Violent behavior by emergency department patients with an involuntary hold status
a b s t r a c t
Background: Violence against health care workers has been increasing. Health care workers in emergency depart- ments (EDs) are highly vulnerable because they provide care for patients who may have Mental illness, behav- ioral problems, or Substance use disorders (alone or in combination) and who are often evaluated during an involuntary hold. Our objective was to identify factors that may be associated with Violent behavior in ED pa- tients during involuntary holds.
Methods: Retrospective review of patients evaluated during an involuntary hold at a suburban acute care hospital ED from January 2014 through November 2015.
Results: Of 251 patients, 22 (9%) had violent incidents in the ED. Violent patients were more likely to have a urine drug screen positive for triCyclic antidepressants (18.2% vs 4.8%, P = 0.03) and to present with substance misuse (68.2% vs 39.7%, P = 0.01), specifically with marijuana (22.7% vs 9.6%, P = 0.06) and alcohol (54.5% vs 24.9%, P = 0.003). ED readmission rates were higher for violent patients (18.2% vs 3.9%, P = 0.02). No significant difference was found between violent patients and nonviolent patients for sex, race, marital status, insurance status, med- ical or psychiatric condition, reason for involuntary hold, or length of stay.
Conclusion: Violent behavior by patients evaluated during an involuntary hold in a suburban acute care hospital ED was associated with tricyclic antidepressant use, substance misuse, and higher ED readmission rates.
(C) 2017
Introduction
Violence against health care workers in hospitals and health care fa- cilities increased from 2.0 events per 100 beds in 2012 to 2.8 events per 100 beds in 2015 [1]. In 2013, approximately 25,000 workplace assaults occurred, and 70% of them were in health care settings [2]. Understand- ing the factors that contribute to the risk of violence by patients is crit- ical for preventing these incidents.
Certain patient and location characteristics are known to be related to a higher incidence of violence. Patients with mental illness, dementia, or substance use disorder have the highest prevalence of violence [3,4],
Abbreviations: ED, emergency department; REDCap, Research Electronic Data Capture.
? Portions of this manuscript have been submitted for publication by Roy A, Lachner C, Dawson NL, Vadeboncoeur TF, Bosworth VA, Rummans TA, et al. Characteristics and out- comes of patients treated on involuntary hold status in the emergency department. Used with permission.
?? Conflict of interest: None.
* Corresponding author at: Division of Hospital Internal Medicine, Mayo Clinic, 4500 San Pablo Rd, Jacksonville, FL 32224, United States.
E-mail address: dawson.nancy11@mayo.edu (N.L. Dawson).
and emergency departments (EDs) and psychiatric facilities are the lo- cations with the greatest risk of violence to health care workers [4]. Studies have shown that up to 100% of ED nurses have experienced some type of workplace violence [4-6].
Patients with mental illness and substance misuse are often seen in the ED for medical evaluations. These patients can be at risk of harm to themselves or others. All states and the District of Columbia have laws that enable a mental health professional or health care provider acting in good faith to revoke a person’s civil liberties when it is believed that the person is at imminent risk to self or others. The Florida Mental Health Act of 1971, commonly known as the Baker Act [7], allows a 72- hour emergency hold, which can be initiated by law enforcement or physicians, for a mental illness patient posing a threat to self or others. Patients with an involuntary hold status are frequently seen in the ED for medical evaluation and can be challenging because they may be un- cooperative and agitated and put health care workers at risk for injury. Most of what is known about the demographic and clinical characteris- tics of patients with an involuntary hold status comes from studies of patients admitted to inpatient psychiatric treatment facilities [8]. Simi- larly, most studies delineating the characteristics of violent patients
http://dx.doi.org/10.1016/j.ajem.2017.08.039
0735-6757/(C) 2017
N.L. Dawson et al. / American Journal of Emergency Medicine 36 (2018) 392–395 393
also involved patients in psychiatric facilities, with the majority of these patients having psychotic disorders, substance misuse history, or a pre- vious history of violence [3,9-11].
The reason that the incidence of violence is higher in the ED than in the rest of the hospital has not been completely explained. Several the- ories have been suggested, including the fact that ED patients are more likely to have both active medical and psychiatric problems. Many stud- ies of violent patients are from psychiatric hospitals or are older studies [3,10-12]; characteristics at other types of hospitals and underlying mo- tivations for violence may have changed over time. More information about patients who commit violent acts during involuntary holds in the ED would be helpful for recognizing and preventing aggressive be- haviors. Therefore, to characterize these patients, we reviewed the re- cords of patients with reports of violence who were seen involuntarily in the ED of an acute care hospital in a suburban location and compared them to patients during involuntary holds without violent acts reported.
Methods
The study was conducted at an academic tertiary care center in Flor- ida. Patients in this study were adults who were involuntarily evaluated in the ED from January 1, 2014, through November 30, 2015. These pa- tients presented to the ED with a mental illness and were deemed a risk of harm to self or others thus were placed on a 72 h involuntary hold. This hold was placed either by law enforcement prior to presentation or by ED providers upon evaluation. If the involuntary hold is placed outside of the hospital, law enforcement officials bring the patient to the ED for a medical evaluation prior to transferring the patient to an in- patient psychiatric facility. Patients who are placed on the hold during the ED stay also receive a complete medical evaluation. These patients are continuously monitored by nursing to prevent harm; patients who are deemed a risk of harm to others may also have law enforcement per- sonnel continuously present. If the complete medical evaluation by the ED provider finds no Medical issues requiring hospitalization, then a re- quest is placed for transfer to an inpatient psychiatric facility in the com- munity, as this study institution does not have an inpatient psychiatric unit. The patients remain in the ED until that transfer is completed. If the patient on hold is found to have a medical issue requiring hospital- ization, they are admitted to the hospital.
If a patient had multiple admissions during the study period, only the first admission was included. Collected demographic data were age, sex, race, religion, marital status, housing, payer source, and tobac- co use. All charts were reviewed and clinical data collected were ED length of stay, reason for involuntary hold, psychiatric disorder, suicide attempt, substance use disorder, serum alcohol level, urine drug screen results, medical disorder, violence in the ED, 30-day ED readmission, and 30-day mortality. Violence included all verbal or physical alterca- tions that were documented in nursing notes, triage notes or provider notes. Any mention of any violent act or angry or abusive language to- ward hospital or law enforcement personnel was considered violent acts. In addition, institutional incident reports were reviewed for the study time period and recorded if they occurred with any of the invol- untary hold patient encounters. All data were retrieved from the elec- tronic health record.
Study data were collected and managed with Research Electronic Data Capture (REDCap), a secure, web-based application designed to support data capture for research studies and hosted at Mayo Clinic. Standard descriptive statistics were used for patient demographic and clinical characteristics. To compare the violent and nonviolent patients, categorical variables were evaluated with the Pearson ?2 test and the Fisher exact test, and continuous variables were evaluated with the Wilcoxon rank sum test. A P value of 0.05 or less was considered statis- tically significant. All statistical analyses were performed with SAS ver- sion 9.3 software (SAS Institute Inc., Cary, North Carolina).
The study was approved by the Mayo Clinic Institutional Review Board.
Results
A total of number of 278 patient encounters with involuntary holds occurred. Out of these, 11 evaluations were excluded as they were readmissions (during the study period) and 16 patients were less than 18 years old leaving a total study population of 251 patient encounters reviewed. Of the 251 patients, 22 (9%) were reported as having violent or aggressive incidents during their stay in the ED. Violent patients and nonviolent patients were not significantly different in sex, race, marital status, or insurance status (Table 1). Violent patients had a longer ED length of stay than nonviolent patients (median, 376 min vs 655.5 min; P = 0.01). Violent patients had more suicide attempts, but the difference was not statistically significant (54 [23.6%] vs 9 [40.9%]; P = 0.07). The reasons for the involuntary hold and the underlying psy- chiatric illnesses were not significantly different between the 2 groups. Table 2 delineates the types of violent events reported by ED personnel: Most ED personnel reported combative behavior, and security person- nel were involved in many of the incidents.
The violent patients were more likely to present with current sub-
stance misuse (15 [68.2%] vs 91 [39.7%]; P = 0.01), with misuse of mar- ijuana (5 [22.7%] vs 22 [9.6%]; P = 0.06) and alcohol (12 [54.5%] vs 57 [24.9%]; P = 0.003) being more common. The increased frequency of marijuana use in these patients was confirmed by its more frequent presence in the urine, and the violent patients were also more likely to have tricyclics present during the urine drug screening (4 [18.2%] vs 11 [4.8%]; P = 0.03).
The 2 groups did not differ in the presence or absence of a concom- itant medical diagnosis nor in the type of underlying medical condition. Virtually all the patients in both groups were discharged to an inpatient psychiatric facility. Only 8 patients had the involuntary hold removed while in the ED, and all those patients were in the nonviolent group. The violent patients were more likely to be readmitted to the ED than those without an aggressive event (4 [18.2%] vs 9 [3.9%]; P = 0.02). Ad- mission to the hospital, disposition after admission and frequency of Hospital readmissions were not significantly different between the 2 pa- tient groups.
Discussion
Among patients requiring involuntary holds to prevent harm to self or others, 9% committed a violent act while in the ED. Other studies have shown a higher prevalence among Psychiatric patients [3,12,13]; how- ever, those studies were done in psychiatric hospitals, and the data are scant for patients presenting to nonpsychiatric facilities. In addition, most of those studies relied on retrospective surveys, which can be sub- ject to recall bias [14,15]. Our data were taken directly from documenta- tion in the electronic health records that was completed immediately after the incident.
Alcohol and marijuana misuse were more common among the vio- lent patients. This finding is consistent with results from previous stud- ies that found that substance use disorder increases the risk of aggression [9,11,12,16]; however, we did not find that use of other drugs of abuse (e.g., opioids, sedatives, hypnotics, or stimulants) was more prevalent among the violent patients, although the number of vi- olent patients who used those drugs was small. We also did not find that any specific underlying psychiatric disorder was more common in the violent patients, although other studies have reported that patients with schizophrenia had a higher risk for aggression [9,10,12]. It is possi- ble that schizophrenic patients are less frequently placed under invol- untary holds than patients with suicide attempts, which were more common in our patient population.
Length of stay in the ED was longer for patients exhibiting violence. This could be explained by the fact that violent acts can delay care, in- cluding drawing blood samples and checking vital signs. In fact, several of the narrative accounts in our study reported difficulties with com- pleting adequate evaluations. Alternatively, increased total time in the
394 N.L. Dawson et al. / American Journal of Emergency Medicine 36 (2018) 392–395
Table 1
Demographics.
Covariate No Yes P-value |
|||||||
Covariate |
No |
Yes |
P-value |
N = 229 |
N= 22 |
Table 1 (continued)
Patient characteristics
N = 229
N= 22
Medical information
113 (49.3%) |
11 (50.0%) |
0.953b |
16 (7.0%) |
0 (0.0%) |
0.374c |
Patient has medical issues
Age (at time of involuntary hold) 0.866a Diabetes mellitus
Median (Q1, Q3) |
42 (29, 54) |
41 (30, 57) |
Hypertension |
36 (15.7%) |
6 (27.3%) |
0.166b |
|
Range |
18, 94 |
18, 71 |
Liver disease |
3 (1.3%) |
0 (0.0%) |
1.000c |
|
Gender |
0.810b |
Coronary artery disease |
7 (3.1%) |
0 (0.0%) |
1.000c |
||
Female |
131 (57.2%) |
12 (54.5%) |
COPD |
4 (1.7%) |
1 (4.5%) |
0.370c |
|
Male |
98 (42.8%) |
10 (45.5%) |
Other |
101 (44.1%) |
11 (50.0%) |
0.595b |
Race
White
Non-White
Religion
None
0.536c
All others |
143 (62.4%) |
14 (63.6%) |
Involuntary hold removed in the ED |
8 (3.5%) |
0 (0.0%) |
1.000c |
|
Marital status |
0.801b |
ED readmission within 30 days |
9 (3.9%) |
4 (18.2%) |
0.019c |
||
Married |
67 (29.3%) |
7 (31.8%) |
ED readmission reason |
N = 9 |
N = 4 |
0.217c |
Not married
162 (70.7%) |
15 (68.2%) |
Related to index admission |
7 (77.8%) |
1 (25.0%) |
|
0.085c |
Unrelated to index ED admission |
2 (22.2%) |
3 (75.0%) |
185 (80.8%) |
17 (77.3%) |
0.691b |
Discharge information |
N = 196 |
N = 18 |
44 (19.2%) |
5 (22.7%) |
Home |
7 (3.6%) |
0 (0.0%) |
|
0.912b |
Inpatient psychiatric/treatment facility |
189 (82.5%) |
18 (100.0%) |
||
86 (37.6%) |
8 (36.4%) |
0 (0.0%) |
Housing
Missing |
9 |
2 Hospital admission information |
|||||
Own/rent house/apartment |
189 (85.9%) |
20 (100.0%) |
Patient was admitted to the hospital |
33 (14.4%) |
4 (18.2%) |
0.544c |
|
Homeless/shelter/institution/living in |
31 (14.1%) |
0 (0.0%) |
Reason for hospital admission |
N = 33 |
N = 4 |
1.000c |
|
another’s home |
Not medically cleared |
24 (72.7%) |
3 (75.0%) |
||||
Payer |
0.929b |
No inpatient psychiatric bed available |
9 (27.3%) |
1 (25.0%) |
|||
Self-pay |
75 (32.8%) |
7 (31.8%) |
Patient was violent in the hospital after |
– |
|||
All others |
5 (9.3%) |
0 (0.0%) |
admission |
Tobacco abuse 109 (47.6%) 8 (36.4%) 0.313b No 33 (100.0%) 4 (100.0%)
ED stay information
Involuntary hold was removed during
6 (18.2%) 1 (25.0%) 1.000c
Emergency department length of stay (hours) |
0.015a |
hospitalization Hospital discharge disposition |
N = 33 |
N = 4 |
1.000c |
||
Median (Q1, Q3) |
6 (4, 10) |
11 (5, 17) |
Home |
6 (18.2%) |
1 (25.0%) |
||
Range |
1, 49 |
1, 30 |
SNF/long term care/LTAC |
2 (6.1%) |
0 (0.0%) |
||
Reason for involuntary hold |
Inpatient psychiatry |
24 (72.7%) |
3 (75.0%) |
evaluation Psychiatry information
Suicidal |
166 (72.5%) |
19 (86.4%) |
0.208c |
Homicidal |
12 (5.2%) |
1 (4.5%) |
1.000c |
Danger to self (not suicidal) |
68 (29.7%) |
7 (31.8%) |
0.503b |
Is unable to understand need for |
6 (2.6%) |
1 (4.5%) |
0.478c |
Current psychiatric disorder Anxiety
131 (57.2%) |
12 (54.5%) |
0.490b |
|
38 (16.6%) |
3 (13.6%) |
0.501c |
|
Schizophrenia |
19 (7.9%) |
1 (4.5%) |
0.458c |
Personality disorder (i.e., borderline, |
7 (3.1%) |
1 (4.5%) |
0.525c |
Homeless shelter
Hospital readmission within 30 days
1 (3.0%) |
0 (0.0%) |
|
– |
||
33 (100.0%) |
4 (100.0%) |
|
1 (3.2%) |
0 (0.0%) |
1.000c |
N = 33 |
N = 4 |
0.309a |
No
Patient died within 30 days of discharge from hospital
Hospitalization length of stay (days)
196 (85.6%) |
19 (86.4%) |
0.250c |
Median (Q1, Q3) |
3 (1, 4) |
1.5 (1.0, 2.5) |
54 (23.6%) |
6 (27.3%) |
0.698b |
Range |
0, 38 |
1, 3 |
Substance abuse (does not include tobacco)
None |
1 (0.4%) |
0 (0.0%) |
1.000c |
Suicide attempt (not just ideation) |
54 (23.6%) |
9 (40.9%) |
0.073b |
Drug information
Current substance abuse (other than tobacco)
Opiates/narcotics |
27 (11.8%) |
2 (9.1%) |
1.000c |
[17-21]. Completing these tasks may be more challenging with violent |
Sedative/hypnotics |
15 (6.6%) |
0 (0.0%) |
0.375c |
patients under involuntary holds because of the immediate need for |
43 (18.8%) 6 (27.3%) 0.337b
93 (40.6%) 15 (68.2%) 0.009b
Wilcoxon.
b Chi-square.
c Fisher exact.
ED could give more opportunity for patients with a propensity for vio- lence to incite an event. Readmission to the ED was also more prevalent among the violent patients. Several factors have been found to lead to readmission, including failed provider handoffs and delayed follow-up
Stimulants 15 (6.6%) 1 (4.5%) 1.000c
Marijuana 22 (9.6%) 5 (22.7%) 0.058b
Alcohol 58 (24.9%) 12 (54.5%) 0.005b
Hallucinogens 5 (2.2%) 0 (0.0%) 1.000c
Other 5 (2.2%) 1 (4.5%) 0.427c
Serum alcohol level obtained? Blood alcohol concentration (mg/dL) N Median (Q1, Q3) |
162 (70.7%) 162 0.0 (0.0, |
19 (86.4%) 19 161.5 (0.0, |
0.141c mediately after the event. This could have caused an underestimation of 0.022a the violent acts. In addition, there was no standard protocol for documenting these events, so this documentation could be incomplete |
165.8) |
289.9) |
or inconsistent. There were a relatively small number of violent acts in |
|
Range |
this sample of patients, so some results may not have achieved signifi- |
safety and urgent placement in a psychiatric facility.
Although this study highlights important issues facing ED staff, it has limitations because of its retrospective nature.. With the retrospective design, some milder violent events may not have been documented im-
cance. A larger sample size could have different results. Finally, the study was conducted in 1 suburban hospital, and results may not be transferrable to all facilities.
Urine drug screen |
180 (78.6%) |
18 (81.8%) |
1.000c |
Amphetamine |
11 (4.8%) |
0 (0.0%) |
0.606c |
Barbiturate |
8 (3.5%) |
0 (0.0%) |
1.000c |
Benzodiazepine |
49 (21.4%) |
6 (27.3%) |
0.525b |
Cocaine |
18 (7.9%) |
2 (9.1%) |
0.690c |
Opiate |
34 (14.8%) |
5 (22.7%) |
0.330b |
PCP |
1 (0.4%) |
0 (0.0%) |
1.000c |
THC |
27 (11.8%) |
8 (36.4%) |
0.001b |
Tricyclic antidepressant |
11 (4.8%) |
4 (18.2%) |
0.032c |
None |
79 (34.5%) |
4 (18.2%) |
0.156c |
In conclusion, we found that in an ED, 9% of patients with invol-
untary holds exhibited violent behavior. Most of these patients were affected by misuse of substances, including alcohol and tetra- hydrocannabinol. Understanding the factors that place health care workers at risk of violence is critical for preventing injury. Since many of these types of patients initially present to community
N.L. Dawson et al. / American Journal of Emergency Medicine 36 (2018) 392–395 395
Table 2
Descriptions of violent episodes reported by ED personnel.
Description of harm episode |
Frequency |
% |
Cumulative frequency |
Cumulative % |
Patient verbally abusive and combative with staff |
7 |
31.8 |
7 |
31.8 |
Patient was agitated, tried to pull out IV, and had to be restrained for safety |
3 |
13.6 |
10 |
45.4 |
Patient combative and attempted to leave the ED |
3 |
13.6 |
13 |
59.1 |
Patient physically violent toward the staff and had to be restrained |
3 |
13.6 |
16 |
72.7 |
Patient escaped campus alone or with family and had to be returned by law enforcement |
2 |
9.1 |
18 |
81.8 |
Patient showed mental stress and anxiety along with agitation |
2 |
9.1 |
20 |
90.9 |
Patient was suicidal and homicidal |
1 |
4.5 |
21 |
95.4 |
Patient was hallucinating and combative and had to be restrained |
1 |
4.5 |
22 |
100.0 |
Abbreviations: ED, emergency department; IV, intravenous catheter.
hospitals, our analysis of the characteristics of patient violence in such a hospital adds to this understanding. Further research into pro- cesses that can prevent violence when a patient with known risk fac- tors presents to the health care facility could improve the safety of both patients and providers. Consequently, development of proto- cols and interventions to manage violent and often intoxicated pa- tients is essential in the ED.
References
- The Joint Commission. Sentinel event data: general information 2Q 2016 update. Oakbrook Terrace (IL): The Joint Commission; 2016.
- Occupational Safety and Health Administration (OSHA). Guidelines for preventing workplace violence for healthcare and social services workers [Internet]. Washington (DC): U.S: Department of Labor and Occupational Safety and Health Administration; 2016, OSHA 3148-06R 2016 [cited 2017 Jan 28] Available from: https://www.osha. gov/Publications/osha3148.pdf.
- Broderick C, Azizian A, Kornbluh R, Warburton K. Prevalence of Physical violence in a forensic psychiatric hospital system during 2011-2013: patient assaults, staff as- saults, and repeatedly violent patients. CNS Spectr Jun 2015;20(3):319-30.
- Pompeii L, Dement J, Schoenfisch A, Lavery A, Souder M, Smith C, et al. Perpetrator, worker and workplace characteristics associated with patient and visitor perpetrat- ed violence (Type II) on hospital workers: a review of the literature and existing oc- cupational injury data. J Saf Res 2013 Feb;44:57-64.
- May DD, Grubbs LM. The extent, nature, and precipitating factors of nurse assault among three groups of registered nurses in a regional medical center. J Emerg Nurs 2002 Feb;28(1):11-7.
- Gerberich SG, Church TR, McGovern PM, Hansen HE, Nachreiner NM, Geisser MS, et al. An Epidemiological study of the magnitude and consequences of work related violence: the Minnesota Nurses’ Study. Occup Environ Med 2004 Jun;61(6): 495-503.
- State of Florida Department of Children and Families Mental Health Program Office. History of the Baker ACT: it’s development and intent [Internet]. Florida: Baker Act Website; May 2002[cited 2017 Mar 21]. Available from: https://www.dcf.state.fl. us/programs/samh/MentalHealth/laws/histba.pdf.
- Kelly BD, Emechebe A, Anamdi C, Duffy R, Murphy N, Rock C. Custody, care and country of origin: demographic and diagnostic admission statistics at an inner-city adult psychiatry unit. Int J Law Psychiatry Jan-Feb 2015;38:1-7.
- Djokic-Pjescic K, Ristic-Dimitrijevic R, Nenadovic M, Jovicic S, Nesic S, Rajcic T, et al. Hetero-aggressive outbursts in patients’ behaviour in Emergency psychiatry. Srp Arh Celok Lek Dec 2011;139(Suppl. 1):57-60.
- McNiel DE, Binder RL, Greenfield TK. Predictors of violence in civilly committed acute psychiatric patients. Am J Psychiatry Aug 1988;145(8):965-70.
- Ochitill HN, Krieger M. Violent behavior among hospitalized medical and surgical patients. South Med J Feb 1982;75(2):151-5.
- Iozzino L, Ferrari C, Large M, Nielssen O, de Girolamo G. prevalence and risk factors of violence by psychiatric acute inpatients: a systematic review and meta-analysis. PLoS One Jun 10 2015;10(6):e0128536.
- Arango C, Calcedo Barba A, Gonzalez-Salvador T, Calcedo Ordonez A. Violence in in- patients with schizophrenia: a prospective study. Schizophr Bull 1999;25(3): 493-503.
- Phillips JP. Workplace violence against health care workers in the United States. N Engl J Med Apr 28 2016;374(17):1661-9.
- Lau JB, Magarey J. Review of research methods used to investigate violence in the emergency department. Accid Emerg Nurs Apr 2006;14(2):111-6.
- Eronen M, Angermeyer MC, Schulze B. The psychiatric epidemiology of violent be- haviour. Soc Psychiatry Psychiatr Epidemiol Dec 1998;33(Suppl. 1):S13-23.
- Roy CL, Poon EG, Karson AS, Ladak-Merchant Z, Johnson RE, Maviglia SM, et al. Pa- tient Safety concerns arising from test results that return after hospital discharge. Ann Intern Med Jul 19 2005;143(2):121-8.
- Ong MS, Magrabi F, Jones G, Coiera E. Last orders: follow-up of tests ordered on the day of hospital discharge. Arch Intern Med Sep 24 2012;172(17):1347-9.
- Moore C, McGinn T, Halm E. Tying up loose ends: discharging patients with unre- solved medical issues. Arch Intern Med Jun 25 2007;167(12):1305-11.
- Forster AJ, Murff HJ, Peterson JF, Gandhi TK, Bates DW. The incidence and severity of adverse events affecting patients after discharge from the hospital. Ann Intern Med Feb 4 2003;138(3):161-7.
- Kripalani S, Roumie CL, Dalal AK, Cawthon C, Businger A, Eden SK, et al. Effect of a pharmacist intervention on clinically important Medication errors after hospital dis- charge: a randomized trial. Ann Intern Med Jul 3 2012;157(1):1-10.