefore, demonstrated an ideal targeted group for hospitalbased intervention. Qualitative studies have shown that survivors of GSWs have substantial morbidity following discharge, including short- and long-term disabilities and with many reporting decreased quality of life.13,14 Subjectively, GSW victims reported overall diminished physical and mental health at 8 mo
145
kao et al recidivism after violent injuries
Table 3 e Violent trauma patient characteristics, recidivists versus nonrecidivists*. Variable Age (y)
Nonrecidivists (n ¼ 852) 22.2 (20.5-23.9)
Recidivistsy (n ¼ 282) 22.0(20.1-24.0)
Table 4 e Multivariate analysis of violent trauma deaths. Variable P-value 0.4
Race White
22.8%
20.2%
Black
62.7%
73.1%
Hispanic
2.40%
1.8%
Asian/Pacific Islander
0.7%
0.0%
American Indian
0.4%
0.4%
Other Male
11.0%
4.6%
87.6%
85.8%
0.01
13.7%
10.4%
Medicaid/ Medicare
35.9%
42.3%
Uninsured/ self pay
45.6%
43.9%
4.80%
61.9%
reference
Black
0.7
0.364-1.203
0.83
Other
0.5
0.184-1.349
0.29
Medicaid/Medicare
-
-
reference
-
-
2.1
1.179-4.227
0.02
Uninsured/self pay
1.1
0.496-2.551
0.16
Other
0.3
0.048-2.057
0.20
Index trauma 0.44
0.23
Assault
Stab wound
reference
-
1.2
0.528-2.731
0.76
1.2
0.452-3.171
ISS
1.1
1.067-1.155
<0.0001
0.82
ICU LOS (d)
1.4
1.168-1.788
<0.0001
ED disposition
2.90%
63.5%
0.98
White
Private
Index trauma Gunshot wound
0.887-1.130
P-value
Race
Gunshot wound
Private
1.0
95% CI
Insurance type
Insurance type
Other
Age (y)
Odds ratio
0.94
Home
reference
Floor
1.8
0.601-5.087
-
0.39
-
ICU
4.5
1.417-14.358
0.005
Or
2.8
0.906-8.888
0.29
Bold values are reached statistical significance where P < 0.05. ISS ¼ Injury Severity Score; LOS ¼ length of stay; ICU ¼ intensive care unit; ED, emergency department; CI ¼ confidence interval.
Stab wound
22.8%
Assault
15.4%
21.6% 14.9%
ISS
4.0 (1.0-10.0)
4.0 (1.0-9.0)
0.14
LOS (d)
1.6 (1.0-4.4)
1.6 (1.0-4.3)
0.68
ICU LOS (d)
0.0 (0.0-2.0)
0.0 (0.0-2.0)
0.25
Admit to floor
40.4%
48.7%
0.64
Admit to ICU
13.4%
9.9%
Operating room
22.2%
22.3%
Discharged from ED/AMA
24.0%
19.1%
ED disposition
Bold values are reached statistical significance where P < 0.05. ISS ¼ Injury Severity Score; LOS ¼ length of stay; ICU ¼ intensive care unit; ED, emergency department; AMA ¼ against medical advice. * Data reflects index trauma. y Recidivism calculated from patients who survived to discharge (n ¼ 1134).
following their injury, with increased pain and decreased functionality.14 These findings may be magnified due to the young age of the patients studied herein. These disabilities, their impact on the patients’ quality of life, and the financial burden associated with them, are significantly increased by the patients’ expected length of life. In addition, these individuals are frequently minorities from low-income, urban areas, as reinforced by the present study, which can accentuate the chronic disabilities or symptoms that limit their quality of life or ability to work. Similar studies in adolescents have shown that early exposure to violence accompanied by
declines in physical and psychological function can predispose as well as perpetuate a cycle of trauma and injury.15-18 As well, present study findings demonstrate 13.8% of patients present later with a suicide attempt or ideation, which highlights the psychological issues surrounding violent trauma and the potential gain from psychologic screening and evaluation in this patient cohort at the initial injury. The public health implications of our study support the need for multidisciplinary, point of care interventions, particularly for high-risk populations such as the one targeted in our patient population. Joint efforts between hospitals and law enforcement should attempt to intercept the cycle of violence in young individuals before repeat violence-related hospitalizations or arrests. Previous bans, such as the 1996 Dickey Amendment, greatly limited federal funding and research dedicated to firearm and violence prevention; this was until the recent media coverage of mass shooting events prompted lifting of the ban in 2012 and increased advocacy for gun control at the state and federal levels.19 The number of publications focused on firearm violence decreased 64% between 1998 and 2012 despite steady firearm-related injury rates.20 One study by Resnick et al. studied the impact of restrictive firearm legislation between states and found that despite decreases in overall firearm fatality and suicide rates, current firearm legislation has made negligible impact on the rates of homicide and African-American firearm-related fatalities. This is an alarming statistic when considering 41.7% of firearm deaths occur in these population subsets.2,19 The present study targets a patient population with demonstrated history of interpersonal
146
j o u r n a l o f s u r g i c a l r e s e a r c h m a y 2 0 1 9 ( 2 3 7 ) 1 4 0 e1 4 7
Table 5 e Multivariate analysis of violent trauma recidivists. Variable Age (y)
Odds ratio
95% CI
1.0
0.925-1.057
P-value 0.73
Race White
reference
-
Black
1.1
0.773-1.620
0.001
-
Other
0.4
0.221-0.735
0.0004
Insurance type Private Medicaid/Medicare
reference
-
1.6
1.007-2.737
0.02
Uninsured/self pay
1.2
0.733-1.950
0.33
Other
0.7
0.255-1.635
0.14
Index trauma Stab wound Gunshot wound Assault
reference
-
1.1
0.775-1.627
0.89
1.3
0798-2.166
0.32
Below median income
1.8
1.221-2.532
0.002
ISS
1.0
0.954-1.011
0.21
LOS (d)
1.0
0.975-1.038
0.71
reference
-
Jail
0.8
0.399-1.700
Nursing facility
0.4
0.106-1.540
0.07
Psych
3.2
1.248-8.192
0.004
Discharge disposition Home
0.53
Bold values are reached statistical significance where P < 0.05. ISS ¼ Injury Severity Score; LOS ¼ length of stay.
violence and is particularly high-risk for future injuries and self-harm. While legislative efforts have focused on firearmrelated injuries due to their lethality and role in mass murder violence, individuals with any history of prior assault, stab, or firearm injury should herald a warning for potential intervention before discharge, including the 25% of patients discharged home from the ED after violent trauma. In addition, the health cost burden is significantly increased in patients with violent trauma, who, on average, had hospital charges of $44,094 per patient. With 83% of this patient cohort uninsured/self pay or funded through Medicaid, the government and hospitals absorb a significant cost burden. According to 2015 estimates by Centers for Disease Control and Prevention, nonfatal hospitalizations from firearms, stab wounds, and blunt assault totaled over $48 billion in medical and workloss costs.2,21 One recent study reported estimates of costs for firearm hospitalizations alone totaled $735 million, an astounding figure when considering the disproportionate number of patients within this cohort who are uninsured or with a primary expected government payer such as Medicaid.6 The economic burden following violent injuries is increased exponentially by recidivist injuries, which incur costs associated with additional hospitalizations. In patients who survived a violent trauma, our study demonstrated average hospital charges in subsequent injuries of $57,031 per patient. Interventions aimed at preventing future injuries in highrisk patients such as our study population, where 1 in 4 will present with trauma recidivism or fatality, have enormous
potential in reducing shared health-care costs as well as the burden of injury in young adults. Hospital-based violence intervention programs (HVIPs) are based on the idea of a golden window of opportunity where victims of trauma are more likely to be responsive to interventions and resources that promote behavioral change. HVIPs have been shown to dramatically reduce the risk of reinjury, with one study citing a fourfold reduction after implementation of changes to screen patients for reinjury risk and provide resources for mentorship, mental health services, and risk factor modifications.22-24 In addition, cost-effectiveness analysis of HVIPs have demonstrated clear benefits in reducing patient and hospital costs, largely due to reduced recidivism rates.25,26 There are several limitations present in this study. In this retrospective analysis of administrative data, additional patients’ injuries may have been listed as unintentional, and it is possible that a patients’ identity may be incorrectly registered. Indeed, if so, it would only lead to an underestimation of the true number of violence-related deaths. Another limitation of our study was use of a single level 1 trauma center to accrue patients; thus, we may not have captured patients who presented to another facility with recurrent injury; again, this would lead to an underestimation of recidivism and increase the 25% recurrent trauma rates and 14% suicide attempt or ideation. However, as the only level 1 trauma center in the metropolitan area and Western North Carolina, any major and many minor traumatic injuries seen at outlying trauma centers/EDs would have been likely transferred for evaluation. Importantly, our trauma center is part of one of the nation’s largest hospital systems, which includes more than 40 surrounding hospitals. It is linked by a single electronic medical record, and any medical evaluation to an outpatient provider or tertiary care center would be captured by the study patient review process. The use of the SSDI captured mortality in a number of patients where death occurred outside of a medical facility, which enhanced the robust follow-up of our study.
Conclusion In young trauma patients sustaining a violent injury, the true burden of injury extends past discharge. These patients have a significantly higher rate of morbidity and mortality, both related to their trauma and potential future injuries. The psychological implications and consequences, including the remarkably high rate of suicide attempts and consideration of such cannot be ignored. These all carry significant implications on the health-care system to identify at-risk patients. Nearly one-quarter of patients who survive a violent injury will present with a second unrelated trauma, with 20% of injuries occurring within the first year. Improved medical, psychological, and social collaborative treatment of these high-risk patients is needed to interrupt the cycle of violent injury.
Acknowledgments Author contributions: All persons who meet authorship criteria are listed as authors, and all authors certify that they
kao et al recidivism after violent injuries
have participated sufficiently in the work to take public responsibility for the content, including participation in the concept, design, analysis, writing, or revision of the manuscript. Furthermore, each author certifies that this material or similar material has not been and will not be submitted to or published in any other publication before its appearance in Journal of Surgical Research. AM Kao, BT Heniford, KR Kasten, and RF Sing contributed to conception and design of study. AM Kao, KA Schlosser, and MR Arnold contributed to acquisition of data. AM Kao, MR Arnold, BT Heniford, and RF Sing contributed to analysis and/ or interpretation of data. AM Kao contributed to drafting of the manuscript. KA Schlosser, MR Arnold, KR Kasten, BD Davis, PD Colavita, RF Sing, and BT Heniford contributed to revision of the manuscript critically for important intellectual content. Sources of funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Disclosure The authors have no relevant financial and personal relationships that could inappropriately influence this work or its conclusions.
references
1. Lyons BH, Fowler KA, Jack SPD, et al. Surveillance for violent deaths d national violent death reporting system, 17 states, 2013. MMWR Surveill Summ. 2016;65:1e42. 2. Prevention UC for DC and. Web-Based Injury statistics query and reporting system. n.d. Available at: http://www.cdc.gov/ injury/wisqars/index.html. Accessed September 12, 2017. 3. Brooke BS, Efron DT, Chang DC, et al. Patterns and outcomes among penetrating trauma recidivists: it only gets worse. J Trauma. 2006;61:16e19. discussion 20. 4. Sims DW, Bivins BA, Obeid FN, et al. Urban trauma: a chronic recurrent disease. J Trauma. 1989;29:940e946. discussion 946-947. 5. Morrissey TB, Byrd CR, Deitch EA. The incidence of recurrent penetrating trauma in an urban trauma center. J Trauma. 1991;31:1536e1538. 6. Spitzer SA, Staudenmayer KL, Tennakoon L, et al. Costs and financial burden of initial hospitalizations for firearm injuries in the United States, 2006-2014. Am J Public Health. 2017;107:770e774. 7. Rowhani-Rahbar A, Zatzick D, Wang J, et al. Firearm-related hospitalization and risk for subsequent violent injury, death, or crime Perpetration. Ann Intern Med. 2015;162:492. 8. Statistics B of J. NCVS Victimization Analysis Tool. Washington, DC; n.d. Available at: www.bjs.gov/index.cfm? ty¼nvat. Accessed November 12, 2017.
147
9. Cooper C, Eslinger D, Nash D, et al. Repeat victims of violence: report of a large concurrent case-control study. Arch Surg. 2000;135:837e843. 10. Rattan R, Parreco J, Namias N, et al. Hidden costs of hospitalization after firearm injury: national analysis of different hospital readmission. Ann Surg. 2017;3:493. 11. Kaufman E, Rising K, Wiebe DJ, et al. Recurrent violent injury: magnitude, risk factors, and opportunities for intervention from a statewide analysis. Am J Emerg Med. 2016;34:1823e1830. 12. Strong BL, Greene CR, Smith GS. Trauma recidivism Predicts long-term mortality Missed opportunities for prevention ( retrospective cohort study ). Ann Surg. 2017;265:847e853. 13. Fahimi J, Larimer E, Hamud-Ahmed W, et al. Long-term mortality of patients surviving firearm violence. Inj Prev. 2016;22:129e134. 14. Greenspan AI, Kellermann AL. Physical and psychological outcomes 8 months after serious gunshot injury. J Trauma. 2002;53:709e716. 15. DiScala C, Sege R. Outcomes in children and young adults who are hospitalized for firearms-related injuries. Pediatrics. 2004;113:1306e1312. 16. Herbert A, Gilbert R, Gonza´lez-Izquierdo A, et al. 10-y risks of death and emergency Re-admission in adolescents Hospitalised with violent, drug- or alcohol-related, or selfInflicted injury: a population-based cohort study. PLoS Med. 2015;12:e1001931. 17. Carter PM, Walton MA, Roehler DR, et al. Firearm violence among high-risk emergency department youth after an assault injury. Pediatrics. 2015;135:805e815. 18. Cunningham RM, Carter PM, Ranney M, et al. Violent reinjury and mortality among youth seeking emergency department care for assault-related injury: a 2-year prospective cohort study. JAMA Pediatr. 2015;169:63e70. 19. Resnick S, Smith RN, Beard JH, et al. Firearm deaths in America. Ann Surg. 2017;266:1. 20. Alcorn T. Trends in research publications about gun violence in the United States, 1960 to 2014. JAMA Intern Med. 2017;177:124. 21. Fowler KA, Dahlberg LL, Haileyesus T, Annest JL. Firearm injuries in the United States. Prev Med (baltim). 2015;79:5e14. 22. Smith R, Dobbins S, Evans A, et al. Hospital-based violence intervention: risk reduction resources that are essential for success. J Trauma Acute Care Surg. 2013;74:976e982. 23. Dicker RA, Jaeger S, Knudson MM, et al. Where do we go from here? interim analysis to forge ahead in violence prevention. J Trauma - Inj Infect Crit Care. 2009;67:1169e1175. 24. Kramer EJ, Dodington J, Hunt A, et al. Violent reinjury risk assessment instrument (VRRAI) for hospital-based violence intervention programs. J Surg Res. 2017;217:177e186.e2. 25. Juillard C, Smith R, Anaya N, et al. Saving lives and saving money: hospital-based violence intervention is cost-effective. J Trauma Acute Care Surg. 2015;78:252e258. 26. Chong VE, Smith R, Garcia A, et al. Hospital-centered violence intervention programs: a cost-effectiveness analysis. Am J Surg. 2015;209:597e603.