Characteristics of Aggressive Behavior among Male Inpatients with Schizophrenia.

BACKGROUND: The incidence of the aggressive behavior is higher among the patients with severe mental disorder such as schizophrenia than the general population. The study of factors related to aggressive behavior has great meaning in designing prevention and intervention methods with this population of patients. AIMS: To understand the characteristics of assaultive behavior of male patients with schizophrenia who have been hospitalized.
METHODS: Using a continuous sampling method, data from 75 male inpatients with a diagnosis of schizophrenia was collected at the psychiatric unit of Central South University Second Xiangya Hospital (Changsha, China) from August 2015 to February 2016.On the third day after hospitalization participants were given a general questionnaire as well as being assessed using the modified overt aggression scale (MOAS), historical clinical risk management-20 (HCR-20) questionnaire, hare psychopathic checklist-revised (PCL-R), and positive and negative syndrome scale (PANSS).Based on results of the MOAS participants were group into an 'aggressive behavior' group (39 cases) and 'non-aggressive behavior' group (36 cases). The differences in socio-demographic characteristics and scores on the other evaluation tools were then compared between these two groups.
RESULTS: Participants in the 'aggressive behavior' group had significantly different scores in the HCR-20 in the H1 (past violence events), H2 (violent events when young), H10 (disobedience in the past), and C4 (impulsiveness) sections; as well in the anti-social section of PCL-R; and significantly higher PANSS scores in the positive symptom, depressive symptoms and paranoid symptom sections than those in the 'non-aggressive behavior' group.
CONCLUSIONS: A combination of adverse and traumatic life events such as a history of violence, vulnerabilities in ones personality (e.g. impulsive or antisocial tendencies) and psychopathology of current illness (e.g. significant anxiety and depressive symptoms) contribute to aggressive behavior in male inpatients with schizophrenia. Our results contribute to the literature that will hopefully aid in ensuring patient and staff safety, as well as providing more information in working with this vulnerable population.

1. Introduction

Aggressive behavior in patients with psychological disorders has always been an important issue in clinical psychiatric work because of the high incidence rate and adverse consequences.[ Understanding factors related to aggressive behaviors in inpatients with mental disorders could aid in the evaluation, prevention, and early intervention in those cases where risk of aggressive behavior is high. There is a higher incidence of aggressive behaviors in those patients who have been diagnosed with schizophrenia [ and in addition, being male has been shown to be a risk factor for aggressive behavior in patients with mental disorders. [ Within this context, the current study seeks to better understand psychosocial and clinical factors related to aggressive behavior in hospitalized male patients with schizophrenia. The national mental health law of China provides for the involuntary hospitalization of those individuals with severe mental illness when there a high risk (or actual behaviors) of harm to self or others. Therefore, in addition to the tools measuring aggressive behaviors and tendencies, we also used this legal definition into account when considering the concept of “aggression”.

2. Participants and Methods

2.1 Study participants

Figure 1 shows the recruitment of study participants. This study used a continuous sampling method to select male inpatients with schizophrenia admitted to the psychiatric unit of the Second Xiangya Hospital of Central South University (Changsha, China) from August 2015 to February 2016. The inclusion criteria were the following: (a) meeting the diagnostic criteria for schizophrenia according to the tenth edition of the international classification of diseases (ICD-10); (b) aged 17 to 60; (c) had normal intelligence, able to communicate with the researchers; (d) informed consent was obtained from patients and their families. The exclusion criteria were the following: a) comorbidity with other mental disorders; b) comorbidity with severe somatic diseases or neurological diseases.

Figure 1.

Flow chart of the study

There were a total of 84 cases collected based on the inclusion and exclusion criteria. Written informed consent was obtained from all patients in this study and their family members, and this study was approved by the ethics committee of the Second Xiangya Hospital. After removing 9 incomplete cases the final number of cases was 75 (a participation rate of 89%). These 75 cases were divided into the “aggressive behavior” group (36 cases) and the “non-aggressive behavior” group (39 cases) according to their scores on each item of the modified overt aggression scales (MOAS). In the MOAS, a score of 0 on each of the items, a score of 0 in the total weighted score, or having a score of 1 or more on the ‘verbal threat’ section of the test were classified as being in the “non-aggressive behavior” group. Those participants receiving a score of 1 or more on test section related to aggression against possessions, self or others were considered as part of the “aggressive behavior” group.

2.2 Measurement tools

2.3.1 Self-administered general information questionnaire

a) Demographic information was collected including the following: age, marital status, education level, history of illness, frequency of past hospitalizations frequency; b) In addition, additional information about the family environment was collected including: early unhealthy family environment, family history of crime, situations involving alcohol or other substance abuse by parents, etc.

2.3.2 Historical clinical risk management-20 (HCR-20)

The HCR-20 contains 20 items: historical factors (10 items), clinical manifestations (5 items), and future risk management (5 items). These 3 aspects involve a comprehensive assessment on the risk of violence in patients with mental disorders. Numerous studies, in both China and elsewhere, demonstrate that the HCR-20 has good reliability and validity.[ This study took the scores in each item of HCR-20 as evaluation indicators (excluding H6 severe mental disorders and H7 mental illness). For each item on the scale there are three possible choices “not present”, “possibly present” and “definitely present”, represented as 0, 1, and 2, respectively.

2.3.3 Psychopathy checklist-revised (PCL-R)

The PCL-R is a scale used to evaluate mental illness as part of the HCR-20 evaluation. PCL-R has been revised by researchers in China, and its reliability and validity have been verified. [ This scale includes 20 items. Each section had a choice of ‘none’, ‘possible or not serious’, and ‘definite’ (i.e 3 selections, scored as 0, 1, and 2, respectively). The scale is divided into four factors including interpersonal factors, emotional factors, lifestyle factors, and antisocial factors, with items 11 and 17 not counted on any factor. [ Each factor score is the total score of the items contained in that section. In this study, we used the factor scores on the scale as evaluation indexes.

2.3.4 Modified Overt Aggression Scale (MOAS)

MOAS is a widely used assessment tool for violent behavior. Xie introduced the MOAS into China and tested the consistency of its raters.[ MOAS assesses frequency and intensity of aggressive behavior of individuals with mental disorders over the past week. The scale has 4 parts: verbal aggression, aggression against objects, aggression against oneself, and physical aggression against others. Each part of the scale is divided into five levels according to the severity of the aggressive behavior, (e.g. score of 0 equals ‘no aggression’ whereas a score of 4 equals the ‘maximum’ level of aggression). In this study, the MOAS scores were used as the basis for whether or not the participant was classified as being in the “aggressive-behavior” group.

2.3.5 Positive and Negative Syndrome Scale (PANSS)

The positive and negative syndrome scale (PANSS) is an assessment tool that is used to assess the presence of psychiatric symptoms in patients with schizophrenia and the severity of each symptom. PANSS is comprised of the positive symptom scale (7 items), negative symptom scale (7 items), general psychopathology scale, and supplemental items (3 items), in total 33 items. These 33 items can also be divided into groups according to the symptoms: (a) lack of response, (b) disturbed thinking, (c) activation syndrome, (d) paranoia, (e) depression, and (f) aggression. This study mainly used the scale score and the symptom score as assessment indicators. Each item on the scale ranges from 1 to 7 based on the frequency and severity of the symptom. The scores for the scale and symptoms were the total score of all included items in those sections.

2.3 Data Entry and Analysis

Data collection was completed by one researcher. All evaluations were completed within 3 days after the participant was hospitalized. Data for the history scale of the HCR-20 and the evaluation period measured by the MOAS and PANSS were the recent week before patients’ hospitalization. The scope of the remaining evaluations was any point during the patient’s life. Sources of evaluation data included patients, the patients’ family members, interviews with the presiding doctor and patients’ medical records.

2.4 Statistical analysis

SPSS version 19.0 statistical software was used to perform data analysis. The socio-demographic information and each assessment tool’s total score and the factor score were described statistically using the mean, standard deviation, constituent ratio, and frequency. t-test, chi-square test, and Mann-Whitney U test were used to test for statistical differences in socio-demographic information and each assessment tool’s total score and factor score between the “aggressive behavior” and “non-aggressive behavior” groups. Any factors of significant difference were put into logistic regression to confirm major risk factors after the use of univariate analysis.

3. Results

3.1 Comparison of general demographic data and unhealthy family environment factors between the aggressive and non-aggressive groups

As shown in table 1, there was no statistical difference between the groups in socio-demographic factors such as age, history of illness, marital status, and educational level. There was no significant difference in any item asking about unhealthy family environment factors between the two groups.

Table 1.

Comparison of socio-demographic and family environment data between the 2 groups

		Aggressive behavior group(n=39)	Non-aggressive behavior group(n=36)	statistics	p
Mean (sd) age		26.4 (6.0)	28.3 (7.2)	t=-1.19	0.234
Mean (sd) history of illness (months)		60(84)	36(54)	Z=-0.77	0.437
Marital status	Single	31	26	χ2=1.69	0.428
	Married/cohabitation	6	5
	Divorced	2	5
Educational level	Primary school	4	4	χ2=5.71	0.335
	Middle school	15	15
	High school	11	14
	College	5	1
	Undergraduate	4	1
	Master’s or above	0	1
Unhealthy family environment in childhood	Present	5	9	χ2=1.04	0.308
	Absent	31	30
Family history of criminal activities	Present	2	1	χ2=0.43	0.470
	Absent	34	38
Alcohol or other substance abuse in parents	Present	1	6	χ2=3.38	0.072
	Absent	34	33
Family history of mental illness	Present	8	5	χ2=-0.75	0.452
	Absent	31	31

3.2 Comparison of the HCR-20 scores between the “aggressive behavior” and “non-aggressive behavior” groups

As shown in table 2, the items that had a significantly higher score in the “aggressive behavior” group than the “non-aggressive behavior” group were: H1 (previous violent event) (mean(sd)=1.13(0.70) v. 0.67(0.76), Z=-2.69, p=0.007), H2 (violent event during childhood) (mean(sd)=1.31 (0.77) v. 0.78 (0.87), Z=-2.66, p=0.008), H10 (history of oppositional or antisocial behavior resulting in: incarceration, admittance to a labor reform program or juvenile reform program (mean(sd)=1.18 (0.51) v. 0.00 (0.00), Z =-2.20, p=0.027), C4 (impulsiveness) (mean(sd)=1.28 (0.79) v. 0.31 (0.58), Z=-4.99, p<0.001). There were significant differences between the two groups in the other items of this measure.

Table 2.

Comparison of mean (sd) HCR-20 total scores and factor scores between the 2 groups (mean [sd])

	Aggressive behavior group(n=39)	Non-aggressive behavior group(n=36)	Z	p
H1: previous violent event	1.13 (0.7)	0.67 (0.86)	-2.69	0.007
H2: violent event when young	1.31 (0.8)	0.78 (0.9)	-2.66	0.008
H3: instable interpersonal relationship	0.97 (0.8)	0.81 (0.7)	-0.85	0.392
H4: employment problem	0.85 (0.8)	0.81 (0.7)	0.17	0.864
H5: substance abuse problem	0.10 (0.4)	0.08 (0.4)	-0.35	0.722
H8: poor adaptation in early years (in the community school, family before 17 years old	0.85 (0.8)	0.53 (0.7)	-1.68	0.092
H9: personality disorder	0.15(0.4)	0.08(0.3)	-0.93	0.351
H10: disobedience in the past (detention, re-education through labor, juvenile reform, reform through labor)	0.18(0.5)	0.00(0.00)	-2.20	0.027
C1: no self insight	1.79(0.4)	1.89(0.4)	-1.40	0.159
C2: severe mental disorders, active symptoms, negative hostility (antisocial, hostility)	0.08(0.3)	0.03(0.2)	-0.94	0.347
C3: severe mental disorder active symptoms	1.87(0.3)	1.94(0.3)	-1.53	0.125
C4: impulsiveness	1.28(0.8)	0.31(0.6)	-4.99	<0.001
C5: no response to treatment (including biological, psychological, social, reform, and so forth)	0.95(0.5)	1.00(0.5)	-0.45	0.649
R1: lack of feasibility of treatment plans	0.90(0.8)	1.14(0.6)	-1.50	0.132
R2: promoting factors in the environment	0.77(1.0)	1.03(0.9)	-1.52	0.128
R3 lack of interpersonal support	0.62(0.8)	0.64(0.8)	-0.11	0.910
R4: disobedient to treatment and handling	0.97(0.9)	0.92(0.8)	0.27	0.787
R5: presence of stress factors	0.33(0.6)	0.33(0.6)	-0.04	0.967

3.3 Comparison of the PCL-R scores between the “aggressive behavior” and “non-aggressive behavior” groups

As shown in table 3, the factor with the highest score (as well as being statistically significant) was the antisocial factor (mean(sd)=1.46 (1.47) v. 0.53 (0.94), Z=-3.40, p=0.001). There was no significant difference between the 2 groups in the other factors on this measure.

Table 3.

Comparison of the PCL-R total scores and the factor scores between the 2 groups (mean [sd])

	Aggressive behavior group(n=39)	Non-aggressive behavior group(n=36)	Z	P
interpersonal factor	0.59(1.14)	0.53(1.18)	-0.36	0.718
coldness factor	2.97(2.55)	2.58(4.45)	-1.38	0.165
life style factor	3.85(2.52)	3.42(2.12)	-0.72	0.468
antisocial factor	1.46(1.47)	0.53(0.94)	-3.40	0.001

3.4 Comparison of the PANSS scores between the “aggressive behavior” and “non-aggressive behavior” groups

As shown in table 4, the factors with higher (as well as significantly different) scores in the “aggressive behavior” group were the following: activation syndrome (mean(sd)=4.90 (2.22) v. 4.08 (1.75)Z=-2.00, p=0.045), paranoia(mean(sd)=8.00(3.28) v. 6.53(2.89), t=-3.05, p=0.043), depression (mean(sd)=7.62(3.78) v. 4.08(1.75), Z=-2.49, p=0.013). There were no significant differences between the 2 groups in other factors on this measure.

Table 4.

Comparison of the PANSS scores between the 2 groups (mean [sd])

	Aggression group(n=39)	Non-aggression group(n=36)	statistics	p
Positive symptom scale	19.44(6.20)	17.42(4.60)	t=-1.59	0.116
Negative symptom scale	18.05(7.82)	19.42(7.68)	t=0.76	0.449
General psychopathology scale	35.10(7.31)	31.94(4.76)	Z=-1.85	0.064
Response deficit symptoms	5.41(2.37)	4.81(2.18)	t=0.24	0.807
Thought disorder syndrome	9.31(3.97)	10.33(3.44)	t=1.19	0.079
Activation syndrome	4.90(2.22)	4.08(1.75)	Z=-2.00	0.045
Paranoid syndrome	8.00(3.28)	6.53(2.89)	t=-3.05	0.043
Depressive syndrome	7.62(3.78)	4.08(1.75)	Z=-2.49	0.013
Aggressive syndrome	9.05(4.10)	5.72(2.50)	Z=-1.13	0.257

3.5 Stepwise regression analysis

Factors found to be significantly different between the two groups were further divided into psychosocial factors and clinical factors. The psychosocial factors included H1 (previous violent events), H2 (violent event during childhood), H10 (history of oppositional or antisocial behavior resulting in: incarceration, admittance to a labor reform program or juvenile reform program), and the antisocial factor on the PCL-R. Clinical factors included C4 (impulsiveness) on the HCR-20 and activation syndrome, paranoia, and depression on the PANSS. The stepwise logistic regression analysis (the p-values for entering the equation and exclusion were 0.05 and 0.01) was performed using “whether or not the participant was in the aggressive behavior group” as the dependent variable and the 4 psychosocial factors mentioned as the covariant. The final factor entered into the regression equation model was the antisocial factor (OR= 2.100, 95% CI: 1.26-3.50); using “whether or not the participant was in the aggressive behavior group” as the dependent variable and the 4 factors clinical factors mentioned as the covariant. The final factors entered from the HCR-20 and PANSS into the regression equation model were C4 (impulsiveness) (OR= 7.134, 95% CI: 2.96-17.21) and depression (OR= 1.291, 95% CI: 1.07-1.56). (See table 5)

Table 5.

Regression model coefficient and its OR value (95%CI)

	B	Exp (B)	p	Exp(B)95%
				Lower limit	Upper limit
historical factors:
antisocial factor	0.742	2.100	0.004	1.26	3.50
constant	-0.589	0.137	0.004
clinical factors:
c4: impulsiveness	1.965	7.134	<0.001	2.96	17.21
depressive syndrome	0.256	1.291	0.007	1.07	1.56
constant	-3.091	0.045	<0.001

4. Discussion

4.1 Main findings

The results of this study did not find that unhealthy family environment or socio-demographic factors such as age, marital status, educational level, and history of illness were related to the aggressive behavior male inpatients with schizophrenia.

Previous studies have shown that past violent behavior was the most important and stable risk factor in predicting hospitalization and aggressive behavior from individuals with various mental disorders in the community.[ The results of this study also lend evidence to this conclusion. This study shows that participants were prone to having conflict with the medical staff and the other patients followed by the occurrence of aggressive behaviors if they had previously attempted running away or quarreled with staff. The results are similar to the overall general clinical experience.

In terms of personality, inpatients with antisocial personality traits were prone to having aggressive behaviors. There are studies that show patients with schizophrenia with comorbid antisocial personality disorder had more brain damage (such as reduced volume of the thalamus) than patients with schizophrenia who have only mild antisocial tendencies. This brain damage can impact the transmission of sensory information and the ability to control impulses and behavior, [ which may lead to increased display of aggressive behaviors.

The manifestation of symptoms in participants who displayed aggressive behaviors was different than those participants who did not display aggressive behaviors. [ In general, hallucinations and delusions were thought to be major factors that influencing the occurrence of the aggressive behavior in schizophrenia.[ However, this study did not find a strong correlation between positive psychotic symptoms and aggressive behavior. Some scholars believe that symptoms such as anxiety, depression, agitation, tension, and anger manifesting in the context of hallucinations and delusions were the key reasons for aggressive behavior in patients with psychosis. [Other psychiatric symptoms related to aggressive behaviors in participants were paranoia, activation syndrome, and impulsiveness, a finding which is consistent with previous studies.[

4.2 Limitations

This study had the following limitations: (a) previous studies had shown that there was gender differences in aggressive behavior when examining individuals with mental disorders, however, this study did not include female participants. Therefore there was no comparison between genders. (b) There was also a limitation in terms of disorders examined in this study. All the participants in this study had a diagnosis of schizophrenia and patients with other mental disorders (including those who had schizophrenia but also comorbid anxiety, depression or mental retardation) were excluded. (c) All participants of this study came from the same psychiatric unit of our general hospital. Further studies should include a more diverse sample including data from patients with schizophrenia in general hospitals, psychiatric hospitals, and forensic hospitals. (d) In this psychiatric unit, patients with a comorbid substance use disorder are relatively rare. Given this study’s small sample size, it could be hard to detect the relationship between aggressive behavior and comorbid substance abuse. (e) ‘Aggressive behavior’ or risk of aggressive behavior was one of the criteria for involuntary admission, however there was no exploration of the correlation between involuntary admission and aggressive behavior. (f) This study was cross-sectional only, there was no further follow up with patients afterwards, therefore the predictive validity of factors we found related to aggressive behavior could not be verified in relation to actual violence in the future.

4.3 Implication

At present, there are quite a few studies (both in China and abroad) that explore the characteristics of aggressive behavior in hospitalized individuals with mental disorders. In general, it was believed that aggressive behavior could be predicted using factors such as socio-demographic characteristics or clinical manifestations.[ However, the situation in China is different than in many places around the world in terms resources, provision of care and legal priorities, therefore research published abroad may not always apply in our local context.[ Previous research published in China on this topic mostly explored patients’ socio-demographic characteristics, substance abuse, family environment, personality traits, or clinical pathology. [ However, these factors were not comprehensive enough.

In view of the results of our study, the following factors should be considered when evaluating male inpatients with schizophrenia for aggressive behavior: (a) close attention should be paid to patients with aggressive history of violent or antisocial behavior; (b) individuals with mental disorders associated with antisocial traits may be prone to aggressive behavior. Additionally, previous studies have suggested that having schizophrenia with comorbid borderline personality disorder may also increase the risk of aggression in patients with mental disorders.[ Therefore, performing a comprehensive personality assessment when patients with mental disorders are hospitalized is necessary. (c) Clinical symptoms (as measured using PANSS) had a higher correlation with aggressive behavior than did other demographic factors. [ Therefore, clinical symptoms should also be taken into account when evaluating how at risk for aggression an individual may be. Furthermore, anxiety and depressive symptoms should be considered in the evaluation process, not merely positive psychotic symptoms.

Factors influencing aggressive behavior among this population are complex and diverse. This study provides data on factors associated with aggressive behavior in male inpatients with a diagnosis of schizophrenia in hopes of improving safety and psychiatric service.