- Email: [email protected]
Factors associated with law enforcement–related use-of-force injury
American Journal of Emergency Medicine (2012) 30, 526–531
www.elsevier.com/locate/ajem
Original Contribution
Factors associated with law enforcement–related use-of-force injury☆ Edward M. Castillo PhD, MPH ⁎, Nitin Prabhakar, Bethi Luu MPH Department of Emergency Medicine, University of California, San Diego Medical Center, San Diego, CA, USA Received 21 October 2010; revised 11 January 2011; accepted 11 January 2011
Abstract Objective: Use-of-force (UOF) techniques are used by law enforcement to gain control of noncompliant subjects. The purpose of this study was to assess factors associated with subject and deputy injuries during law enforcement UOF. Methods: This is a retrospective study of nonlethal UOF events from January to June 2009 by a single law enforcement agency serving a population of 3 million. A standard data collection tool, which included basic demographic data, the type of force used, subject response, and if there were any injuries to the subject or deputies involved, was used by deputies for each UOF event. Descriptive statistics were used to describe the specific subject and incident details. Univariate and multivariate analysis was used to identify factors potentially associated with subject and deputy injuries. Results: There were 1174 UOF incidents recorded during the study period. A total of 282 incidents (24%) involved no physical force, 135 (11.5%) involved less lethal methods, 620 (52.8%) involved other physical restraint methods, and 137 (11.7%) involved both less lethal and other physical methods. Factors with the largest independent associations with subject injury were physical resistance by the subject (odds ratio [OR], 2.65; 95% confidence interval [CI], 1.49-4.74) and force used to prevent a violent felony (OR, 2.15; 95% CI, 1.24-3.71). When the subject had a weapon (OR, 4.15; 95% CI, 1.5311.23) and physical resistance by the subject (OR, 4.15; 95% CI, 1.24-13.94) had the largest associations with deputy injury. Conclusions: This study identifies situational characteristics potentially associated with subject and deputy injuries during UOF events. © 2012 Elsevier Inc. All rights reserved.
1. Introduction Use-of-force (UOF) techniques are used by law enforcement agencies as a mean to gain control of noncompliant subjects. Use of force includes the continuum of force ☆
Presentation. A poster of the initial results of this study was presented as a poster at the American College of Emergency Physician Research Forum in Las Vegas, NV, on September 2010. ⁎ Corresponding author. Department of Emergency Medicine, UC San Diego Medical Center 200, San Diego, CA 92103, USA. Tel.: +1 619 543 6463; fax: +1 619 543 3115. E-mail address: [email protected] (E.M. Castillo). 0735-6757/$ – see front matter © 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2011.01.017
ranging from simply the psychologic concept of an officer being present and verbal commands to displaying a conducted energy device (CED) or baton, going hands on a subject, striking a subject with hands or an object, using oleoresin capsicum (OC) or pepper spray, to the use of a CED or firearm. The most common and successful types of force used in this continuum are nonphysical instruction and verbal commands. When physical force must be used to restrain a subject, methods aimed to limit injuries in both the subject and deputies are often used. Many law enforcement agencies are moving toward more “less lethal” means of force when possible to control a subject, such as OC spray
Use of force injuries and CEDs. Less lethal means, also referred to as “nonlethal,” are methods that are used to gain control of a suspect while minimizing fatalities. The majority of research in the past decade investigating UOF events has focused on the safety of these specific methods in controlled situations [1-9], but only a few have investigated the UOF in the field [10-16], and most of them focus on CED use and safety [10,12-15]. Recent studies investigating law enforcement UOF in the field have focused on the impact of new less lethal weapons on injury risk as well as the general frequency of injury. A study by Jenkinson et al [10] investigated the relative risk of injuries to officers and suspects of police UOF options, specifically with the introduction of CEDs. Another similar article by Smith et al [11] studied the risk of injury by UOF methods in 2 separate law enforcement agencies. Both of these studies concluded that CED use was associated with a lower risk of officer and subject injuries compared to other UOF methods. Another more recent article by MacDonald et al [14] investigated the effect of less lethal weapons on injuries in police UOF events in 12 law enforcement agencies. Although this study included a large number of agencies and events, they were unable to evaluate the impact of specific situational aspects of the UOF incident. The only study that was able to report the specific situational aspects of the UOF incident was that of Strote et al [16]. Their study investigated UOF injures and medical outcomes recorded by the Seattle Police Department, but their focus was on specific outcomes and not the factors associated with the actual events leading to injury. The purpose of this study is to assess the factors associated with subject and deputy injuries during law enforcement UOF events.
2. Methods This is a retrospective cross-sectional study using nonlethal UOF data collected by the San Diego County Sheriff's Department, which is the fourth largest sheriff's department in the nation. It covers approximately 4200 square miles and serves a population of about 3 million. The department has 8 stations and 10 substations. They provide law enforcement services to all unincorporated areas of the county and 9 cities and provide security for 10 courthouses. The study was approved by the University of California, San Diego Human Research Protection Program.
2.1. Data collection and processing All nonlethal UOF incidents recorded by the San Diego County Sheriff's Department from January 2009 through June 2009 were included in this study. The department collects data on every UOF event by its personnel. Deputies complete a standard incident report when UOF is implemented that include demographic and situational information regarding the incident. Data were abstracted from these
527 reports and entered into a Microsoft Access 2007 Database (Microsoft Inc, Redmond, WA) by a single trained reviewer. The abstraction tool included specific information regarding subject age and sex, the nature of the call, behavioral characteristics of the subject, the primary reason for using force, if the subject had a weapon and which type (gun, knife, weapon of opportunity, and other), the number of deputies on the scene, level of resistance displayed by the subject, the time it took to restrain the subject, the method of force used along with the number of contacts and duration if applicable, and if the subject or deputy was injured during the incident.
2.2. Data analysis The main outcomes were deputy-recorded subject and deputy injuries that occurred during a law enforcement UOF incident. Behavioral characteristics were categorized into suspected alcohol or drug intoxication or not, as reported by the officers. A suspect with a weapon was categorized as weapon yes or no. Level of resistance displayed by the subject was categorized into 2 groups that included no physical resistance (none, psychologic intimidation, verbal noncompliance, and passive resistance) and any type of physical resistance (active resistance, assaultive behavior, or aggravated active aggression, which is when a weapon is used by the subject). The time it took to restrain the subject was dichotomized into less than or greater than 1 minute. Use-of-force methods were categorized as nonphysical, less lethal methods (OC spray, pepperball launcher, CEDs, rifles and shotguns using less lethal shells, or riot shield), and other physical methods (grab, takedown, control hold, pressure point, strike, carotid, impact weapon such as a baton, canine, or the forcible stop of a fleeing vehicle). Descriptive statistics, including frequencies and means, were used to describe specific subject and incident details including the UOF method used. Among the incidents where physical methods occurred, excluding incidents where both less lethal and other physical methods were reported, differences in suspect and situational characteristics by injury status for subjects and deputies were assessed with a t test for age or χ2 or Fisher exact test for categorical variables. Independent associations with injury for suspects and deputies were assessed using logistic regression. Variables with a P value b.05 in the univariate analysis were included in the regression models. Odds ratios (ORs), 95% confidence intervals (CIs), and associated P values are presented. Data were analyzed using SPSS version 17 (SPSS Inc, Chicago, IL).
3. Results 3.1. Characteristics of study subjects There were a total of 1174 UOF incidents recorded by the San Diego County Sheriff's Department during the study
528
E.M. Castillo et al.
period. A disturbance was the most common reason for a deputy response (64.8%). A total of 282 incidents (24%) involved no physical force, 135 (11.5%) involved less lethal methods, 620 (52.8%) involved other physical methods, and 137 (11.7%) involved both less lethal and other physical methods. Characteristics of UOF events involving the 755 subjects where less lethal method or other physical methods occurred are reported in Table 1. The mean age of subjects was 30.1 years (SD, 12.2), and 572 (75.8%) were male. Suspected substance use was noted in 375 incidents (49.7%), the subject had a weapon in 33 cases (4.4%), and physical resistance displayed by the subject was reported in 618 (81.9%) of the 755 incidents. Additional force was required after initial force failed to restrain the subject in 213 (28.4%) incidents, and it took less than 1 minute to restrain 691 (91.5%) of the subjects. Situational characteristics for the 755 physical law enforcement UOF events are reported in Table 2. A disturbance was the most common reason for response and was noted in 503 incidents (66.6%), followed by vehicle related, with 67 (8.9%). Domestic violence was noted for 65 calls (8.6%). To “defend self or another” and “subject's safety” were the most common primary reasons for the UOF and were noted in 457 (60.5%) and 338 (44.8%) incidents, respectively. The most common techniques used to control the subject were other physical methods, such as empty-hand control without using devices, tools, or weapons, and occurred in 620 events (82.1%).
4. Main results Differences in subject and situational characteristics by subject and deputy injuries are presented in Table 3. Subjects Table 1 Characteristics of 755 physical law enforcement UOF events, January 1 to June 30, 2009 Characteristics Age (y) Characteristics Sex Male Female Suspected substance use Subject had weapon Physical resistance by subject Additional force required a Time to restrain subject b1 min ≥1 min Subject injured b Deputy injured b a b
6 missing. When physical force was used.
Mean
SD
30.1
12.2
n
%
572 183 375 33 618 213
75.8 24.2 49.7 4.4 81.9 28.4
691 64 165 64
91.5 8.5 21.9 8.5
Table 2 Situational characteristics of 755 physical law enforcement UOF events, January 1 to June 30, 2009
Nature of call Assault Breaking and entering Disturbance Domestic violence Vehicle related No. of deputies on scene a 1-2 3-4 N5 Primary reason for using force To effect an arrest To prevent a violent felony To defend self or other Subject's safety To prevent escape/evasion During high-risk incident To effect a lawful detention Type of force used Less lethal method Other method
n
%
22 19 503 65 67
2.9 2.5 66.6 8.6 8.9
360 217 176
47.8 28.8 23.4
271 79 457 338 249 136 206
35.9 10.5 60.5 44.8 33.0 18.0 27.3
135 620
17.9 82.1
Nature of call and primary reason for using force are not mutually exclusive categories. a 2 missing.
and deputies who were injured during UOF events were more likely to be in situations where the subject displayed physical resistance, required additional force after initial force failed to restrain the subject, when more deputies were on the scene, and when the deputy used force to defend themselves or others (P b .05). Cases in which subjects were injured, but deputies were not, occurred more often in incidents where the subject was male, when the time needed to restrain the subject was greater than 1 minute, during disturbance calls, when force was used to prevent a violent felony and during a high-risk event that may cause someone injury or death, and when less lethal methods were used by the deputy to gain compliance (P b .05). Injuries among subjects and deputies were less common during incidents that involved suspected substance use (P b .05). Subjects and deputies who were and were not injured did not differ by suspect age, suspected subject substance use, if the suspect had a weapon, an assault call, breaking and entering call, a domestic violence call, or using force to either effect an arrest or prevent escape (P N .05). Regression analysis examining the associations of subject and situational characteristics and the risk of injury among subjects and deputies are shown in Table 4. After adjusting for covariates, the variables with the largest associations with subject injury were physical resistance by the subject (OR, 2.65; 95% CI, 1.49-4.74) and to prevent a violent felony (OR, 2.15; 95% CI, 1.24-3.71). Incidents where the subject was not suspected of substance use were also associated with subject injury (OR, 1.96; 95% CI, 1.34-2.86).The time to restrain the
Use of force injuries
529
Table 3 Differences in subject and situational characteristics by subject and deputy injuries among 755 physical law enforcement UOF events, January 1 to June 30, 2009 Subject injured
Age (y) b25 ≥25 Male sex Suspected substance use Subject had weapon Physical resistance by subject Additional force required Time to restrain subject b1 min ≥1 min Nature of call Assault Breaking and entering Disturbance Domestic violence Vehicle related No. of deputies on scene 1-2 3-4 N5 Primary reason for using force Effect an arrest Prevent a violent felony Defend self or other Subject's safety Prevent escape/evasion During high-risk incident Effect a lawful detention Technique used to gain compliance Less lethal method Other method
Deputy injury
Yes (n = 165)
No (n = 590)
n
%
n
%
68 97 141 64 9 146 66
41.2 58.8 85.5 38.8 5.5 88.5 40.0
235 355 431 311 24 472 147
39.8 60.2 73.1 52.7 4.1 80.0 25.2
143 22
86.7 13.3
548 42
92.9 7.1
3 5 123 9 11
1.8 3.0 75.5 5.5 6.7
19 14 380 56 56
3.2 2.4 64.4 9.5 9.5
61 54 49
37.2 32.9 29.9
299 163 127
50.8 27.7 21.6
55 31 116 72 45 45 42
33.3 18.8 70.3 43.6 27.3 27.3 25.5
216 48 341 266 204 91 164
36.6 8.1 57.8 45.1 34.6 15.4 27.8
44 121
26.7 73.3
91 499
15.4 84.6
P
Yes (n = 64)
No (n = 691)
n
%
n
%
33 31 49 21 6 61 33
51.6 48.4 76.6 32.8 9.4 95.3 51.6
270 421 523 354 27 557 180
39.1 60.9 75.7 51.2 3.9 80.6 26.3
55 9
85.9 14.1
636 55
92.0 8.0
4 3 47 2 1
6.3 4.7 73.4 3.1 1.6
18 16 456 63 66
2.6 2.3 66.0 9.1 9.6
20 18 25
31.7 28.6 39.7
340 199 151
49.3 28.8 21.9
23 5 50 24 24 7 14
35.9 7.8 78.1 37.5 37.5 10.9 21.9
248 74 407 314 225 129 192
35.9 10.7 58.9 45.4 32.6 18.7 27.8
4 60
6.3 93.8
131 560
19.0 81.0
.749
subject, a disturbance call, the number of deputies at the scene, and the primary reason for UOF was to defend self or others were no longer significant and were removed from the model. After adjusting for covariates, the variables with the largest associations with deputy injury were when the subject had a weapon (OR, 4.15; 95% CI, 1.53-11.23) and physical resistance by the subject (OR, 4.15; 95% CI, 1.24-13.94). As with subject injury, incidents where the subject was not suspected of substance use were also associated with deputy injury (OR, 2.39; 95% CI, 1.35-4.21). The number of deputies at the scene and the type of force used by the deputy were no longer significant in the model and were removed.
5. Discussion This study reports factors associated with suspect and deputy injuries during UOF events. The range of force used
.001 .002 .441 .012 b.001 .011
.440 .582 .015 .117 .259 .007
.438 b.001 .004 .741 .078 b.001 .550 .001
P .051
.876 .005 .041 0.002 b.001 .094
.108 .246 .227 .157 .035 .003
.994 .668 .003 .222 .421 .124 .310 .010
by a deputy to gain control of a suspect can range from law enforcement physical presence to displaying or using a weapon. This study focused on situations where only physical force was used during an incident. Of the 755 uses of physical force over the 6-month period that we analyzed, a few factors were identified that were independently associated with both subject and deputy injuries and included if a person was not suspected of substance use, if the subjects actively resisted, and if additional force was required because initial UOF did not work. A male subject, when force is used to prevent a violent felony, a high-risk situation that may cause someone injury or death, and when a less lethal method was used were also independently associated with injury among subjects as was when the suspect had a weapon and to defend self or other for deputies. These results suggest that the odds of injury among subjects and deputies increases the longer a situation is out of control and provides further justification for policies that
530
E.M. Castillo et al.
Table 4 Associations of select factors with risk of injury among subjects and deputies during 755 physical law enforcement UOF events, January 1 to June 30, 2009 Subject injured Characteristic Sex (male/female) Weapon (yes/no) Possible substance use (no/yes) Physical resistance by subject (yes/no) Additional force required (yes/no) Prevent a violent felony (yes/no) Defend self or other (yes/no) High-risk situation (yes/no) Type of force (less lethal/other) a
OR 1.74
Deputy injury
95% CI (lower, higher) 1.06, 2.85
P .028
a
1.96 2.65 2.05 2.15
2.86 4.74 3.03 3.71
b.001 .001 b.001 .006
1.05, 2.67 1.25, 3.30
.031 .004
1.34, 1.49, 1.39, 1.24,
a
1.67 2.03
OR
95% CI (lower, higher)
P
1.53, 1.35, 1.24, 1.63,
.005 .003 .021 b.001
a
4.15 2.39 4.15 2.81
11.23 4.21 13.94 4.84
a
1.97
1.05, 3.70
.036
a a
Not included in the model because the characteristics were not significant in the univariate comparisons or no longer significant in regression model.
support the use of less lethal, yet effective, tools to gain control of suspects in a timely manner. These results also provide more information regarding the type of calls where injuries may be more likely to occur, such as with disturbances. However, although it would be intuitive to think potential alcohol intoxication or substance use may lead to situations where injuries are more common, it does not appear to be associated with subject or deputy injury during UOF events in this population and should be studied further. Although this study focused on all types of nonfatal injuries, a recent study by Ho et al [17] suggests a potential increase in the risk of sudden death during incidents of physical resistance or fleeing. In their study, the authors conclude that simulations of physical resistance had the greatest changes in acidosis and catecholamine markers, which could be potentially contributing or causal mechanisms in sudden arrest-related deaths. There are only a few recently published studies that have investigated the association between law enforcement UOF events and injury in the field [10-16], and most of them focus on the use of CEDs [10,12-15]. These less lethal UOF weapons have become widely used by many law enforcement agencies throughout the world and have been shown to cause minimal injury to subjects [13-15] and decrease injury among deputies [10,14]. Although the current study did not focus specifically on these devices, they are currently used by the San Diego County Sheriff's Department and were the most often used less lethal weapon system used by the department, comprising over 70% of this type of force. However, although the use of less lethal weapons was associated with subject injury, it was not possible to determine the cause or extent of injury from the current data. Department policy requiring all CED activations to be assessed by a physician could overinflate the odds of injury from this method. In addition, the use of these methods did not show to be protective against deputy injury. This was likely because of the relatively small number of uses and deputy injuries and not the lack of benefit of these systems. A study by Smith et al [11] investigated the impact of various UOF methods and resistance on officer and suspect
injuries. In this study, investigators obtained data on 1645 UOF incidents from 2 different law enforcement agencies and modeled the odds of injury and severity of injury. They reported that CED was associated with decreased odds of suspect injury in 1 agency, whereas the use of OC spray was associated with decreased odds of suspect injury in the other. However, the use of hands on tactics to control a suspect was associated with increased odds of injury among both the suspect and deputy. Although the study provides valuable information regarding the odds of injury by UOF method, they did not control for the situation characteristics that potentially contribute to injury such as type of resistance or type of call. The most comprehensive study investigating the law enforcement UOF safety and injury was done by Strote et al [16]. This was a retrospective study investigating UOF events by the Seattle Police Department among 888 subjects. In this study, they reported 187 injured subjects (21.1%), which was similar to the injury rate (20.5%) found in the current study. Also in both studies, confirmed or suspected substance use and mental impairment were relatively common. However, the study by Strote et al defined injury as being seen in an emergency department within 24 hours of the incident. In our study, we defined injury as deputy indicated regardless of treatment and included those who were treated for their injuries and those who were not. This study has limitations that must be considered when interpreting the results. As a retrospective data review, data elements were not prospectively collected, and missing or incomplete data may introduce bias into the results. However, the UOF report form completed by deputies includes standard data elements for these types of events and is required for each incident immediately after the incident. Data could also be misinterpreted or miscoded during the abstraction process. To limit this, data were abstracted by a trained researcher and completed on site with San Diego County Sheriff's Department personnel available for assistance and interpretation when needed. The results may not be generalizable to other populations, as data were abstracted from only one law enforcement agency that spans a wide
Use of force injuries enforcement area that includes rural and urban locations. In addition, the records used for this study were all based on deputy reports and specific data elements, including the true cause or extent of an injury could not be confirmed. Although UOF events are recorded on standardized report forms, confirmation of specific factors such as suspected substance use could not be confirmed, or interpretation of what is defined as an injury could not be determined. The result of sampling bias due to deputy assignments to higher call areas with an increased potential for UOF incidents could also not be controlled for or assessed. In conclusion, this study identifies situational characteristics potentially associated with subject and deputy injuries during UOF events. Although causality cannot be demonstrated in this study and temporal factors cannot be assessed, these results suggest that quick control of suspects may decrease suspect and deputy injuries during these situations. However, more studies on the frequency of law enforcement UOF and medical outcomes are needed to improve our understanding of these situations and related injuries.
Acknowledgments The authors thank the San Diego County Sheriff's Department for their participation in this project.
References [1] Chan TC, Vilke GM, Neuman T, Clausen J, Schmidt P, Snowden T, et al. Does oleoresin capsicum “pepper spray” exposure impact cardiovascular function in human subjects? Acad Emerg Med 2001;8(5):442. [2] Ho JD, Miner JR, Lakireddy DR, et al. Cardiovascular and physiologic effects of conducted electrical weapon discharge in resting adults. Acad Emerg Med 2006;13:589-95. [3] Ho JD, Dawes DM, Bultman LL, Thacker JL, Skinner LD, Bahr JM, et al. Respiratory effect of prolonged electrical weapon application on human volunteers. Acad Emerg Med 2007;14:197-201.
531 [4] Vilke GM, Sloane CM, Bouton KD, Kolkhorst FW, Levine SD, Neuman TS, et al. Physiological effects of a conducted electrical weapon on human subjects. Ann Emerg Med 2007;50(5):569-75. [5] Vilke GM, Sloane C, Levine S, Neuman T, Castillo E, Chan TC. Twelve-lead electrocardiogram monitoring of subjects before and after voluntary exposure to the Taser X26. Am J Emerg Med 2008;26(1): 1-4. [6] Sloane CM, Chan TC, Levine SD, Dunford JV, Neuman T, Vilke GM, et al. Measurement of subjects exposed to the Taser X-26(R). J Emerg Med 2008;35(1):29-32. [7] Dawes DM, Ho JD, Johnson MA, Lundin E, Janchar TA, Miner JR. 15-second conducted electrical weapon exposure does not cause core body temperature elevation in non-environmentally stressed resting adults. Forensic Sci Int 2008;176:253-7. [8] Ho JD, Dawes DM, Bultman LL, Moscati RM, Janchar TA, Miner JR. Prolonged TASER use on exhausted humans does not worsen markers of acidosis. Am J Emerg Med 2009;27(4):413-8. [9] Vilke GM, Sloane CM, Suffecool A, Kolkhorst FW, Neuman TS, Castillo EM, et al. Physiologic effects of the TASER after exercise. Acad Emerg Med 2009;16(8):704-10. [10] Jenkinson E, Neeson C, Bleetman A. The relative risk of police use-offorce options: evaluating the potential for deployment of electronic weaponry. J Clin Forensic Med 2006;13(5):229-41. [11] Smith MR, Kaminski RJ, Rojek J, Alpert JP, Mathis J. The impact of conducted energy devices and other types of force and resistance on officer and suspect injuries. Policing 2007;30(3):423-46. [12] Vilke GM, Johnson 3rd WD, Castillo EM, Sloane C, Chan TC. Tactical and subject considerations of in-custody deaths proximal to use of conductive energy devices. Am J Forensic Med Pathol 2009;30(1):23-5. [13] Bozeman WP, Hauda 2nd WE, Heck JJ, Graham Jr DD, Martin BP, Winslow JE. Safety and injury profile of conducted electrical weapons used by law enforcement officers against criminal suspects. Ann Emerg Med 2009;53(4):480-9. [14] MacDonald JM, Kaminski RJ, Smith MR. The effect of less-lethal weapons on injuries in police use-of-force events. Am J Public Health 2009;99(12):2268-74. [15] Strote J, Walsh M, Angelidis M, Basta A, Hutson HR. Conducted electrical weapon use by law enforcement: an evaluation of safety and injury. J Trauma 2010;68(5):1239-46. [16] Strote J, Verzemnieks E, Walsh M, Hutson HR. Use of force by law enforcement: an evaluation of safety and injury. J Trauma 2010;69(5): 1288-93. [17] Ho JD, Dawes DM, Nelson RS, Lundin EJ, Ryan FJ, Overton KG, et al. Acidosis and catecholamine evaluation following simulated law enforcement “use of force” encounters. Acad Emerg Med 2010;17(7): e60-8.
www.elsevier.com/locate/ajem
Original Contribution
Factors associated with law enforcement–related use-of-force injury☆ Edward M. Castillo PhD, MPH ⁎, Nitin Prabhakar, Bethi Luu MPH Department of Emergency Medicine, University of California, San Diego Medical Center, San Diego, CA, USA Received 21 October 2010; revised 11 January 2011; accepted 11 January 2011
Abstract Objective: Use-of-force (UOF) techniques are used by law enforcement to gain control of noncompliant subjects. The purpose of this study was to assess factors associated with subject and deputy injuries during law enforcement UOF. Methods: This is a retrospective study of nonlethal UOF events from January to June 2009 by a single law enforcement agency serving a population of 3 million. A standard data collection tool, which included basic demographic data, the type of force used, subject response, and if there were any injuries to the subject or deputies involved, was used by deputies for each UOF event. Descriptive statistics were used to describe the specific subject and incident details. Univariate and multivariate analysis was used to identify factors potentially associated with subject and deputy injuries. Results: There were 1174 UOF incidents recorded during the study period. A total of 282 incidents (24%) involved no physical force, 135 (11.5%) involved less lethal methods, 620 (52.8%) involved other physical restraint methods, and 137 (11.7%) involved both less lethal and other physical methods. Factors with the largest independent associations with subject injury were physical resistance by the subject (odds ratio [OR], 2.65; 95% confidence interval [CI], 1.49-4.74) and force used to prevent a violent felony (OR, 2.15; 95% CI, 1.24-3.71). When the subject had a weapon (OR, 4.15; 95% CI, 1.5311.23) and physical resistance by the subject (OR, 4.15; 95% CI, 1.24-13.94) had the largest associations with deputy injury. Conclusions: This study identifies situational characteristics potentially associated with subject and deputy injuries during UOF events. © 2012 Elsevier Inc. All rights reserved.
1. Introduction Use-of-force (UOF) techniques are used by law enforcement agencies as a mean to gain control of noncompliant subjects. Use of force includes the continuum of force ☆
Presentation. A poster of the initial results of this study was presented as a poster at the American College of Emergency Physician Research Forum in Las Vegas, NV, on September 2010. ⁎ Corresponding author. Department of Emergency Medicine, UC San Diego Medical Center 200, San Diego, CA 92103, USA. Tel.: +1 619 543 6463; fax: +1 619 543 3115. E-mail address: [email protected] (E.M. Castillo). 0735-6757/$ – see front matter © 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2011.01.017
ranging from simply the psychologic concept of an officer being present and verbal commands to displaying a conducted energy device (CED) or baton, going hands on a subject, striking a subject with hands or an object, using oleoresin capsicum (OC) or pepper spray, to the use of a CED or firearm. The most common and successful types of force used in this continuum are nonphysical instruction and verbal commands. When physical force must be used to restrain a subject, methods aimed to limit injuries in both the subject and deputies are often used. Many law enforcement agencies are moving toward more “less lethal” means of force when possible to control a subject, such as OC spray
Use of force injuries and CEDs. Less lethal means, also referred to as “nonlethal,” are methods that are used to gain control of a suspect while minimizing fatalities. The majority of research in the past decade investigating UOF events has focused on the safety of these specific methods in controlled situations [1-9], but only a few have investigated the UOF in the field [10-16], and most of them focus on CED use and safety [10,12-15]. Recent studies investigating law enforcement UOF in the field have focused on the impact of new less lethal weapons on injury risk as well as the general frequency of injury. A study by Jenkinson et al [10] investigated the relative risk of injuries to officers and suspects of police UOF options, specifically with the introduction of CEDs. Another similar article by Smith et al [11] studied the risk of injury by UOF methods in 2 separate law enforcement agencies. Both of these studies concluded that CED use was associated with a lower risk of officer and subject injuries compared to other UOF methods. Another more recent article by MacDonald et al [14] investigated the effect of less lethal weapons on injuries in police UOF events in 12 law enforcement agencies. Although this study included a large number of agencies and events, they were unable to evaluate the impact of specific situational aspects of the UOF incident. The only study that was able to report the specific situational aspects of the UOF incident was that of Strote et al [16]. Their study investigated UOF injures and medical outcomes recorded by the Seattle Police Department, but their focus was on specific outcomes and not the factors associated with the actual events leading to injury. The purpose of this study is to assess the factors associated with subject and deputy injuries during law enforcement UOF events.
2. Methods This is a retrospective cross-sectional study using nonlethal UOF data collected by the San Diego County Sheriff's Department, which is the fourth largest sheriff's department in the nation. It covers approximately 4200 square miles and serves a population of about 3 million. The department has 8 stations and 10 substations. They provide law enforcement services to all unincorporated areas of the county and 9 cities and provide security for 10 courthouses. The study was approved by the University of California, San Diego Human Research Protection Program.
2.1. Data collection and processing All nonlethal UOF incidents recorded by the San Diego County Sheriff's Department from January 2009 through June 2009 were included in this study. The department collects data on every UOF event by its personnel. Deputies complete a standard incident report when UOF is implemented that include demographic and situational information regarding the incident. Data were abstracted from these
527 reports and entered into a Microsoft Access 2007 Database (Microsoft Inc, Redmond, WA) by a single trained reviewer. The abstraction tool included specific information regarding subject age and sex, the nature of the call, behavioral characteristics of the subject, the primary reason for using force, if the subject had a weapon and which type (gun, knife, weapon of opportunity, and other), the number of deputies on the scene, level of resistance displayed by the subject, the time it took to restrain the subject, the method of force used along with the number of contacts and duration if applicable, and if the subject or deputy was injured during the incident.
2.2. Data analysis The main outcomes were deputy-recorded subject and deputy injuries that occurred during a law enforcement UOF incident. Behavioral characteristics were categorized into suspected alcohol or drug intoxication or not, as reported by the officers. A suspect with a weapon was categorized as weapon yes or no. Level of resistance displayed by the subject was categorized into 2 groups that included no physical resistance (none, psychologic intimidation, verbal noncompliance, and passive resistance) and any type of physical resistance (active resistance, assaultive behavior, or aggravated active aggression, which is when a weapon is used by the subject). The time it took to restrain the subject was dichotomized into less than or greater than 1 minute. Use-of-force methods were categorized as nonphysical, less lethal methods (OC spray, pepperball launcher, CEDs, rifles and shotguns using less lethal shells, or riot shield), and other physical methods (grab, takedown, control hold, pressure point, strike, carotid, impact weapon such as a baton, canine, or the forcible stop of a fleeing vehicle). Descriptive statistics, including frequencies and means, were used to describe specific subject and incident details including the UOF method used. Among the incidents where physical methods occurred, excluding incidents where both less lethal and other physical methods were reported, differences in suspect and situational characteristics by injury status for subjects and deputies were assessed with a t test for age or χ2 or Fisher exact test for categorical variables. Independent associations with injury for suspects and deputies were assessed using logistic regression. Variables with a P value b.05 in the univariate analysis were included in the regression models. Odds ratios (ORs), 95% confidence intervals (CIs), and associated P values are presented. Data were analyzed using SPSS version 17 (SPSS Inc, Chicago, IL).
3. Results 3.1. Characteristics of study subjects There were a total of 1174 UOF incidents recorded by the San Diego County Sheriff's Department during the study
528
E.M. Castillo et al.
period. A disturbance was the most common reason for a deputy response (64.8%). A total of 282 incidents (24%) involved no physical force, 135 (11.5%) involved less lethal methods, 620 (52.8%) involved other physical methods, and 137 (11.7%) involved both less lethal and other physical methods. Characteristics of UOF events involving the 755 subjects where less lethal method or other physical methods occurred are reported in Table 1. The mean age of subjects was 30.1 years (SD, 12.2), and 572 (75.8%) were male. Suspected substance use was noted in 375 incidents (49.7%), the subject had a weapon in 33 cases (4.4%), and physical resistance displayed by the subject was reported in 618 (81.9%) of the 755 incidents. Additional force was required after initial force failed to restrain the subject in 213 (28.4%) incidents, and it took less than 1 minute to restrain 691 (91.5%) of the subjects. Situational characteristics for the 755 physical law enforcement UOF events are reported in Table 2. A disturbance was the most common reason for response and was noted in 503 incidents (66.6%), followed by vehicle related, with 67 (8.9%). Domestic violence was noted for 65 calls (8.6%). To “defend self or another” and “subject's safety” were the most common primary reasons for the UOF and were noted in 457 (60.5%) and 338 (44.8%) incidents, respectively. The most common techniques used to control the subject were other physical methods, such as empty-hand control without using devices, tools, or weapons, and occurred in 620 events (82.1%).
4. Main results Differences in subject and situational characteristics by subject and deputy injuries are presented in Table 3. Subjects Table 1 Characteristics of 755 physical law enforcement UOF events, January 1 to June 30, 2009 Characteristics Age (y) Characteristics Sex Male Female Suspected substance use Subject had weapon Physical resistance by subject Additional force required a Time to restrain subject b1 min ≥1 min Subject injured b Deputy injured b a b
6 missing. When physical force was used.
Mean
SD
30.1
12.2
n
%
572 183 375 33 618 213
75.8 24.2 49.7 4.4 81.9 28.4
691 64 165 64
91.5 8.5 21.9 8.5
Table 2 Situational characteristics of 755 physical law enforcement UOF events, January 1 to June 30, 2009
Nature of call Assault Breaking and entering Disturbance Domestic violence Vehicle related No. of deputies on scene a 1-2 3-4 N5 Primary reason for using force To effect an arrest To prevent a violent felony To defend self or other Subject's safety To prevent escape/evasion During high-risk incident To effect a lawful detention Type of force used Less lethal method Other method
n
%
22 19 503 65 67
2.9 2.5 66.6 8.6 8.9
360 217 176
47.8 28.8 23.4
271 79 457 338 249 136 206
35.9 10.5 60.5 44.8 33.0 18.0 27.3
135 620
17.9 82.1
Nature of call and primary reason for using force are not mutually exclusive categories. a 2 missing.
and deputies who were injured during UOF events were more likely to be in situations where the subject displayed physical resistance, required additional force after initial force failed to restrain the subject, when more deputies were on the scene, and when the deputy used force to defend themselves or others (P b .05). Cases in which subjects were injured, but deputies were not, occurred more often in incidents where the subject was male, when the time needed to restrain the subject was greater than 1 minute, during disturbance calls, when force was used to prevent a violent felony and during a high-risk event that may cause someone injury or death, and when less lethal methods were used by the deputy to gain compliance (P b .05). Injuries among subjects and deputies were less common during incidents that involved suspected substance use (P b .05). Subjects and deputies who were and were not injured did not differ by suspect age, suspected subject substance use, if the suspect had a weapon, an assault call, breaking and entering call, a domestic violence call, or using force to either effect an arrest or prevent escape (P N .05). Regression analysis examining the associations of subject and situational characteristics and the risk of injury among subjects and deputies are shown in Table 4. After adjusting for covariates, the variables with the largest associations with subject injury were physical resistance by the subject (OR, 2.65; 95% CI, 1.49-4.74) and to prevent a violent felony (OR, 2.15; 95% CI, 1.24-3.71). Incidents where the subject was not suspected of substance use were also associated with subject injury (OR, 1.96; 95% CI, 1.34-2.86).The time to restrain the
Use of force injuries
529
Table 3 Differences in subject and situational characteristics by subject and deputy injuries among 755 physical law enforcement UOF events, January 1 to June 30, 2009 Subject injured
Age (y) b25 ≥25 Male sex Suspected substance use Subject had weapon Physical resistance by subject Additional force required Time to restrain subject b1 min ≥1 min Nature of call Assault Breaking and entering Disturbance Domestic violence Vehicle related No. of deputies on scene 1-2 3-4 N5 Primary reason for using force Effect an arrest Prevent a violent felony Defend self or other Subject's safety Prevent escape/evasion During high-risk incident Effect a lawful detention Technique used to gain compliance Less lethal method Other method
Deputy injury
Yes (n = 165)
No (n = 590)
n
%
n
%
68 97 141 64 9 146 66
41.2 58.8 85.5 38.8 5.5 88.5 40.0
235 355 431 311 24 472 147
39.8 60.2 73.1 52.7 4.1 80.0 25.2
143 22
86.7 13.3
548 42
92.9 7.1
3 5 123 9 11
1.8 3.0 75.5 5.5 6.7
19 14 380 56 56
3.2 2.4 64.4 9.5 9.5
61 54 49
37.2 32.9 29.9
299 163 127
50.8 27.7 21.6
55 31 116 72 45 45 42
33.3 18.8 70.3 43.6 27.3 27.3 25.5
216 48 341 266 204 91 164
36.6 8.1 57.8 45.1 34.6 15.4 27.8
44 121
26.7 73.3
91 499
15.4 84.6
P
Yes (n = 64)
No (n = 691)
n
%
n
%
33 31 49 21 6 61 33
51.6 48.4 76.6 32.8 9.4 95.3 51.6
270 421 523 354 27 557 180
39.1 60.9 75.7 51.2 3.9 80.6 26.3
55 9
85.9 14.1
636 55
92.0 8.0
4 3 47 2 1
6.3 4.7 73.4 3.1 1.6
18 16 456 63 66
2.6 2.3 66.0 9.1 9.6
20 18 25
31.7 28.6 39.7
340 199 151
49.3 28.8 21.9
23 5 50 24 24 7 14
35.9 7.8 78.1 37.5 37.5 10.9 21.9
248 74 407 314 225 129 192
35.9 10.7 58.9 45.4 32.6 18.7 27.8
4 60
6.3 93.8
131 560
19.0 81.0
.749
subject, a disturbance call, the number of deputies at the scene, and the primary reason for UOF was to defend self or others were no longer significant and were removed from the model. After adjusting for covariates, the variables with the largest associations with deputy injury were when the subject had a weapon (OR, 4.15; 95% CI, 1.53-11.23) and physical resistance by the subject (OR, 4.15; 95% CI, 1.24-13.94). As with subject injury, incidents where the subject was not suspected of substance use were also associated with deputy injury (OR, 2.39; 95% CI, 1.35-4.21). The number of deputies at the scene and the type of force used by the deputy were no longer significant in the model and were removed.
5. Discussion This study reports factors associated with suspect and deputy injuries during UOF events. The range of force used
.001 .002 .441 .012 b.001 .011
.440 .582 .015 .117 .259 .007
.438 b.001 .004 .741 .078 b.001 .550 .001
P .051
.876 .005 .041 0.002 b.001 .094
.108 .246 .227 .157 .035 .003
.994 .668 .003 .222 .421 .124 .310 .010
by a deputy to gain control of a suspect can range from law enforcement physical presence to displaying or using a weapon. This study focused on situations where only physical force was used during an incident. Of the 755 uses of physical force over the 6-month period that we analyzed, a few factors were identified that were independently associated with both subject and deputy injuries and included if a person was not suspected of substance use, if the subjects actively resisted, and if additional force was required because initial UOF did not work. A male subject, when force is used to prevent a violent felony, a high-risk situation that may cause someone injury or death, and when a less lethal method was used were also independently associated with injury among subjects as was when the suspect had a weapon and to defend self or other for deputies. These results suggest that the odds of injury among subjects and deputies increases the longer a situation is out of control and provides further justification for policies that
530
E.M. Castillo et al.
Table 4 Associations of select factors with risk of injury among subjects and deputies during 755 physical law enforcement UOF events, January 1 to June 30, 2009 Subject injured Characteristic Sex (male/female) Weapon (yes/no) Possible substance use (no/yes) Physical resistance by subject (yes/no) Additional force required (yes/no) Prevent a violent felony (yes/no) Defend self or other (yes/no) High-risk situation (yes/no) Type of force (less lethal/other) a
OR 1.74
Deputy injury
95% CI (lower, higher) 1.06, 2.85
P .028
a
1.96 2.65 2.05 2.15
2.86 4.74 3.03 3.71
b.001 .001 b.001 .006
1.05, 2.67 1.25, 3.30
.031 .004
1.34, 1.49, 1.39, 1.24,
a
1.67 2.03
OR
95% CI (lower, higher)
P
1.53, 1.35, 1.24, 1.63,
.005 .003 .021 b.001
a
4.15 2.39 4.15 2.81
11.23 4.21 13.94 4.84
a
1.97
1.05, 3.70
.036
a a
Not included in the model because the characteristics were not significant in the univariate comparisons or no longer significant in regression model.
support the use of less lethal, yet effective, tools to gain control of suspects in a timely manner. These results also provide more information regarding the type of calls where injuries may be more likely to occur, such as with disturbances. However, although it would be intuitive to think potential alcohol intoxication or substance use may lead to situations where injuries are more common, it does not appear to be associated with subject or deputy injury during UOF events in this population and should be studied further. Although this study focused on all types of nonfatal injuries, a recent study by Ho et al [17] suggests a potential increase in the risk of sudden death during incidents of physical resistance or fleeing. In their study, the authors conclude that simulations of physical resistance had the greatest changes in acidosis and catecholamine markers, which could be potentially contributing or causal mechanisms in sudden arrest-related deaths. There are only a few recently published studies that have investigated the association between law enforcement UOF events and injury in the field [10-16], and most of them focus on the use of CEDs [10,12-15]. These less lethal UOF weapons have become widely used by many law enforcement agencies throughout the world and have been shown to cause minimal injury to subjects [13-15] and decrease injury among deputies [10,14]. Although the current study did not focus specifically on these devices, they are currently used by the San Diego County Sheriff's Department and were the most often used less lethal weapon system used by the department, comprising over 70% of this type of force. However, although the use of less lethal weapons was associated with subject injury, it was not possible to determine the cause or extent of injury from the current data. Department policy requiring all CED activations to be assessed by a physician could overinflate the odds of injury from this method. In addition, the use of these methods did not show to be protective against deputy injury. This was likely because of the relatively small number of uses and deputy injuries and not the lack of benefit of these systems. A study by Smith et al [11] investigated the impact of various UOF methods and resistance on officer and suspect
injuries. In this study, investigators obtained data on 1645 UOF incidents from 2 different law enforcement agencies and modeled the odds of injury and severity of injury. They reported that CED was associated with decreased odds of suspect injury in 1 agency, whereas the use of OC spray was associated with decreased odds of suspect injury in the other. However, the use of hands on tactics to control a suspect was associated with increased odds of injury among both the suspect and deputy. Although the study provides valuable information regarding the odds of injury by UOF method, they did not control for the situation characteristics that potentially contribute to injury such as type of resistance or type of call. The most comprehensive study investigating the law enforcement UOF safety and injury was done by Strote et al [16]. This was a retrospective study investigating UOF events by the Seattle Police Department among 888 subjects. In this study, they reported 187 injured subjects (21.1%), which was similar to the injury rate (20.5%) found in the current study. Also in both studies, confirmed or suspected substance use and mental impairment were relatively common. However, the study by Strote et al defined injury as being seen in an emergency department within 24 hours of the incident. In our study, we defined injury as deputy indicated regardless of treatment and included those who were treated for their injuries and those who were not. This study has limitations that must be considered when interpreting the results. As a retrospective data review, data elements were not prospectively collected, and missing or incomplete data may introduce bias into the results. However, the UOF report form completed by deputies includes standard data elements for these types of events and is required for each incident immediately after the incident. Data could also be misinterpreted or miscoded during the abstraction process. To limit this, data were abstracted by a trained researcher and completed on site with San Diego County Sheriff's Department personnel available for assistance and interpretation when needed. The results may not be generalizable to other populations, as data were abstracted from only one law enforcement agency that spans a wide
Use of force injuries enforcement area that includes rural and urban locations. In addition, the records used for this study were all based on deputy reports and specific data elements, including the true cause or extent of an injury could not be confirmed. Although UOF events are recorded on standardized report forms, confirmation of specific factors such as suspected substance use could not be confirmed, or interpretation of what is defined as an injury could not be determined. The result of sampling bias due to deputy assignments to higher call areas with an increased potential for UOF incidents could also not be controlled for or assessed. In conclusion, this study identifies situational characteristics potentially associated with subject and deputy injuries during UOF events. Although causality cannot be demonstrated in this study and temporal factors cannot be assessed, these results suggest that quick control of suspects may decrease suspect and deputy injuries during these situations. However, more studies on the frequency of law enforcement UOF and medical outcomes are needed to improve our understanding of these situations and related injuries.
Acknowledgments The authors thank the San Diego County Sheriff's Department for their participation in this project.
References [1] Chan TC, Vilke GM, Neuman T, Clausen J, Schmidt P, Snowden T, et al. Does oleoresin capsicum “pepper spray” exposure impact cardiovascular function in human subjects? Acad Emerg Med 2001;8(5):442. [2] Ho JD, Miner JR, Lakireddy DR, et al. Cardiovascular and physiologic effects of conducted electrical weapon discharge in resting adults. Acad Emerg Med 2006;13:589-95. [3] Ho JD, Dawes DM, Bultman LL, Thacker JL, Skinner LD, Bahr JM, et al. Respiratory effect of prolonged electrical weapon application on human volunteers. Acad Emerg Med 2007;14:197-201.
531 [4] Vilke GM, Sloane CM, Bouton KD, Kolkhorst FW, Levine SD, Neuman TS, et al. Physiological effects of a conducted electrical weapon on human subjects. Ann Emerg Med 2007;50(5):569-75. [5] Vilke GM, Sloane C, Levine S, Neuman T, Castillo E, Chan TC. Twelve-lead electrocardiogram monitoring of subjects before and after voluntary exposure to the Taser X26. Am J Emerg Med 2008;26(1): 1-4. [6] Sloane CM, Chan TC, Levine SD, Dunford JV, Neuman T, Vilke GM, et al. Measurement of subjects exposed to the Taser X-26(R). J Emerg Med 2008;35(1):29-32. [7] Dawes DM, Ho JD, Johnson MA, Lundin E, Janchar TA, Miner JR. 15-second conducted electrical weapon exposure does not cause core body temperature elevation in non-environmentally stressed resting adults. Forensic Sci Int 2008;176:253-7. [8] Ho JD, Dawes DM, Bultman LL, Moscati RM, Janchar TA, Miner JR. Prolonged TASER use on exhausted humans does not worsen markers of acidosis. Am J Emerg Med 2009;27(4):413-8. [9] Vilke GM, Sloane CM, Suffecool A, Kolkhorst FW, Neuman TS, Castillo EM, et al. Physiologic effects of the TASER after exercise. Acad Emerg Med 2009;16(8):704-10. [10] Jenkinson E, Neeson C, Bleetman A. The relative risk of police use-offorce options: evaluating the potential for deployment of electronic weaponry. J Clin Forensic Med 2006;13(5):229-41. [11] Smith MR, Kaminski RJ, Rojek J, Alpert JP, Mathis J. The impact of conducted energy devices and other types of force and resistance on officer and suspect injuries. Policing 2007;30(3):423-46. [12] Vilke GM, Johnson 3rd WD, Castillo EM, Sloane C, Chan TC. Tactical and subject considerations of in-custody deaths proximal to use of conductive energy devices. Am J Forensic Med Pathol 2009;30(1):23-5. [13] Bozeman WP, Hauda 2nd WE, Heck JJ, Graham Jr DD, Martin BP, Winslow JE. Safety and injury profile of conducted electrical weapons used by law enforcement officers against criminal suspects. Ann Emerg Med 2009;53(4):480-9. [14] MacDonald JM, Kaminski RJ, Smith MR. The effect of less-lethal weapons on injuries in police use-of-force events. Am J Public Health 2009;99(12):2268-74. [15] Strote J, Walsh M, Angelidis M, Basta A, Hutson HR. Conducted electrical weapon use by law enforcement: an evaluation of safety and injury. J Trauma 2010;68(5):1239-46. [16] Strote J, Verzemnieks E, Walsh M, Hutson HR. Use of force by law enforcement: an evaluation of safety and injury. J Trauma 2010;69(5): 1288-93. [17] Ho JD, Dawes DM, Nelson RS, Lundin EJ, Ryan FJ, Overton KG, et al. Acidosis and catecholamine evaluation following simulated law enforcement “use of force” encounters. Acad Emerg Med 2010;17(7): e60-8.