ORIGINAL ARTICLE
Children are safer in states with strict firearm laws: A National Inpatient Sample study Arash Safavi, MD, Peter Rhee, MD, Viraj Pandit, MD, Narong Kulvatunyou, MD, Andrew Tang, MD, Hassan Aziz, MD, Donald Green, MD, Terence O’Keeffe, MBChB, Gary Vercruysse, MD, Randall S. Friese, MD, and Bellal Joseph, MD, Tucson, Arizona
Firearm control laws vary across the United States and remain state specific. The aim of this study was to determine the relationship between variation in states’ firearm control laws and the risk of firearm-related injuries in pediatric population. We hypothesized that strict firearm control laws impact the incidence of pediatric firearm injury. METHODS: All patients with trauma Ecodes and those 18 years or younger were identified from the 2009 Nationwide Inpatient Sample. Individual states’ firearm control laws were evaluated and scored based on background checks on firearm sales, permit requirements, assault weapon and large-capacity magazine ban, mandatory child safety lock requirements, and regulations regarding firearms in college and workplaces. States were then dichotomized into strict firearm laws (SFLs) and nonYstrict firearm laws (non-SFLs) state based on median total score. The primary outcome measure was incidence of firearm injury. Data were compared between the two groups using simple linear regression analysis. RESULTS: A total of 60,224 pediatric patients with trauma-related injuries across 44 states were included. Thirty-three states were categorized as non-SFL and 11 as SFL. Two hundred eighty-six (0.5%) had firearm injuries, of which 31 were self-inflicted. Mean firearm injury rates per 1,000 trauma patients was higher in the non-SFL states (mean [SD]: SFL, 2.2 [1.6]; non-SFL, 5.9 [5.6]; p = 0. 001). Being in a non-SFL state increased the mean firearm injury rate by 3.75 (A coefficient, 3.75; 95% confidence interval, 0.25Y7.25; p = 0.036). CONCLUSION: Children living in states with strict firearm legislation are safer. Efforts to improve and standardize national firearm control laws are warranted. (J Trauma Acute Care Surg. 2014;76: 146Y151. Copyright * 2014 by Lippincott Williams & Wilkins) LEVEL OF EVIDENCE: Prognostic study, level III. KEY WORDS: Pediatric firearm injuries; firearm legislation; strict firearm law states; firearm injury; National Inpatient Sample. BACKGROUND:
I
n the United States, firearms have become a dangerous reality in the daily lives of the children. According to the 2011 Youth Risk Behavior Survey, 5.1% of US children use and carry guns.1 The incidence of firearm injuries among children in the United States is the highest of all industrialized countries in the world.2,3 According to the Center for Disease Control and Prevention, the incidence of firearm injuries among children has increased in the past few years, with 9,981 firearm injuries reported in 2009 alone.4 Moreover, firearm injury has become a serious health problem and currently the second leading cause of mortality among children in the United States.5 The December 14, 2012, mass shooting of 20 children in Sandy Hook Elementary School in Newtown, Connecticut, renewed the national debate for stricter firearm laws across the country. Since this event, a number of shootings involving children have dominated national headlines including the February 26, 2012, death of 17-year-old Trayvon Martin in Submitted: August 2, 2013, Revised: October 4, 2013, Accepted: October 7, 2013. From the Division of Trauma, Critical Care, Emergency Surgery, and Burns, Department of Surgery, University of Arizona, Tucson, Arizona. This study was presented as a Quickshot at the 72nd annual meeting of the American Association for the Surgery of Trauma, September 18Y21, 2013, in, San Francisco, California. Address for reprints: Bellal Joseph, MD, University of Arizona, Department of Surgery Division of Trauma, Critical Care, Emergency Surgery, and Burns 1501 N Campbell Ave, Room 5411 PO Box 245063 Tucson, AZ 85724; email: [email protected]. DOI: 10.1097/TA.0b013e3182ab10fb
146
Sanford, Florida. Currently, there are more than 300 federal gun laws established in the United States that regulate the sale, possession, and use of firearms and ammunition.6,7 The current legislature also establishes a licensing system for gun dealers and regulates transactions and record keeping by these dealers;7 however, there is a wide variability in the implementation of these federal laws across the different states. States across the country have the autonomy to supplement the federal firearm regulations and also have the right to establish their own legislature based on the assessment of gun violence in their individual state. Few studies have assessed the statewide differences in firearm legislation and incidence of firearm injuries; however, these studies have focused on the entire population (youth and adults) as a single category.7Y9 To our knowledge, until now, no studies have specifically assessed the impact of child-specific firearms legislation on the incidence of firearm injuries among children. The aim of this study was to determine the relationship between variation in state gun laws and the risk of firearmrelated injuries in the pediatric population. We hypothesized that more rigorous gun control laws would impact the incidence of pediatric firearm injury.
PATIENTS AND METHODS Data Source This is a 1-year (2009) retrospective cohort analysis of the National Inpatient Sample (NIS) database, which is maintained J Trauma Acute Care Surg Volume 76, Number 1
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
J Trauma Acute Care Surg Volume 76, Number 1
by the Agency for Healthcare Research and Quality as part of the Healthcare Cost and Utilization Project (HCUP). The NIS database is the largest all-payer inpatient care database publically available in the United States. It covers 95% of the US population and includes comprehensive abstracted discharged data. The data in the NIS are derived from a stratified sample of 20% of the discharges from all community hospitals (nonfederal, short-term, general, and specialty hospitals) in the United States.10 The data are weighted back to help make population estimates of the various parameters. For the year 2009, the NIS contains information for 7.4 million weighted discharges from 4,121 hospitals across 44 states.10 For our study, the use of NIS database was conformed to the data use agreement from HCUP. This study was reviewed by the University of Arizona Institutional Review Board and was determined to be exempt from the need for approval.
Patient Population Trauma patients were identified from the 2009 NIS database using the DRG International Classification of DiseasesV9th Rev. Trauma patients (Ecodes 800Y959) 18 years or younger were included. Patients with missing information on age, mechanism of injury, and location of hospital were excluded.
Data Points Collected The following data points were retrieved from the 2009 NIS database: demographic characteristics (age, sex, and race), day of admission (weekend, weekday), type of admission (elective, emergent), mechanism of injury, firearm injury, hospital length of stay, and in-hospital mortality. Patients with firearm-related injury were abstracted using the following Ecodes: suicide and self-inflicted injury (E955.0, E955.2, E955.3, and E955.4), assault by firearm and explosives (E965.0, E965.1, E965.2, and E965.3), and accidents caused by firearm and air gun missile (E922.0, E922.1, E922.2, E922.3, and E922.4).
State Firearm Legislation and Groups We extracted 2009 firearm control policies in all 44 NIS available states. To quantify state-level variations in gun regulations, we used data from the Brady Campaign to Prevent Gun Violence and the Brady Center to Prevent Gun Violence (referred to collectively herein as the Brady Center) in addition to the individual states’ government resources. The Brady Center has tracked firearm legislation annually since 2007 and prepares legislative scorecards for every state each year.11 Individual states’ firearm control laws were evaluated, and each state was scored based on five broad categories of firearm laws pertinent to the pediatric population: (i) background checks on firearm sales (all firearms including gun shows or handguns only), (ii) permit requirements (mandatory safety training, license to possess), (iii) bans on assault weapon and large-capacity magazines, (iv) mandatory child safety lock requirements (integrated locks, external locks), and (v) regulations on firearm in public places such as college campuses and workplaces. Each category had a potential subtotal of 4 points and a total potential aggregate score of 20 points. The higher the total aggregate score for each state, the stricter the firearm laws in that state. We then calculated the median score
Safavi et al.
for all the states. The median states’ scores were bimodaly distributed. The intersection between two normal distribution (score of 8) was used as a cutoff to categorize states as strict and nonYstrict firearm law states. States were then dichotomized into strict firearm laws (SFLs) or nonYstrict firearm laws (nonSFL) state based on the median score. Eleven states were categorized as SFL states, while the remaining (n = 33) were categorized as non-SFL states. Table 1 describes the scoring system for the firearm legislature score. The primary outcome of interest was incidence of firearm injury. Secondary outcome was firearm-related mortality rate and hospital length of stay.
Statistical Analysis Data are reported as mean (SD) for continuous descriptive variables, median (range) for ordinal descriptive variables, and as proportions for categorical variables. We performed one-way analysis of variance, Pearson’s W2 test (categorical variables), and independent t test (continuous variables) to compare demographics and outcomes between SFL and non-SFL. To compare the two groups using standardized population, we analyzed our data based on mean firearm injury rates per 1,000 trauma patients by simple linear regression analysis. For our study, we considered p e 0.05 as statistically significant. All statistical analyses were performed using SPSS (version 20; SPSS, Inc., Chicago, IL).
RESULTS A total of 60,224 pediatric trauma patients were included, of whom 286 (0.5%) had firearm-related injuries. The mean (SD) age was 9.7 (6.5) years, 34,990 (58.1%) were male, and 28,727 (47.7%) were whites. Of the 286 patients with TABLE 1. Scoring System for Firearm Legislation Legislation Intent
Description of Measures
Background checks Universal background checks All firearm (2) Universal background checks Handguns only (2) Permit to purchase Firearm purchase (2) Permit required to purchase firearms Safety training (2) Safety training and/or testing required Improve child safety Child safety locks Integrated/external locks sold on handguns Integrated locks (2) External locks (2) Regulations to restrict guns in public places No guns in workplace (2) Employers not required to allow firearms in parking lots No guns in college campus (2) Colleges are not required to allow firearms on campus Assault weapon Assault weapon ban (2) Regulations of firearms with military style features Large capacity ban (2) Maximum no. rounds per magazine e 15
* 2014 Lippincott Williams & Wilkins
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
147
J Trauma Acute Care Surg Volume 76, Number 1
Safavi et al.
TABLE 2. Demographics All Trauma Patients
n = 60,224
Age, mean (SD), y e6, % (n) 7Y12, % (n) Q13, % (n) Sex, % (n) Males Females Race, % (n) White African American Hospital LOS, mean (SD) Mortality, % (n) Patients With Firearm Injury Age, y, mean (SD) e6, % (n) 7Y12, % (n) Q13, % (n) Sex, % (n) Males Females Race, % (n) White African American Hospital LOS, mean (SD) Mortality, % (n)
9.7 (6.5) 36.3 (21,861) 19.8 (11,924) 43.9 (26,439) 58.1 (34,990) 40.9 (25,234) 47.7 (28,727) 12.2 (7,347) 5.1 (13.9) 1.1 (662) n = 286 14.5 (4) 5.9 (17) 14.7 (42) 79.4 (227) 84.6 (242) 15.4 (44) 36.4 (104) 26.2 (75) 6.7 (20.9) 8 (3)
LOS, length of stay.
firearm-related injuries, 255 (89.1%) were firearm-related assaults, and the remaining (n = 31) were self-inflicted injuries. Table 2 describes the demographics and outcomes of the study population. Of the patients with firearm-related injuries (n = 286), 220 patients were in non-SFL states, while the remaining (n = 66) were in states with SFLs. Patients in non-SFL states were younger ( p = 0.02) than patients in non-SFL states. There was no difference in sex (male) ( p = 0.4) and race ( p = 01) between patients in SFL and non-SFL states. The overall mortality rate in patients with firearm injuries was 8% (3 of 286). There was no difference ( p = 0.3) in mortality rate in SFL and non-SFL states. Table 3 highlights the differences in demographics and outcomes between patients in SFL and non-SFL states. Table 4 highlights the score, incidence of trauma, and firearm injuries in each state. California (score, 16) had the highest score and was the state with the strictest firearm laws, while Utah (score, 0) had the lowest scores and was the least strict firearm state. The mean and median among 44 states was 5 and median 4, respectively. Figure 1 demonstrates the distribution and rate of firearm injury in SFL and non-SFL states. Louisiana, a nonstrict gun law state, had the highest mean rate of firearm injuries (22.5 per 1,000), while New Jersey, an SFL state, had the lowest (1.6 per 1,000) mean rate of firearm accidents. States with non-SFLs had higher mean firearm injury rate per 1,000 trauma patients (mean [SD]: SFL, 2.2 [1.6]; 148
non-SFL, 5.9 [5.6]; p = 0. 001), compared with states with nonSFLs. On a simple linear regression analysis, we found that being in non-strict gun law state increased the mean firearm injury rate by 3.75 (A coefficient, 3.75; 95% confidence interval, 0.25Y7.25; p = 0.036).
DISCUSSION This study highlights that children-specific firearm legislations are effective in controlling the incidence of firearmrelated injuries among children. Children in non-SFL states are more likely to have a firearm-related injury in comparison with children in SFL states. Younger children were more likely to be involved in a firearm injury in non-SFL states compared with SFL states. We believe that improving and standardizing national and state firearm control laws with special focus on child safety laws are warranted. The United States has the highest rate of firearm-related deaths among children younger than 18 years in comparison with other industrialized countries.2,3 The current debate over firearm control focuses on the effectiveness of rigorous firearm control legislations. Advocates of increased legislative control claim that strict laws will reduce death rates and crime rates by restricting firearm possession by minors and by unreliable adults. However, opponents argue that strict laws do not prevent crime sufficiently to justify the restrictions placed on responsible citizens. We believe that the important point of this debate is not the number of legislations and formulating new legislations but understanding the true impact of the current legislations and how these legislations can safeguard the people they intended to protect. In our study, we classified states into SFL and non-SFL states based on children-specific firearm legislations in each state. We scored states based on the presence of children-specific firearm legislation in each state and assessed the cumulative impact of all the firearm legislations in a state on the incidence of firearm-related injuries. Studies assessing the impact of firearm legislature have routinely focused on individual components of the firearm laws and have failed to take into account the cumulative effects of these firearm legislations.8,12,13 Lee et al.8 in a review of firearm-related injuries in children assessed the impact TABLE 3. SFL Versus Non-SFL States Variables Age, mean (SD) e6, % (n) 7Y12, % (n) Q13, % (n) Sex, % (n) Males Females Race, % (n) White African American Hospital LOS, mean (SD) Mortality, % (n)
SFL (n = 66)
Non-SFL (n = 220)
p
15.4 (3.3) 3 (2) 9.1 (6) 87.9 (58)
14.3 (4.1) 6.8 (15) 16.4 (36) 76.8 (169)
0.02
87.9 (58) 12.1 (8)
83.6 (184) 16.4 (36)
40.9 (27) 22.7 (15) 4.4 (6.5) 4.5% (3)
40.5 (89) 21.8 (48) 7.3 (23.6) 9.1% (20)
0.4
0.1
0.08 0.3
LOS, length of stay.
* 2014 Lippincott Williams & Wilkins
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
J Trauma Acute Care Surg Volume 76, Number 1
Safavi et al.
TABLE 4. Incidence of Trauma and Firearm Injury in Each State SFL (n = 11) State California Connecticut Hawaii Illinois Massachusetts Maryland Michigan New Jersey New York Pennsylvania Rhode Island
Non-SFL (n = 33) Score
Trauma
Firearm
State
Score
Trauma
Firearm
State
Score
Trauma
Firearm
16 10 12 10 12 10 8 16 12 8 12
7,053 338 10 2,641 700 664 578 1,185 4,753 4,791 70
21 1 0 6 2 0 3 2 18 13 0
Arkansas Arizona Colorado Florida Georgia Iowa Indiana Kansas Kentucky Louisiana Maine Minnesota Missouri Montana North Carolina Nebraska
4 2 4 2 2 4 2 2 2 2 2 2 4 2 4 4
352 3,026 1,587 5,151 892 300 1,120 442 376 1,152 80 1,018 1,056 254 1,935 102
4 9 7 29 9 4 8 5 4 26 0 3 5 0 10 0
New Hampshire New Mexico Nevada Ohio Oklahoma Oregon South Carolina South Dakota Tennessee Texas Utah Virginia Vermont Washington West Virginia Wisconsin Wyoming
4 4 4 4 2 2 4 4 4 2 0 4 4 4 4 2 2
682 213 780 2,998 2,994 1,161 318 24 789 4,093 198 1,339 12 1,483 121 1,283 110
1 4 3 16 16 3 3 0 11 27 0 7 0 6 1 1 0
of child access protection, safe storage of guns, and stand-yourground laws on the incidence of firearm injuries in children. Sen and Panjamapirom12 in a national exploratory study assessed
the impact of only background checks on the incidence of firearm-related injury and mortality. In another study, Ludwig and Cook13 evaluated the impact of waiting periods and background
Figure 1. Distribution and rate of firearm injury in SFL and non-SFL states. * 2014 Lippincott Williams & Wilkins
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
149
J Trauma Acute Care Surg Volume 76, Number 1
Safavi et al.
checks for firearm purchases and demonstrated reduction in firearm suicide rates among individuals 55 years or older. A recently published study by Fleegler et al.9 is the only study that has assessed the true impact of all the firearm legislation in a state and the incidence of firearm-related injury. However, this study failed to take into account the firearm legislation specific to children and assess the impact of these legislation on the incidence of firearm injuries among children. Several studies have evaluated the association between firearm legislations and the incidence of firearm injuries and deaths in the United States.9,14,15 Rosengart et al.14 demonstrated an increased incidence of firearm homicides in states with no restrictions on carrying concealed weapons. Fleegler et al. demonstrated a decreased incidence of firearm-related casualties in states with higher number of firearm laws.14 In contrast, Lott15 did not find any difference in the incidence of firearm injuries in states with and without SFLs. However, none of these studies adequately assessed the cumulative impact of firearm legislations on the incidence of firearm-related injuries. In our study, we found that SFLs in the states were associated with a lower rate of firearm injuries in children. This variability in results among studies can be attributed to different methodologies of assessing firearm laws in each state and to the heterogeneity of the study population. We found some heterogeneity in firearm incidence rates among the states within each category of the legislative strength scores (e.g., Nebraska has nonstrict gun control laws and zero firearm incidence). We anticipated such heterogeneity and hence conducted this study assessing the firearm legislation across all the 44 states in the United States. The presence of a trend in higher mortality and longer stay in non-SFL states compared with SFL states despite the lack of statistical difference was most likely caused by our sample size, and it characterizes a Type II error. Our study has limitations. First, we were unable to control for the enforcement of firearm laws, which may vary between states. Second, the state gun control strength score used in our analysis has not been validated, and we are unaware of any such scoring systems that have been validated. To minimize bias in our state scoring, we gave an equal set of weights to each category to avoid biasing the results by our classification in favor or against gun control. Finally, certain types of state and local gun control legislation were not considered, and the comparison among the states did not take into account the additional gun regulations of local governments within the state. Nonetheless, this study is distinctive in the sense that it examines the cumulative impact of firearm legislation linked to pediatric population rather than single laws in a mixed adult and pediatric cohort.
AUTHORSHIP B.J., A.S., V.P., R.S.F., and P.R. designed this study. B.J., A.S., V.P., N.K., A.T., H.A., and R.S.F. searched the literature. B.J., A.S., V.P., A.T., H.A., and T.O. abstracted the data. B.J., A.S., V.P., H.A., R.S.F., and P.R. analyzed the data. All other authors participated in the data interpretation and manuscript preparation.
DISCLOSURE The authors declare no conflicts of interest.
REFERENCES 1. Center for Disease Control and Prevention. Youth risk behavior surveillanceVUnited States, 2011. MMWR Surveill Summ. 2012;61(4): 1Y162. 2. Fingerhut LA, Warner M. Injury chartbook. In: Health, United States, 1196-97. Hyattsville, MD: National Center for Health Statistics; 1997. 3. Richardson EG, Hemenway D. Homicide, suicide, and unintentional firearm fatality: comparing the United States with other high-income countries, 2003. J Trauma. 2011;70:238. 4. Web-based Injury Statistics Query and Reporting System. Centers for Disease Control and Prevention, National Center for Injury Prevention and Control. www.cdc.gov/ncipc/wisqars (Accessed on July 23, 2013). 5. Centers for Disease Control and Prevention. Injury prevention and control: Data and statistics (WISQARS). http://www.cdc.gov/injury/wisqars/index. html (Accessed on July 23, 2013). 6. Hutchinson A. Report of the National School Shield Task Force. Hutchinson A. April 2, 2013. http://www.nraschoolshield.com/NSS_Final_ FULL.pdf. Accessed July 5, 2013. 7. Vernick JS, Hepburn LM. Twenty thousand gun-control laws? In: Ludwig J, Cook PI, eds. Evaluating Gun Policy: Effects on Crime and Violence. Washington, DC: Brookings Institution Press; 2003:456. 8. Lee J, Moriarty KP, Tashjian DB, Patterson LA. Guns and states: Pediatric firearm injury. J Trauma Acute Care Surg. 2013;75(1):50Y53. 9. Fleegler EW, Lee LK, Monuteaux MC, Hemenway D, Mannix R. Firearm legislation and firearm-related fatalities in the United States. JAMA Intern Med. 2013;173(9):732Y740. 10. National Inpatient sample. http://www.hcup-us.ahrq.gov/nisoverview.jsp. Accessed July 5, 2013. 11. Brady Campaign to Prevent Gun Violence. Legislative scorecards. http:// www.bradycampaign.org/stategunlaws. Accessed July 4,2013. 12. Sen B, Panjamapirom A. State background checks for gun purchase and firearm deaths: an exploratory study. Preventive Medicine. 2012;55(4):346Y350. 13. Ludwig J, Cook PJ. Homicide and suicide rates associated with implementation of the Brady Handgun Violence Prevention Act. JAMA. 2000; 284(5):585Y591. 14. Rosengart M, Cummings P, Nathens A. An evaluation of state firearm regulations and homicide and suicide death rates. Inj Prev. 2005;11:77Y83. 15. Lott J Jr. More Guns, Less Crime: Understanding Crime and Gun Control Laws. 3rd ed. Chicago, IL: University of Chicago Press; 2010.
DISCUSSION CONCLUSION Firearm control legislations are effective in reducing the number of firearm injuries in children. Children living in SFL states are safer and are less likely to be involved in firearm accidents. Efforts to decrease pediatric firearm injuries should focus on improving and standardizing national firearm control laws aimed specifically at child safety. 150
Dr. Michael L. Nance (Philadelphia, Pennsylvania): I appreciate the opportunity to review this abstract and the manuscript. I have two questions for the authors. First, the authors used an un-validated method of scoring state firearms laws. They authors created their model that assigned an equal weight to the five categories of laws, such as assault weapons ban or safety training. The Brady Center, from whom the background data was obtained, has created their own score that assigns weights to the laws based on their belief that the legislation is of varying * 2014 Lippincott Williams & Wilkins
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
J Trauma Acute Care Surg Volume 76, Number 1
efficacy and, therefore, of varying significance. Why did the authors to use an un-validated, weighted score for their analyses? Second, the authors used the National Inpatient Sample, which included more than 60,000 injured children. Remarkably, there were only 286 firearm injuries reported. This small number was then further subdivided by the 44 states to determine rates of injury by state. Twenty-five of the 44 states reported fewer than five injuries. That seems unbelievably low and in those states just one or two more reported injuries would increase the reported rates quite significantly, by 100% or 200%. Can these numbers be believed? The data are very provocative. I think this is a very important line of research and I encourage them to continue their work. Dr. Arash Safavi (Tucson, Arizona): Thank you very much. In terms of your first question about using the scoring, we looked for a validated scoring system but we couldn’t find any. There were some scorings, for example by the Brady
Safavi et al.
Campaign, which is campaigning for restrictive gun control laws. Their scoring system includes adults and pediatric population and it’s not validated either. We chose not to use their scoring because, first of all, it does not specifically address children and, secondly, it is unvalidated and couldn’t add any value to the scoring we designed. In terms of your second question, the reason that some of the states did not have any firearm injuries in our study is mainly because NIS database does not collect every single patient and it records 20% of hospitalized patients. In addition, as I mentioned, the database only includes inpatients. For instance, if there was a patient who died at the scene, that’s not recorded in NIS. I think the most important things is the sample size and sample representation of the population. In other words, as long as you have a reliable sample such as NIS and you have enough sample size that give you enough power for your primary outcome, then the result of the study is reliable and that’s what we had.
* 2014 Lippincott Williams & Wilkins
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
151
Children are safer in states with strict firearm laws: A National Inpatient Sample study Arash Safavi, MD, Peter Rhee, MD, Viraj Pandit, MD, Narong Kulvatunyou, MD, Andrew Tang, MD, Hassan Aziz, MD, Donald Green, MD, Terence O’Keeffe, MBChB, Gary Vercruysse, MD, Randall S. Friese, MD, and Bellal Joseph, MD, Tucson, Arizona
Firearm control laws vary across the United States and remain state specific. The aim of this study was to determine the relationship between variation in states’ firearm control laws and the risk of firearm-related injuries in pediatric population. We hypothesized that strict firearm control laws impact the incidence of pediatric firearm injury. METHODS: All patients with trauma Ecodes and those 18 years or younger were identified from the 2009 Nationwide Inpatient Sample. Individual states’ firearm control laws were evaluated and scored based on background checks on firearm sales, permit requirements, assault weapon and large-capacity magazine ban, mandatory child safety lock requirements, and regulations regarding firearms in college and workplaces. States were then dichotomized into strict firearm laws (SFLs) and nonYstrict firearm laws (non-SFLs) state based on median total score. The primary outcome measure was incidence of firearm injury. Data were compared between the two groups using simple linear regression analysis. RESULTS: A total of 60,224 pediatric patients with trauma-related injuries across 44 states were included. Thirty-three states were categorized as non-SFL and 11 as SFL. Two hundred eighty-six (0.5%) had firearm injuries, of which 31 were self-inflicted. Mean firearm injury rates per 1,000 trauma patients was higher in the non-SFL states (mean [SD]: SFL, 2.2 [1.6]; non-SFL, 5.9 [5.6]; p = 0. 001). Being in a non-SFL state increased the mean firearm injury rate by 3.75 (A coefficient, 3.75; 95% confidence interval, 0.25Y7.25; p = 0.036). CONCLUSION: Children living in states with strict firearm legislation are safer. Efforts to improve and standardize national firearm control laws are warranted. (J Trauma Acute Care Surg. 2014;76: 146Y151. Copyright * 2014 by Lippincott Williams & Wilkins) LEVEL OF EVIDENCE: Prognostic study, level III. KEY WORDS: Pediatric firearm injuries; firearm legislation; strict firearm law states; firearm injury; National Inpatient Sample. BACKGROUND:
I
n the United States, firearms have become a dangerous reality in the daily lives of the children. According to the 2011 Youth Risk Behavior Survey, 5.1% of US children use and carry guns.1 The incidence of firearm injuries among children in the United States is the highest of all industrialized countries in the world.2,3 According to the Center for Disease Control and Prevention, the incidence of firearm injuries among children has increased in the past few years, with 9,981 firearm injuries reported in 2009 alone.4 Moreover, firearm injury has become a serious health problem and currently the second leading cause of mortality among children in the United States.5 The December 14, 2012, mass shooting of 20 children in Sandy Hook Elementary School in Newtown, Connecticut, renewed the national debate for stricter firearm laws across the country. Since this event, a number of shootings involving children have dominated national headlines including the February 26, 2012, death of 17-year-old Trayvon Martin in Submitted: August 2, 2013, Revised: October 4, 2013, Accepted: October 7, 2013. From the Division of Trauma, Critical Care, Emergency Surgery, and Burns, Department of Surgery, University of Arizona, Tucson, Arizona. This study was presented as a Quickshot at the 72nd annual meeting of the American Association for the Surgery of Trauma, September 18Y21, 2013, in, San Francisco, California. Address for reprints: Bellal Joseph, MD, University of Arizona, Department of Surgery Division of Trauma, Critical Care, Emergency Surgery, and Burns 1501 N Campbell Ave, Room 5411 PO Box 245063 Tucson, AZ 85724; email: [email protected]. DOI: 10.1097/TA.0b013e3182ab10fb
146
Sanford, Florida. Currently, there are more than 300 federal gun laws established in the United States that regulate the sale, possession, and use of firearms and ammunition.6,7 The current legislature also establishes a licensing system for gun dealers and regulates transactions and record keeping by these dealers;7 however, there is a wide variability in the implementation of these federal laws across the different states. States across the country have the autonomy to supplement the federal firearm regulations and also have the right to establish their own legislature based on the assessment of gun violence in their individual state. Few studies have assessed the statewide differences in firearm legislation and incidence of firearm injuries; however, these studies have focused on the entire population (youth and adults) as a single category.7Y9 To our knowledge, until now, no studies have specifically assessed the impact of child-specific firearms legislation on the incidence of firearm injuries among children. The aim of this study was to determine the relationship between variation in state gun laws and the risk of firearmrelated injuries in the pediatric population. We hypothesized that more rigorous gun control laws would impact the incidence of pediatric firearm injury.
PATIENTS AND METHODS Data Source This is a 1-year (2009) retrospective cohort analysis of the National Inpatient Sample (NIS) database, which is maintained J Trauma Acute Care Surg Volume 76, Number 1
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
J Trauma Acute Care Surg Volume 76, Number 1
by the Agency for Healthcare Research and Quality as part of the Healthcare Cost and Utilization Project (HCUP). The NIS database is the largest all-payer inpatient care database publically available in the United States. It covers 95% of the US population and includes comprehensive abstracted discharged data. The data in the NIS are derived from a stratified sample of 20% of the discharges from all community hospitals (nonfederal, short-term, general, and specialty hospitals) in the United States.10 The data are weighted back to help make population estimates of the various parameters. For the year 2009, the NIS contains information for 7.4 million weighted discharges from 4,121 hospitals across 44 states.10 For our study, the use of NIS database was conformed to the data use agreement from HCUP. This study was reviewed by the University of Arizona Institutional Review Board and was determined to be exempt from the need for approval.
Patient Population Trauma patients were identified from the 2009 NIS database using the DRG International Classification of DiseasesV9th Rev. Trauma patients (Ecodes 800Y959) 18 years or younger were included. Patients with missing information on age, mechanism of injury, and location of hospital were excluded.
Data Points Collected The following data points were retrieved from the 2009 NIS database: demographic characteristics (age, sex, and race), day of admission (weekend, weekday), type of admission (elective, emergent), mechanism of injury, firearm injury, hospital length of stay, and in-hospital mortality. Patients with firearm-related injury were abstracted using the following Ecodes: suicide and self-inflicted injury (E955.0, E955.2, E955.3, and E955.4), assault by firearm and explosives (E965.0, E965.1, E965.2, and E965.3), and accidents caused by firearm and air gun missile (E922.0, E922.1, E922.2, E922.3, and E922.4).
State Firearm Legislation and Groups We extracted 2009 firearm control policies in all 44 NIS available states. To quantify state-level variations in gun regulations, we used data from the Brady Campaign to Prevent Gun Violence and the Brady Center to Prevent Gun Violence (referred to collectively herein as the Brady Center) in addition to the individual states’ government resources. The Brady Center has tracked firearm legislation annually since 2007 and prepares legislative scorecards for every state each year.11 Individual states’ firearm control laws were evaluated, and each state was scored based on five broad categories of firearm laws pertinent to the pediatric population: (i) background checks on firearm sales (all firearms including gun shows or handguns only), (ii) permit requirements (mandatory safety training, license to possess), (iii) bans on assault weapon and large-capacity magazines, (iv) mandatory child safety lock requirements (integrated locks, external locks), and (v) regulations on firearm in public places such as college campuses and workplaces. Each category had a potential subtotal of 4 points and a total potential aggregate score of 20 points. The higher the total aggregate score for each state, the stricter the firearm laws in that state. We then calculated the median score
Safavi et al.
for all the states. The median states’ scores were bimodaly distributed. The intersection between two normal distribution (score of 8) was used as a cutoff to categorize states as strict and nonYstrict firearm law states. States were then dichotomized into strict firearm laws (SFLs) or nonYstrict firearm laws (nonSFL) state based on the median score. Eleven states were categorized as SFL states, while the remaining (n = 33) were categorized as non-SFL states. Table 1 describes the scoring system for the firearm legislature score. The primary outcome of interest was incidence of firearm injury. Secondary outcome was firearm-related mortality rate and hospital length of stay.
Statistical Analysis Data are reported as mean (SD) for continuous descriptive variables, median (range) for ordinal descriptive variables, and as proportions for categorical variables. We performed one-way analysis of variance, Pearson’s W2 test (categorical variables), and independent t test (continuous variables) to compare demographics and outcomes between SFL and non-SFL. To compare the two groups using standardized population, we analyzed our data based on mean firearm injury rates per 1,000 trauma patients by simple linear regression analysis. For our study, we considered p e 0.05 as statistically significant. All statistical analyses were performed using SPSS (version 20; SPSS, Inc., Chicago, IL).
RESULTS A total of 60,224 pediatric trauma patients were included, of whom 286 (0.5%) had firearm-related injuries. The mean (SD) age was 9.7 (6.5) years, 34,990 (58.1%) were male, and 28,727 (47.7%) were whites. Of the 286 patients with TABLE 1. Scoring System for Firearm Legislation Legislation Intent
Description of Measures
Background checks Universal background checks All firearm (2) Universal background checks Handguns only (2) Permit to purchase Firearm purchase (2) Permit required to purchase firearms Safety training (2) Safety training and/or testing required Improve child safety Child safety locks Integrated/external locks sold on handguns Integrated locks (2) External locks (2) Regulations to restrict guns in public places No guns in workplace (2) Employers not required to allow firearms in parking lots No guns in college campus (2) Colleges are not required to allow firearms on campus Assault weapon Assault weapon ban (2) Regulations of firearms with military style features Large capacity ban (2) Maximum no. rounds per magazine e 15
* 2014 Lippincott Williams & Wilkins
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
147
J Trauma Acute Care Surg Volume 76, Number 1
Safavi et al.
TABLE 2. Demographics All Trauma Patients
n = 60,224
Age, mean (SD), y e6, % (n) 7Y12, % (n) Q13, % (n) Sex, % (n) Males Females Race, % (n) White African American Hospital LOS, mean (SD) Mortality, % (n) Patients With Firearm Injury Age, y, mean (SD) e6, % (n) 7Y12, % (n) Q13, % (n) Sex, % (n) Males Females Race, % (n) White African American Hospital LOS, mean (SD) Mortality, % (n)
9.7 (6.5) 36.3 (21,861) 19.8 (11,924) 43.9 (26,439) 58.1 (34,990) 40.9 (25,234) 47.7 (28,727) 12.2 (7,347) 5.1 (13.9) 1.1 (662) n = 286 14.5 (4) 5.9 (17) 14.7 (42) 79.4 (227) 84.6 (242) 15.4 (44) 36.4 (104) 26.2 (75) 6.7 (20.9) 8 (3)
LOS, length of stay.
firearm-related injuries, 255 (89.1%) were firearm-related assaults, and the remaining (n = 31) were self-inflicted injuries. Table 2 describes the demographics and outcomes of the study population. Of the patients with firearm-related injuries (n = 286), 220 patients were in non-SFL states, while the remaining (n = 66) were in states with SFLs. Patients in non-SFL states were younger ( p = 0.02) than patients in non-SFL states. There was no difference in sex (male) ( p = 0.4) and race ( p = 01) between patients in SFL and non-SFL states. The overall mortality rate in patients with firearm injuries was 8% (3 of 286). There was no difference ( p = 0.3) in mortality rate in SFL and non-SFL states. Table 3 highlights the differences in demographics and outcomes between patients in SFL and non-SFL states. Table 4 highlights the score, incidence of trauma, and firearm injuries in each state. California (score, 16) had the highest score and was the state with the strictest firearm laws, while Utah (score, 0) had the lowest scores and was the least strict firearm state. The mean and median among 44 states was 5 and median 4, respectively. Figure 1 demonstrates the distribution and rate of firearm injury in SFL and non-SFL states. Louisiana, a nonstrict gun law state, had the highest mean rate of firearm injuries (22.5 per 1,000), while New Jersey, an SFL state, had the lowest (1.6 per 1,000) mean rate of firearm accidents. States with non-SFLs had higher mean firearm injury rate per 1,000 trauma patients (mean [SD]: SFL, 2.2 [1.6]; 148
non-SFL, 5.9 [5.6]; p = 0. 001), compared with states with nonSFLs. On a simple linear regression analysis, we found that being in non-strict gun law state increased the mean firearm injury rate by 3.75 (A coefficient, 3.75; 95% confidence interval, 0.25Y7.25; p = 0.036).
DISCUSSION This study highlights that children-specific firearm legislations are effective in controlling the incidence of firearmrelated injuries among children. Children in non-SFL states are more likely to have a firearm-related injury in comparison with children in SFL states. Younger children were more likely to be involved in a firearm injury in non-SFL states compared with SFL states. We believe that improving and standardizing national and state firearm control laws with special focus on child safety laws are warranted. The United States has the highest rate of firearm-related deaths among children younger than 18 years in comparison with other industrialized countries.2,3 The current debate over firearm control focuses on the effectiveness of rigorous firearm control legislations. Advocates of increased legislative control claim that strict laws will reduce death rates and crime rates by restricting firearm possession by minors and by unreliable adults. However, opponents argue that strict laws do not prevent crime sufficiently to justify the restrictions placed on responsible citizens. We believe that the important point of this debate is not the number of legislations and formulating new legislations but understanding the true impact of the current legislations and how these legislations can safeguard the people they intended to protect. In our study, we classified states into SFL and non-SFL states based on children-specific firearm legislations in each state. We scored states based on the presence of children-specific firearm legislation in each state and assessed the cumulative impact of all the firearm legislations in a state on the incidence of firearm-related injuries. Studies assessing the impact of firearm legislature have routinely focused on individual components of the firearm laws and have failed to take into account the cumulative effects of these firearm legislations.8,12,13 Lee et al.8 in a review of firearm-related injuries in children assessed the impact TABLE 3. SFL Versus Non-SFL States Variables Age, mean (SD) e6, % (n) 7Y12, % (n) Q13, % (n) Sex, % (n) Males Females Race, % (n) White African American Hospital LOS, mean (SD) Mortality, % (n)
SFL (n = 66)
Non-SFL (n = 220)
p
15.4 (3.3) 3 (2) 9.1 (6) 87.9 (58)
14.3 (4.1) 6.8 (15) 16.4 (36) 76.8 (169)
0.02
87.9 (58) 12.1 (8)
83.6 (184) 16.4 (36)
40.9 (27) 22.7 (15) 4.4 (6.5) 4.5% (3)
40.5 (89) 21.8 (48) 7.3 (23.6) 9.1% (20)
0.4
0.1
0.08 0.3
LOS, length of stay.
* 2014 Lippincott Williams & Wilkins
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
J Trauma Acute Care Surg Volume 76, Number 1
Safavi et al.
TABLE 4. Incidence of Trauma and Firearm Injury in Each State SFL (n = 11) State California Connecticut Hawaii Illinois Massachusetts Maryland Michigan New Jersey New York Pennsylvania Rhode Island
Non-SFL (n = 33) Score
Trauma
Firearm
State
Score
Trauma
Firearm
State
Score
Trauma
Firearm
16 10 12 10 12 10 8 16 12 8 12
7,053 338 10 2,641 700 664 578 1,185 4,753 4,791 70
21 1 0 6 2 0 3 2 18 13 0
Arkansas Arizona Colorado Florida Georgia Iowa Indiana Kansas Kentucky Louisiana Maine Minnesota Missouri Montana North Carolina Nebraska
4 2 4 2 2 4 2 2 2 2 2 2 4 2 4 4
352 3,026 1,587 5,151 892 300 1,120 442 376 1,152 80 1,018 1,056 254 1,935 102
4 9 7 29 9 4 8 5 4 26 0 3 5 0 10 0
New Hampshire New Mexico Nevada Ohio Oklahoma Oregon South Carolina South Dakota Tennessee Texas Utah Virginia Vermont Washington West Virginia Wisconsin Wyoming
4 4 4 4 2 2 4 4 4 2 0 4 4 4 4 2 2
682 213 780 2,998 2,994 1,161 318 24 789 4,093 198 1,339 12 1,483 121 1,283 110
1 4 3 16 16 3 3 0 11 27 0 7 0 6 1 1 0
of child access protection, safe storage of guns, and stand-yourground laws on the incidence of firearm injuries in children. Sen and Panjamapirom12 in a national exploratory study assessed
the impact of only background checks on the incidence of firearm-related injury and mortality. In another study, Ludwig and Cook13 evaluated the impact of waiting periods and background
Figure 1. Distribution and rate of firearm injury in SFL and non-SFL states. * 2014 Lippincott Williams & Wilkins
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
149
J Trauma Acute Care Surg Volume 76, Number 1
Safavi et al.
checks for firearm purchases and demonstrated reduction in firearm suicide rates among individuals 55 years or older. A recently published study by Fleegler et al.9 is the only study that has assessed the true impact of all the firearm legislation in a state and the incidence of firearm-related injury. However, this study failed to take into account the firearm legislation specific to children and assess the impact of these legislation on the incidence of firearm injuries among children. Several studies have evaluated the association between firearm legislations and the incidence of firearm injuries and deaths in the United States.9,14,15 Rosengart et al.14 demonstrated an increased incidence of firearm homicides in states with no restrictions on carrying concealed weapons. Fleegler et al. demonstrated a decreased incidence of firearm-related casualties in states with higher number of firearm laws.14 In contrast, Lott15 did not find any difference in the incidence of firearm injuries in states with and without SFLs. However, none of these studies adequately assessed the cumulative impact of firearm legislations on the incidence of firearm-related injuries. In our study, we found that SFLs in the states were associated with a lower rate of firearm injuries in children. This variability in results among studies can be attributed to different methodologies of assessing firearm laws in each state and to the heterogeneity of the study population. We found some heterogeneity in firearm incidence rates among the states within each category of the legislative strength scores (e.g., Nebraska has nonstrict gun control laws and zero firearm incidence). We anticipated such heterogeneity and hence conducted this study assessing the firearm legislation across all the 44 states in the United States. The presence of a trend in higher mortality and longer stay in non-SFL states compared with SFL states despite the lack of statistical difference was most likely caused by our sample size, and it characterizes a Type II error. Our study has limitations. First, we were unable to control for the enforcement of firearm laws, which may vary between states. Second, the state gun control strength score used in our analysis has not been validated, and we are unaware of any such scoring systems that have been validated. To minimize bias in our state scoring, we gave an equal set of weights to each category to avoid biasing the results by our classification in favor or against gun control. Finally, certain types of state and local gun control legislation were not considered, and the comparison among the states did not take into account the additional gun regulations of local governments within the state. Nonetheless, this study is distinctive in the sense that it examines the cumulative impact of firearm legislation linked to pediatric population rather than single laws in a mixed adult and pediatric cohort.
AUTHORSHIP B.J., A.S., V.P., R.S.F., and P.R. designed this study. B.J., A.S., V.P., N.K., A.T., H.A., and R.S.F. searched the literature. B.J., A.S., V.P., A.T., H.A., and T.O. abstracted the data. B.J., A.S., V.P., H.A., R.S.F., and P.R. analyzed the data. All other authors participated in the data interpretation and manuscript preparation.
DISCLOSURE The authors declare no conflicts of interest.
REFERENCES 1. Center for Disease Control and Prevention. Youth risk behavior surveillanceVUnited States, 2011. MMWR Surveill Summ. 2012;61(4): 1Y162. 2. Fingerhut LA, Warner M. Injury chartbook. In: Health, United States, 1196-97. Hyattsville, MD: National Center for Health Statistics; 1997. 3. Richardson EG, Hemenway D. Homicide, suicide, and unintentional firearm fatality: comparing the United States with other high-income countries, 2003. J Trauma. 2011;70:238. 4. Web-based Injury Statistics Query and Reporting System. Centers for Disease Control and Prevention, National Center for Injury Prevention and Control. www.cdc.gov/ncipc/wisqars (Accessed on July 23, 2013). 5. Centers for Disease Control and Prevention. Injury prevention and control: Data and statistics (WISQARS). http://www.cdc.gov/injury/wisqars/index. html (Accessed on July 23, 2013). 6. Hutchinson A. Report of the National School Shield Task Force. Hutchinson A. April 2, 2013. http://www.nraschoolshield.com/NSS_Final_ FULL.pdf. Accessed July 5, 2013. 7. Vernick JS, Hepburn LM. Twenty thousand gun-control laws? In: Ludwig J, Cook PI, eds. Evaluating Gun Policy: Effects on Crime and Violence. Washington, DC: Brookings Institution Press; 2003:456. 8. Lee J, Moriarty KP, Tashjian DB, Patterson LA. Guns and states: Pediatric firearm injury. J Trauma Acute Care Surg. 2013;75(1):50Y53. 9. Fleegler EW, Lee LK, Monuteaux MC, Hemenway D, Mannix R. Firearm legislation and firearm-related fatalities in the United States. JAMA Intern Med. 2013;173(9):732Y740. 10. National Inpatient sample. http://www.hcup-us.ahrq.gov/nisoverview.jsp. Accessed July 5, 2013. 11. Brady Campaign to Prevent Gun Violence. Legislative scorecards. http:// www.bradycampaign.org/stategunlaws. Accessed July 4,2013. 12. Sen B, Panjamapirom A. State background checks for gun purchase and firearm deaths: an exploratory study. Preventive Medicine. 2012;55(4):346Y350. 13. Ludwig J, Cook PJ. Homicide and suicide rates associated with implementation of the Brady Handgun Violence Prevention Act. JAMA. 2000; 284(5):585Y591. 14. Rosengart M, Cummings P, Nathens A. An evaluation of state firearm regulations and homicide and suicide death rates. Inj Prev. 2005;11:77Y83. 15. Lott J Jr. More Guns, Less Crime: Understanding Crime and Gun Control Laws. 3rd ed. Chicago, IL: University of Chicago Press; 2010.
DISCUSSION CONCLUSION Firearm control legislations are effective in reducing the number of firearm injuries in children. Children living in SFL states are safer and are less likely to be involved in firearm accidents. Efforts to decrease pediatric firearm injuries should focus on improving and standardizing national firearm control laws aimed specifically at child safety. 150
Dr. Michael L. Nance (Philadelphia, Pennsylvania): I appreciate the opportunity to review this abstract and the manuscript. I have two questions for the authors. First, the authors used an un-validated method of scoring state firearms laws. They authors created their model that assigned an equal weight to the five categories of laws, such as assault weapons ban or safety training. The Brady Center, from whom the background data was obtained, has created their own score that assigns weights to the laws based on their belief that the legislation is of varying * 2014 Lippincott Williams & Wilkins
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
J Trauma Acute Care Surg Volume 76, Number 1
efficacy and, therefore, of varying significance. Why did the authors to use an un-validated, weighted score for their analyses? Second, the authors used the National Inpatient Sample, which included more than 60,000 injured children. Remarkably, there were only 286 firearm injuries reported. This small number was then further subdivided by the 44 states to determine rates of injury by state. Twenty-five of the 44 states reported fewer than five injuries. That seems unbelievably low and in those states just one or two more reported injuries would increase the reported rates quite significantly, by 100% or 200%. Can these numbers be believed? The data are very provocative. I think this is a very important line of research and I encourage them to continue their work. Dr. Arash Safavi (Tucson, Arizona): Thank you very much. In terms of your first question about using the scoring, we looked for a validated scoring system but we couldn’t find any. There were some scorings, for example by the Brady
Safavi et al.
Campaign, which is campaigning for restrictive gun control laws. Their scoring system includes adults and pediatric population and it’s not validated either. We chose not to use their scoring because, first of all, it does not specifically address children and, secondly, it is unvalidated and couldn’t add any value to the scoring we designed. In terms of your second question, the reason that some of the states did not have any firearm injuries in our study is mainly because NIS database does not collect every single patient and it records 20% of hospitalized patients. In addition, as I mentioned, the database only includes inpatients. For instance, if there was a patient who died at the scene, that’s not recorded in NIS. I think the most important things is the sample size and sample representation of the population. In other words, as long as you have a reliable sample such as NIS and you have enough sample size that give you enough power for your primary outcome, then the result of the study is reliable and that’s what we had.
* 2014 Lippincott Williams & Wilkins
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
151
