Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 Dec 30.
Published in final edited form as: Psychiatry Res. 2016 Sep 26;246:203–210. doi: 10.1016/j.psychres.2016.09.039

A Longitudinal Analysis of the Overlap between Violence and Victimization among Adults with Mental Illnesses

Kiersten L Johnson 1,*, Sarah L Desmarais 2, Stephen J Tueller 3, Kevin J Grimm 4, Marvin S Swartz 5, Richard A Van Dorn 6
PMCID: PMC5161544  NIHMSID: NIHMS821763  PMID: 27721058

Abstract

Prior research suggests considerable overlap of violence perpetration and victimization among adults with mental illnesses. However, there has been no examination of how the likelihood of being a victim and/or perpetrator of violence may change over time, nor consideration of clinically-relevant factors affecting these transitions. In a pooled sample of adults with mental illnesses (N=3,473) we employed latent transition analysis to: (a) determine prevalence of four violence and victimization classifications (i.e., non-victim/non-perpetrator, victim only, perpetrator only, and victim-perpetrator) over a 6-month period; (b) calculate the likelihood that adults with mental illnesses will remain in or transition between these classifications over time; and (c) assess the effects of recent substance use, psychiatric symptoms, and suicidal behaviors on transitions over time. At each time point, the majority of participants identified as non-victim/non-perpetrators, followed by victim-perpetrators, victims only, and perpetrators only. Analyses also revealed many individuals transitioned between classifications over time. These distinct pathways towards, and away from, violent outcomes were, in part, a function of recent violence and/or victimization, as well as substance use, psychiatric symptoms, and suicidal behaviors. Findings inform the identification of adults with mental illnesses at risk of violence and victimization and highlight points of intervention.

1. Introduction

Compared to the general population, adults with mental illnesses are at heightened risk of violent outcomes, including violence perpetration (Corrigan & Watson, 2005; Swanson et al., 1990; Van Dorn et al., 2012) and violent victimization (Goodman et al., 1999; Khalifeh et al., 2015; Teplin et al., 2005). Review of the empirical literature in this population shows that there are many common risk factors for violence and victimization, such as substance use and psychiatric symptoms (e.g., Goodman et al., 2001; Silver et al., 2011; Van Dorn et al., 2012), and evidence points towards the co-occurrence of violent outcomes (e.g., Johnson et al., 2016; Swanson et al., 2002). Theoretical perspectives, such as the general theory of crime (Gottfredson & Hirschi, 1990), I3 theory (Finkel, 2014), and the general aggression model (DeWall, Anderson, & Bushman, 2011) suggest that individual characteristics, including alcohol and drug use and exacerbated symptom profiles, may work to inhibit self-control, thus increasing the risk of both perpetration and victimization (Piquero, MacDonald, Dobrin, Daigle, & Cullen, 2005; Pratt, Turanovic, Fox, & Wright, 2013). However, there have been only a handful of investigations of the overlap between violence and victimization in adults with mental illnesses (Jennings et al., 2012; Johnson et al., 2015). Consequently, whether some individuals are at greater risk of both violence and victimization compared to only one outcome – or neither – at any given time is unknown. Moreover, no research has assessed if the likelihood of being a perpetrator and/or victim of violence changes over time, and in response to dynamic, clinically-relevant characteristics.

The few studies investigating violence and victimization in the same timeframe among adults with mental illnesses have consistently observed an overlap of violent outcomes (cf., Havassy & Mericle, 2013; Hiday et al., 2001; Silver et al., 2011). For example, secondary data on 4,474 adults with mental illnesses were examined to identify the baseline prevalence and characteristics of a typology of violence and victimization defined by four groups: non-victim/non-perpetrator, victim only, perpetrator only, and victim-perpetrator (Johnson et al., 2015). Results showed that the majority of the sample reported neither violence nor victimization in a 6-month time period. The remaining participants were more likely to report both perpetration and victimization than either outcome alone. Similar analyses conducted on a sample of 345 discharged psychiatric patients over 12 months also identified a robust overlap between outcomes: of those reporting any violence, over one-third reported both perpetration and victimization (Roaldset & Bjørkly, 2015).

Taken together, findings of extant research support an association between violence and victimization in adults with mental illnesses. However, these studies typically have been cross-sectional in nature, thus defining their intersection as static; participants are classified as victim and/or perpetrator at only one point in time. Other studies only examine the predictive impact of one violent event on a subsequent event, rather than transitions in violence classifications over time. In fact, these roles likely are dynamic in nature and may change in response to variations in clinically-relevant factors that can be targeted via risk management efforts.

If adults with mental illnesses do transition across classifications of victim and/or perpetrator, identifying characteristics associated with an escalation or de-escalation in violent outcomes will be key to the development of risk management strategies and treatment plans. Proximal factors – that is, those that are currently present or occurred in the recent past – may be of particular relevance, as transitions between classifications of violence likely reflect changes in clinical functioning. Specifically, alcohol and drug use are known correlates of violence (Swanson et al., 2006; Van Dorn et al., 2012; Witt et al., 2013) and victimization (Goodman et al., 2001; Hiday et al., 1999; Teasdale, 2009) among adults with mental illnesses, and recent findings support their association with both outcomes in the same sample (Johnson et al., 2015; Roaldset & Bjørkly, 2015). Similarly, although the empirical evidence regarding strength and direction of associations is somewhat mixed, psychiatric symptoms are widely regarded as relevant to risk of violence and victimization in this population (Bjørkly, 2002; Douglas et al., 2009; Fazel et al., 2009; Roaldset & Bjørkly, 2015; Teasdale, 2009; Van Dorn et al., 2016). Finally, suicidal behaviors, including ideation and attempts, have been linked to increased risk of violence (Hillbrand, 2001; Witt et al., 2013; Witt et al., 2014) and victimization (Brown et al., 2013) in adults with mental illnesses; recent findings have also demonstrated stronger associations with violence and victimization than associations observed with either outcome alone (Roaldset & Bjørkly, 2015). In sum, the factors examined in this study are clinically-relevant and may be targeted in treatment to minimize high or maintain low violence risk.

1.1 The present study

The goals of the present study were threefold. First, we sought to examine the overlap between violence and victimization over time in a large, heterogeneous sample of adults with mental illnesses (N=3,473). Second, we sought to determine whether adults with mental illnesses transition between classifications of victim and/or perpetrator of violence over time. Third, we sought to identify whether proximal, clinically-relevant indicators are associated with transitions between or persistence in these classifications over time. To that end, we employed latent transition analysis (LTA) to: (1) determine the prevalence of violence and victimization as defined by four classifications – non-victim/non-perpetrator, victim only, perpetrator only, and victim-perpetrator – over the 6-month study period; (b) calculate the likelihood that adults with mental illnesses will remain in or transition between these classifications over 3-month and 6-month periods; and (c) assess the effects of recent substance use, psychiatric symptoms, and suicidal behaviors on transitions from non-victim/non-perpetrator and victim-perpetrator classifications over 3-month and 6-month periods.

2. Methods

2.1 Data

Data were pooled from four studies of adults with mental illnesses: (1) Facilitated Psychiatric Advance Directive Study (n=473; Swanson et al., 2006); (2) MacArthur Mental Disorder and Violence Risk Study (n=1,136; Steadman et al., 1998); (3) Schizophrenia Care and Assessment Program (n=404; Swanson et al., 2004); and (4) Clinical Antipsychotic Trials of Intervention Effectiveness Study (n=1,460; Lieberman et al., 2005). These studies enrolled a range of participants, from exacerbated inpatients to partially remitted outpatients. A total of 3,473 adults with mental illnesses were included at baseline. Longitudinal analyses across 3- and 6-month time frames included samples of 908 and 2,512 participants, respectively. Sampling and inclusion criteria for each data source are described briefly below.

2.1.1 Facilitated Psychiatric Advance Directive Study (F-PAD)

A random sample of clients prescreened for eligibility was obtained from two North Carolina mental health systems. Inclusion criteria were: (a) 18–65 years of age; (b) schizophrenia-spectrum or major mood disorder; and (c) in treatment at the time of the study. Data collection began in 2003; assessments used in the present study were conducted at baseline and 6-month follow-up.

2.1.2 MacArthur Mental Disorder and Violence Risk Study (MacRisk)

Eligible participants were sampled from three acute inpatient facilities, with recruitment adjusted to maintain a consistent distribution of age, sex, and ethnicity across sites. Inclusion criteria were: (a) English-speaking White, Black, or Hispanic patients; (b) 18–40 years of age; and (c) schizophrenia-spectrum, depression, mania, brief reactive psychosis, or delusional disorder. Data were collected through participant and collateral interviews and hospital abstractions beginning in 1992; the present study used assessments conducted at baseline, 3-month, and 6-month follow-ups.

2.1.3 Schizophrenia Care and Assessment Program (SCAP)

Participants were recruited from treatment facilities across North Carolina using both sequential inpatient admissions and a random selection of outpatients, with eligibility limited to: (a) 18 years of age or older; (b) schizophrenia; and (c) current service use. Data collection began in 1997; we include assessments from baseline and 6-month follow-up in the present analyses.

2.1.4 Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE)

Participants were recruited from 57 sites across the United States. Inclusion criteria were: (a) 18–65 years of age; (b) schizophrenia; and (c) ability to take oral antipsychotics. Data collection began in 2001. The present study includes assessments from baseline and 6-month follow-up, as well as any end-of-phase assessments conducted prior to 6-month follow-up.

The current study protocol was approved by IRBs from North Carolina State University, RTI International, and Arizona State University. All participants in the source studies gave written informed consent.

2.2 Measures

2.2.1 Outcome variables

Prevalence and severity of community violence and victimization was assessed in all studies using the MacArthur Community Violence Screening Instrument (MCVSI; Steadman et al., 1998). The MCVSI is comprised of eight questions each for violence and victimization, derived from the Revised Conflict Tactics Scale (CTS2; Straus et al., 1996). The questions assess: (a) pushing, grabbing, or shoving; (b) kicking, biting, or choking; (c) slapping; (d) throwing an object; (e) hitting with a fist or object; (f) sexual assault; (g) threatening with a weapon in hand; and (h) using a weapon. For each item, participants are first asked if someone did this to them, then if they did this to someone else. In addition to its inclusion in the parent studies, the MCVSI has been used in other studies assessing community violence (e.g., Davidson et al., 2009; Michie & Cooke, 2006). For the present analyses, we dichotomized responses of “violence” (e.g., choking) and “other aggressive acts” (e.g., throwing an object) to obtain prevalence of violent outcomes by 3- and 6-month follow-ups.1

2.2.2 Covariates

Alcohol and drug use were assessed at baseline with multiple measures across studies, including the CAGE questionnaire (Mayfield et al., 1974), urine drug screens, self-report, the Alcohol and Drug Use Scales (Drake et al., 1990), and the Structured Clinical Interview for DSM-IV (SCID; First et al., 1996). We harmonized data across studies with 3-level summary variables (0=abstinence, 1=non-problematic use, 2=problematic use).

Psychiatric symptoms were assessed at baseline with the Positive and Negative Syndrome Scale (Kay et al., 1987) in the CATIE and SCAP, and the BPRS in the F-PAD and MacRisk. The MacRisk additionally used the MacArthur-Maudsley Delusions Assessment Schedule (Appelbaum et al., 1999). Briefly, a factor analytic cross-validation approach was employed with exploratory factor analyses conducted on a 50% random sample of data where four factors were identified.2 These factors retained all items, and included affective (e.g., anxiety, depression, hostility), positive (e.g., delusions, suspiciousness, grandiosity), negative (e.g., social withdrawal, blunted affect, emotional withdrawal), and disorganized cognitive processing (DCP) (e.g., poor impulse control, excitement, mannerisms & posturing) symptoms. This model was then evaluated and supported using a confirmatory factor model with the remainder of the data, with good model fit (CFI = 0.915, RMSEA = 0.078, TLI = 0.906) (Van Dorn et al., 2016). Factor scores for each latent variable were then created from unidimensional factor models using expected a posteriori (EAP) estimates. EAP estimates are calculated as the mean of the posterior predicted distribution of scores for an individual based on his/her response pattern and the estimated model parameters. The factor scores account for incomplete data and are based on the items to which the individual responded.

Indicators of suicidal ideation and suicide attempt were harmonized across studies. For each outcome, we created a dichotomous variable (yes/no) of any ideation or attempt in the six months prior to baseline.

2.3 Data analysis

Descriptive statistics were calculated for all covariates at baseline using SPSS v.22. Subsequent analyses were conducted with Mplus 7.31 (Muthén & Muthén, 1998–2014). Model specification for the LTA was informed by prior work identifying a typology of violence and victimization among adults with mental illnesses (Johnson et al., 2015); accordingly, two parameter restrictions were imposed. First, groups were specified using fixed thresholds, such that group membership was directly derived from reported violence and victimization (e.g., an individual reporting victimization, but not violence, was identified as a victim only at the corresponding time point). Second, these thresholds were held to be equal over time, signifying measurement invariance.

We conducted an initial model in LTA to obtain prevalence of the four groups at baseline, 3-month follow-up, and 6-month follow-up and calculate probabilities of transitioning between these groups from baseline to 3-month follow-up and from baseline to 6-month follow-up. We then ran a series of LTA models to calculate multinomial logistic regression coefficients linking baseline covariates to transitions between groups over the 3-month and 6-month periods. Specifically, we assessed the bivariate, predictive effects of baseline measures of recent substance use, psychiatric symptoms, and suicidal behaviors on transition probabilities. Separate models were conducted with non-victim/non-perpetrator and victim-perpetrator classifications used as reference classes. Mplus uses full information maximum likelihood to estimate transition probabilities. Listwise deletion was used in cases of missing data for regression analyses. Given the number of effects, we implemented a Benjamini and Hochberg adjustment, which is a false discovery rate adjustment for p-values and is more appropriate than family-wise error rate adjustments (Glickman, Rao, & Schultz, 2014).

3. Results

Frequencies and means for covariates across non-missing and missing cases are reported in Table 1. At baseline, the majority of the sample indicated abstinence from alcohol and drugs. Problematic use of alcohol was reported by approximately one-fourth of participants; problematic use of drugs, by about one-fifth. Roughly one-third of the sample reported suicidal ideation, and over one-tenth reported an actual attempt. Baseline characteristics of the participants who were present at follow-up interviews differed from those who were not present in several ways (see Table 1); in general, those present at 3- and 6-month follow-up exhibited poorer psychosocial functioning at baseline. These factors were included as predictors in subsequent analyses. All parent studies reported significant levels of violence and victimization at all three time points.

Table 1.

Frequencies and Means of Covariates.

Full Sample
(N=3,473)		 3-Month Follow-Up
(N=908)		 6-Month Follow-Up
(N=2,512)
Non-Missing Missing Non-Missing Missing
n % n % n % p n % n % p
Study <.001 <.001
  F-PAD 473 13.6 0 0.0 473 18.4 381 15.2 92 9.6
  MacRisk 1,136 32.7 845 93.1 291 11.3 855 34.0 281 29.2
  SCAP 404 11.6 0 0.0 404 15.8 339 13.5 65 6.7
  CATIE 1,460 42.0 63 6.9 1,397 54.5 936 37.3 525 54.5
Baseline Alcohol Use <.001 <.001
  Abstinence 2,474 55.3 270 29.8 2,204 61.8 1,286 51.3 1,188 60.4
  Non-problematic use 858 19.2 254 28.0 604 16.9 534 21.3 324 16.5
  Problematic use 1,141 25.5 382 42.2 759 21.3 686 27.4 455 23.1
Baseline Drug Use <.001 <.001
  Abstinence 3,022 67.7 470 51.9 2,552 71.7 1,629 65.1 1,393 70.9
  Non-problematic use 476 10.7 143 15.8 333 9.4 301 12.0 175 8.9
  Problematic use 969 21.7 293 32.3 676 19.0 572 22.9 397 20.2
Baseline Suicidal Ideation <.001 .006
  No 2,913 65.2 301 33.2 2,612 73.4 1,588 63.5 1,325 67.5
  Yes 1,552 34.8 606 66.8 946 26.6 913 36.5 639 32.5
Baseline Suicide Attempt <.001 <.001
  No 3,928 88.0 588 64.9 3,340 93.9 2,145 85.8 1,783 90.8
  Yes 536 12.0 318 35.1 218 6.1 355 14.2 181 9.2
Baseline Symptoms (M±SD)
  Affective .26±.89 .59±.99 .18±.84 <.001 .30±.90 .21±.88 <.001
  Positive .09±.93 −.11±1.05 .15±.89 <.001 .19±.94 −.03±.91 <.001
  Negative −.07±.92 −.32±.88 −.01±.92 <.001 .04±95 −.22±.87 <.001
  DCP −.02±.88 −.36±.77 .07±.89 <.001 .09±.91 −.15±.83 <.001
Open in a new tab

Notes. %=valid percent. Means of variables were compared across missing and non-missing participants by t tests, and proportions were compared by chi square tests.

Results of the first LTA model (without covariates) are presented in Table 2. Specifically, the top section shows the prevalence of classifications at baseline, 3-month follow-up, and 6-month follow-up observed in participants at each time point. The majority of participants identified as non-victim/non-perpetrator at all three time points, followed by victim-perpetrator, victim only, and perpetrator only.

Table 2.

Prevalence and transition probabilities of observed class over time.

Prevalence Observed Class
NVNP VO PO VP
n % n % n % n %
Baseline 2,172 62.5 456 13.2 182 5.3 647 18.7
3-Month Follow-Up 499 55.0 130 14.3 72 7.9 207 22.8
6-Month Follow-Up 1,781 70.9 191 7.6 115 4.6 425 16.9
Transition probabilities 3-Month Follow-Up
NVNP VO PO VP
Baseline NVNP .899 .045 .017 .040
VO .577 .211 .055 .156
PO .591 .053 .193 .162
VP .457 .142 .082 .319
6-Month Follow-Up
NVNP VO PO VP
Baseline NVNP .831 .051 .034 .083
VO .533 .170 .053 .243
PO .547 .052 .151 .251
VP .422 .115 .069 .393
Open in a new tab

Notes. %=valid percent. NVNP=non-victim/non-perpetrator; VO=victim only; PO=perpetrator only; VP=victim-perpetrator. Bolded transition values represent the probability of staying in the same class.

The bottom section of Table 2 shows transition probabilities between baseline and 3-month follow-up and baseline and 6-month follow-up. The bolded, diagonal values represent stability in a group from one time point to the next. For example, 89.9% (n = 1,953) of those who were classified as non-victim/non-perpetrator at baseline maintained that classification through three months. Off-diagonal values indicate movement across groups. For example, 4.5% (n = 98) of those classified as non-victim/non-perpetrator at baseline were re-classified as victim only by three months. Estimates for both transition periods (i.e., from baseline to 3-month follow-up and from baseline to 6-month follow-up) indicated that non-victim/non-perpetrators at baseline were likely to persist in that role over time. In contrast, victims and/or perpetrators were more likely to move across groups than to remain stable; these participants were most likely to report no violent outcome at follow-up assessments or, less frequently, both violence and victimization.

Table 3 presents odds ratios for the effects of covariates on transitions away from non-victim/non-perpetrator and victim-perpetrator classifications between baseline and 3-month follow-up. In comparison to the LTA without covariates, all covariate models reported lower AIC and BIC values and significant changes in −2 log likelihood values (ps < .001), indicating improved model fit. When non-victim/non-perpetrator at baseline served as reference in regression analyses, all covariates (with the exception of affective and positive symptoms) demonstrated significant predictive effects over time. As shown in the leftmost columns, alcohol use increased the risk of victimization or violence and victimization by 3-month follow-up; drug use increased the risk of victimization only. Increased negative symptoms were associated with a decreased likelihood of experiencing both violence and victimization, whereas increased DCP symptoms were associated with a decreased likelihood of perpetration only. Reports of suicidal behaviors at baseline were associated with increased risk of victimization only, as well as violence and victimization, by 3-month follow-up.

Table 3.

Effects of covariates on class transitions between baseline and 3-month follow-up.

NVNP at baseline VP at baseline
VO PO VP NVNP VO PO
Substance use OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI
  Alcohol use 1.56** 1.21–2.01 1.31 0.79–2.17 2.58*** 1.91–3.48 0.75 0.58–0.98 0.69 0.48–1.00 0.54* 0.34–0.85
  Drug use 1.46* 1.10–1.92 1.25 0.74–2.10 1.35 1.01–1.81 0.88 0.70–1.12 0.93 0.67–1.30 0.89 0.58–1.37
Psych. symptoms
  Affective 1.49 1.06–2.09 1.94 1.11–3.40 1.22 0.86–1.74 0.75* 0.60–0.94 1.06 0.76–1.47 0.96 0.68–1.35
  Positive 0.73 0.49–1.07 0.59 0.30–1.16 0.77 0.55–1.07 1.07 0.86–1.34 1.54* 1.14–2.09 1.15 0.79–1.67
  Negative 0.67 0.48–0.93 0.70 0.46–1.05 0.63* 0.47–0.84 1.15 0.90–1.46 1.18 0.83–1.68 1.34 0.89–2.02
  DCP 0.68 0.47–1.00 0.28*** 0.14–0.56 0.70 0.52–0.95 1.56** 1.20–2.04 1.35 0.93–1.96 1.24 0.83–1.87
Suicidal behavior
  Ideation 2.22* 1.33–3.72 2.18 0.90–5.29 2.43** 1.42–4.18 0.67 0.43–1.06 0.96 0.50–1.83 1.44 0.59–3.52
  Attempt 4.43*** 2.39–8.24 3.47 1.21–9.97 3.47*** 1.75–6.89 0.66 0.42–1.04 0.59 0.31–1.14 0.78 0.35–1.73
Open in a new tab

Notes. n=908.

*

p < .05

**

p < .01

***

p < .001.

NVNP=non-victim/non-perpetrator; VO=victim only; PO=perpetrator only; VP=victim-perpetrator.

Top row indicates reference group in multinomial regression; the row below indicates groups at 3-month follow-up.

When the baseline reference group consisted of victim-perpetrators, alcohol use, affective, positive, and DCP symptoms were associated with transitions from baseline to 3-month follow-up. Alcohol use significantly decreased the likelihood of violence in the absence of victimization. Decreased affective symptoms were associated with a transition to no violent outcomes by 3-month follow-up. Increased positive symptoms were associated with reports of only victimization by 3-month follow-up, and increased DCP symptoms were associated with reporting no perpetration nor victimization.

Table 4 presents odds ratios for the effects of covariates on transitions away from non-victim/non-perpetrator and victim-perpetrator classifications between baseline and 6-month follow-up. When participants identified as non-victim/non-perpetrator at baseline, increased negative and DCP symptoms lessened the odds of reporting both violence and victimization at 6-month follow-up. As was the case with 3-month analyses, suicidal ideation and suicide attempt increased the risk of victimization as well as violence and victimization from baseline to 6-month follow-up. Notably, baseline alcohol and drug use did not demonstrate significant predictive effects in increasing the risk of someone moving from the non-violent/non-perpetrator classification to one of the violence classifications by 6-month follow-up.

Table 4.

Effects of covariates on class transitions between baseline and 6-month follow-up.

NVNP at baseline VP at baseline
VO PO VP NVNP VO PO
Substance use OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI OR 95% CI
  Alcohol use 1.08 0.73–1.58 0.95 0.63–1.44 1.40 0.98–2.01 0.29*** 0.21–0.41 0.48** 0.32–0.74 0.55* 0.36–0.85
  Drug use 0.91 0.56–1.48 1.12 0.75–1.69 1.06 0.70–1.61 0.30*** 0.22–0.42 0.53** 0.36–0.79 0.61* 0.40–0.92
Psych. symptoms
  Affective 1.47 0.93–2.33 0.99 0.63–1.56 1.39 0.98–1.97 0.30*** 0.20–0.45 0.57* 0.38–0.85 0.69 0.41–1.15
  Positive 0.74 0.47–1.18 0.72 0.47–1.08 0.71 0.44–1.17 4.17* 1.45–12.01 1.17 0.61–2.24 1.39 0.61–3.18
  Negative 0.72 0.47–1.10 0.68 0.47–0.99 0.72* 0.54–0.96 5.84*** 3.05–11.16 2.84* 1.38–5.82 2.36* 1.20–4.61
  DCP 0.74 0.46–1.20 0.83 0.53–1.31 0.59* 0.40–0.87 7.25*** 3.69–14.24 1.61 0.79–3.28 2.11 0.91–4.89
Suicidal behavior
  Ideation 2.88* 1.38–5.99 1.49 0.66–3.36 4.85*** 2.38–9.89 0.05*** 0.03–0.11 0.38* 0.18–0.79 0.27* 0.11–0.68
  Attempt 3.18* 1.36–7.43 1.48 0.42–5.22 10.26*** 5.21–20.22 0.07*** 0.04–0.14 0.30* 0.13–0.71 0.24* 0.09–0.65
Open in a new tab

Notes. n=2,512.

*

p < .05

**

p < .01

***

p < .001.

NVNP=non-victim/non-perpetrator; VO=victim only; PO=perpetrator only; VP=victim-perpetrator.

Top row indicates reference group in multinomial regression; the row below indicates groups at 6-month follow-up.

When victim-perpetrators at baseline served as reference group, all covariates demonstrated varying degrees of significant predictive effects. Specifically, both alcohol and drug use decreased the likelihood that participants would transition to another classification over the 6-month period. There was an inverse relationship between affective symptoms and transitioning to non-victim/non-perpetrator or victim only by 6-month follow-up when someone started as a victim-perpetrator; conversely, increased negative symptoms were associated with transitioning from victim-perpetrator at baseline to another classification by six months. Heightened positive and DCP symptoms both increased the likelihood of moving from the victim-perpetrator classification at baseline to the non-victim/non-perpetrator classification by six months. Suicidal behaviors significantly decreased the likelihood that participants would move to any other classification by 6-month follow-up.

4. Discussion

4.1 Summary of findings

At baseline, the majority of participants identified as non-victim/non-perpetrators (62.5%), followed by victim-perpetrators (18.7%), victims only (13.2%), and perpetrators only (5.3%). Though this ordering remained unchanged at 3- and 6-month follow-up, transition probabilities yielded by LTA demonstrated that many individuals did transition between classifications of victim and/or perpetrator. Prior research has demonstrated heterogeneity in violent events for this population, and with similar rates across violence classifications (Silver et al., 2011; Roaldset & Bjørkly, 2015). We additionally find that involvement in violent events changes over time, and in estimable ways that are associated with clinically-relevant factors. Specifically, observed escalations and de-escalations in violent outcomes over different time periods – and the role of substance use, psychiatric symptoms, and suicidal behaviors in those transitions – highlight potential intervention points for this population.

Results showed that the role of clinical factors in the escalation of violent outcomes – that is, their effects in transitions away from non-victim/non-perpetrator status at baseline – demonstrated a greater number of significant effects over the 3-month, rather than 6-month period, though the direction of effects remained the same across both time frames. By 3-month follow-up, results showed that increased alcohol use, drug use, and suicidal behaviors increased the risk of reporting victimization or both violence and victimization. Decreased negative and DCP symptoms were also associated with a heightened likelihood of reporting both violent outcomes. By 6-month follow-up, increased suicidal behaviors and decreased negative and DCP symptoms remained positively associated with risk of violent outcomes. Specifically, suicidal behaviors increased the likelihood of victimization or violence and victimization, and decreased negative and DCP symptoms increased the risk of reporting both violence and victimization.

In contrast, results demonstrated that the role of clinical factors in the de-escalation of violent outcomes over time – that is, their effects in transitions away from victim-perpetrator status at baseline – was more prominent over the 6-month period, though again, the direction of effects were mirrored across both time periods. By 3-month follow-up, results showed that alcohol use decreased the likelihood of transitioning to perpetrator only status. Increased affective symptoms lessened the likelihood of transitioning to non-victim/non-perpetrator status; increased positive and DCP symptoms were associated with de-escalations to victimization and no violent outcomes, respectively. Over the 6-month follow-up period, increased alcohol and drug use, affective symptoms, and suicidal behaviors lessened the likelihood of transitioning to another classification, whereas increased positive, negative, and DCP symptoms heightened the likelihood of de-escalation.

These findings, which offer insight into the role of clinical factors in the escalation or de-escalation of violent outcomes over time, are consistent with some prior studies examining the correlates of violence and victimization (albeit as separate outcomes and without a specific focus on transitions between violence classifications). For example, alcohol and drug use have been identified as correlates of violence (Swanson et al., 2006; Van Dorn et al., 2012; Witt et al., 2013) and victimization (Goodman et al., 2001; Hiday et al., 1999). Additionally, recent findings showed significant effects on both outcomes in the same sample (Roaldset & Bjørkly, 2015). Our analyses demonstrate that alcohol use is involved with transitions towards violence, on the one hand, but also that decreased alcohol use is associated with transitions away from violence, on the other hand. As such, integrated interventions, such as dual diagnosis motivational interviewing, may minimize high or maintain low risk of violence and victimization associated with alcohol use (Drake, Mueser, Brunette, & McHugo, 2004; Swanson, Pantalon, & Cohen, 1999).

Prior research has also considered the role of psychiatric symptoms in predicting violence (Johnson et al., 2016), and violence and victimization (Roaldset & Bjørkly, 2015; Silver et al., 2011), though the operational definitions of violence and victimization differed in the latter studies. Our study investigated the effects of a 4-factor model of psychiatric symptoms (Van Dorn et al., 2016) on both violence and victimization. Overall, our findings show that decreased positive, negative, and DCP symptoms and increased affective symptoms were generally associated with worse outcomes and thus provide multiple targets for risk management efforts. Indeed, significant effects of psychiatric symptoms in the escalation and de-escalation of violent outcomes were more likely than that of alcohol and drug use to emerge across both 3- and 6-month time frames. Cognitive behavioral and psychopharmacological interventions to address internalizing and externalizing symptoms across all four factors may work to decrease risk of violent outcomes (Van Dorn et al., 2016). The affective factor, in particular, likely has direct and indirect effects on violence risk. For example, hostility (one of the included items) is associated with violence, medication noncompliance, impaired insight, and substance use (Bartels, Drake, Wallach, & Freeman, 1991; Czobor et al., 2015; Lindenmayer et al., 2009; Volavka, Czobor, Citrome, & Van Dorn, 2014; Volavka et al., 2016). As a result, an increase in affective symptoms should prompt further investigation of clinical functioning across other domains to inform assessment of violence risk and the subsequent treatment response.

Prior research has also observed associations of suicidal behaviors with violence (Hillbrand, 2001; Witt et al., 2013; Witt et al., 2014) and victimization (Brown et al., 2013). The present study is among the first to assess the effects of suicidal ideation and attempt on both outcomes in the same sample (see Roaldset & Bjørkly, 2015). Our findings highlight the role of suicidal behaviors in the escalation from no violent outcomes to victimization-only or both violence and victimization over both 3- and 6-month periods, as well as in the de-escalation from both violent outcomes to another violence classification over a 6-month period. This latter finding suggests that the effects of suicidal behaviors on short-term violence and victimization may be tempered by the recent occurrence of violent outcomes. Nevertheless, the relationship between suicidal behaviors and violent outcomes provides another target for clinical intervention, vis-à-vis clinical assessment and provision of tailored treatment (e.g., medication monitoring, cognitive behavioral therapy).

Finally, our tests of statistical significance were based on a Benjamini and Hochberg adjustment. This adjustment is notable as substance use, psychiatric symptoms, and suicide-related behaviors were significantly associated with additional transitions without the adjustment. For example, while affective symptoms were inversely and significantly associated with transitioning from victim-perpetrator to non-victim/non-perpetration (i.e., away from violent behaviors) at 3-month follow-up with the Benjamini and Hochberg adjustment, a similar trend was found for affective symptoms and an escalation towards violence perpetration and victimization as separate outcomes in the same follow-up period. Specifically, with odds ratios of 1.49 (NVNP to VO) and 1.94 (NVNP to PO), affective symptoms for these two outcomes had adjusted p-values of .059 and unadjusted p-values of .022 and .019, respectively. In another example, the unadjusted effect for suicide attempts in the escalation towards violence perpetration was significant (p=.021) with an odds ratio of 3.47, whereas the adjusted effect was not (p=.059). Additional trends were noted for alcohol use in the transition away from violent behaviors at 3-month follow-up; while alcohol use was inversely and significantly associated with the transition from victim-perpetrator to perpetrator only, unadjusted effects for the transitions to non-victim/non-perpetrator (OR=0.75) and victim only (OR=0.69) were also significant with p-values of .033 and .047, respectively; adjusted p-values were .083 and .107. While we maintain that our implementation of adjusted p-values was methodologically appropriate given the number of effects examined simultaneously, examination of the unadjusted effects is also worthwhile to inform future research and begin to develop a cumulative science related to clinically-relevant factors that can be addressed within effective risk management strategies.

4.2 Implications

Our findings show that adults with mental illnesses vary in their involvement in violent events. Notably, the majority of our sample did not report violence or victimization; those who did were more likely to transition from one classification to another over the 6-month study period, thus illustrating that one’s classification as a victim and/or perpetrator is not fixed. This enhanced perspective on the “dangerousness” of adults with mental illnesses – or rather, the diversity of their violence or victimization – may translate to improved public perceptions and an increase in service use in this population (Corrigan, 2004). Furthermore, modifiable, clinicallyrelevant factors are associated with escalations or de-escalations in violence and victimization; clinicians can use these associations as points of leverage in treatment.

In showing that recent violence and/or victimization create distinct pathways to future violent outcomes, findings offer important insights for clinical practice. Notably, non-victim/non-perpetrators and victim-perpetrators exhibited particular susceptibility to violent outcomes as a function of clinical characteristics over 3-month and 6-month follow-ups, respectively. Consideration of these pathways, as well as the differing effects of substance use, psychiatric symptoms, and suicidal behaviors, may assist clinicians in identifying those at risk of future violence and/or victimization. Our findings support mounting evidence of the importance of proximal, clinically-relevant factors in routine and ongoing risk assessment and management approaches (Johnson et al., 2016; McNiel et al., 2003; Sadeh et al., 2014). Indeed, clinical characteristics were especially predictive of an escalation in violent outcomes over the shorter, 3-month follow-up period. Moreover, the change in risk of violent outcomes was signified by multiple factors, suggesting that violence risk assessment is not only an iterative process, but one that requires actively attending to how a change in one domain of functioning affects or is reflected in changes in another.

Furthermore, findings from the present study can assist clinicians in the tailoring of treatment according to recent involvement in violent outcomes. For example, for a patient who has recently been involved in violent outcomes, both as perpetrator and victim, short-term violence prevention may be most effective when incorporating evidence-based practices to reduce alcohol use and manage high affective symptoms while attending to the effects of low negative and DCP symptoms. Violence prevention over a longer period of time (e.g., six months or longer) may additionally include treatment to reduce drug use and suicidal behaviors, while still attending to improvements in positive symptoms and how that affects other domains of functioning. In contrast, for someone who has not been involved in recent violence or victimization, short-term treatment to effectively maintain low violence risk may work best when implementing evidence-based practices for substance use and suicidal behaviors, with attention also being paid to negative and DCP symptomatology. Over longer periods of time, clinical focus is best directed at symptom management (particularly negative and DCP symptoms) and suicidal behaviors.

4.3 Limitations

Findings of the present study should be considered in light of its limitations. First, violence and victimization data were derived from self-report and are thus susceptible to effects of social desirability, recall bias, and other errors. Though self-report has been found to be a reliable and valid method in violence and victimization research (Huizinga & Elliott, 1986; Van Dorn et al., 2010), other sources (e.g., hospital and arrest records) may capture additional violent events. Second, attrition and variation in assessment times across studies resulted in missing outcome data at 3-month and 6-month follow-ups. Comparisons between 3- and 6-month follow-ups revealed significant differences in missing and non-missing cases, such that participants present at follow-up assessments had poorer psychosocial functioning at baseline. Results may thus not generalize to lower risk populations. Third, we assess the incidence – not the context – of violent outcomes. As a result, we are unable to determine, in the case of victim-perpetrators, if violence preceded victimization, victimization preceded violence, or violence and victimization occurred simultaneously. Across studies, however, violence in self-defense was not counted as perpetration. Fourth, the included studies all take place in the United States. Findings may not generalize to countries with differing societal contexts and mental health systems. Fifth, we focused on several proximal, clinically-relevant characteristics, but there are other factors that may provide further insight into transitions across violence classifications, such as homelessness and access to treatment. Moreover, static factors such as sex may warrant consideration in further research, though, in the present study, chi-square test results indicated that sex did not affect transitions across violence classifications; this is illustrated in the Supplemental Figure. Finally, we conducted bivariate analyses to assess the individual effects of factors at both follow-up assessments. Further research is needed to examine unique effects of or potential interactions among factors that increase or decrease risk.

4.4 Conclusions

This paper examines the classification of victims and/or perpetrators over time among adults with mental illnesses, and assesses the effects of proximal, clinically-relevant variables on transitions between classifications over 3- and 6-month periods. Findings point to heterogeneity in risk of violent outcomes. Notably, many individuals transitioned between classifications of victim and/or perpetrator during 3-month and 6-month follow-ups. Indeed, distinct pathways to future violent outcomes emerged as a function of recent violence and/or victimization, as well as modifiable clinically-relevant factors, including psychiatric symptoms, substance use, and suicidal behaviors. Consideration of these pathways and associated factors may assist in the identification of adults with mental illnesses at acute or persistent risk of violence and victimization. Furthermore, our findings on the differing effects of substance use, psychiatric symptoms, and suicidal behaviors offer insight into how interventions may be tailored to minimize high or maintain low violence risk and, ultimately, improve outcomes for adults with mental illnesses.

Supplementary Material

Highlights.

  • Experiences with violent outcomes vary widely among adults with mental illnesses.

  • An individual’s classification as a victim and/or perpetrator may change over time.

  • Clinical characteristics are associated with changes in classification over time.

  • Interventions targeting clinical characteristics may prevent violent outcomes.

Acknowledgments

Role of the funding source

This work was supported by the National Institute of Mental Health (grant number R01MH093426, PI: Dr. Van Dorn). The sponsor did not have any role in the conduct of the research.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

Due to varied assessment schedules across parent studies, only CATIE and MacRisk participant data were available for inclusion in 3-month analyses.

2

The MacArthur Mandated Community Treatment Study (Monahan et al., 2005), in addition to the four longitudinal studies, was used in the calculation of factor scores for psychiatric symptoms.

Conflict of interest

None

References

  1. Appelbaum PS, Robbins PC, Roth LH. Dimensional approach to delusions: Comparison across types and diagnoses. Am. J. Psychiatry. 1999;156:1938–1943. doi: 10.1176/ajp.156.12.1938. [DOI] [PubMed] [Google Scholar]
  2. Bartels SJ, Drake RE, Wallach MA, Freeman DH. Characteristic hostility in schizophrenic outpatients. Schizophr. Bull. 1991;17:163–171. doi: 10.1093/schbul/17.1.163. [DOI] [PubMed] [Google Scholar]
  3. Bjørkly S. Psychotic symptoms and violence toward others: A literature review of some preliminary findings. Aggress. Violent Beh. 2002;7:605–631. [Google Scholar]
  4. Brown R, DuMont J, MacDonald S, Bainbridge D. A comparative analysis of victims of sexual assault with and without mental health histories: Acute and follow-up care characteristics. J. Forensic Nurs. 2013;9:76–83. doi: 10.1097/JFN.0b013e31828106df. [DOI] [PubMed] [Google Scholar]
  5. Corrigan PW. How stigma interferes with mental health care. Am. Psychol. 2004;59:614–625. doi: 10.1037/0003-066X.59.7.614. [DOI] [PubMed] [Google Scholar]
  6. Corrigan PW, Watson AC. Understanding the impact of stigma on people with mental illness. World Psychiatry. 2002;1:16–20. [PMC free article] [PubMed] [Google Scholar]
  7. Czobor P, Van Dorn RA, Citrome L, Kahn RS, Fleischhacker WW, Volavka J. Treatment adherence in schizophrenia: A patient-level meta-analysis of combined CATIE and EUFEST studies. Eur. Neuropsychopharmacol. 2015;25:1158–1166. doi: 10.1016/j.euroneuro.2015.04.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Davidson KM, Tyrer P, Tata P, Cooke D, Gumley A, Ford I, Walker A, Bezlyak V, Seivewright H, Robertson H, Crawford MJ. Cognitive behaviour therapy for violent men with antisocial personality disorder in the community: An exploratory randomized controlled trial. Psychol. Med. 2009;39:569–577. doi: 10.1017/S0033291708004066. [DOI] [PubMed] [Google Scholar]
  9. Desmarais SL, Van Dorn RA, Johnson KL, Grimm KJ, Douglas KS, Swartz M. Community violence perpetration and victimization among adults with mental illnesses. Am. J. Public Health. 2014;104:2342–2349. doi: 10.2105/AJPH.2013.301680. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. DeWall CN, Anderson CA, Bushman BJ. The general aggression model: Theoretical extensions to violence. Psychol. Viol. 2011;1:245–258. [Google Scholar]
  11. Douglas KS, Guy LS, Hart SD. Psychosis as a risk factor for violence to others: A meta-analysis. Psychol. Bull. 2009;135:679–706. doi: 10.1037/a0016311. [DOI] [PubMed] [Google Scholar]
  12. Drake RE, Mueser KT, Brunette MF, McHugo GJ. A review of treatments for people with severe mental illnesses and co-occurring substance use disorders. Psychiatr. Rehabil. J. 2004;27:360. doi: 10.2975/27.2004.360.374. [DOI] [PubMed] [Google Scholar]
  13. Drake RE, Osher FC, Noordsy DL, Hurlbut SC, Teague GB, Beaudett MS. Diagnosis of alcohol use disorders in schizophrenia. Schizophrenia Bull. 1990;16:57–67. doi: 10.1093/schbul/16.1.57. [DOI] [PubMed] [Google Scholar]
  14. Fazel S, Gulati G, Linsell L, Geddes JR, Grann M. Schizophrenia and violence: Systematic review and meta-analysis. PLoS Med. 2009;6:e1000120. doi: 10.1371/journal.pmed.1000120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Finkel EJ. The I3 model: Metatheory, theory, and evidence. Adv. Exp. Soc. Psychol. 2014;49:1–104. [Google Scholar]
  16. First MB, Gibbon M, Spitzer RL, Williams JBW. User’s guide for the structured clinical interview for DSM-IV axis I Disorders—Research version. New York: New York State Psychiatric Institute; 1996. [Google Scholar]
  17. Goodman LA, Salyers MP, Mueser KT, Rosenberg SD, Swartz M, Essock SM, Swanson J. Recent victimization in women and men with severe mental illness: Prevalence and correlates. J. Trauma. Stress. 2001;14:615–632. doi: 10.1023/A:1013026318450. [DOI] [PubMed] [Google Scholar]
  18. Goodman L, Thompson K, Weinfurt K. Reliability of violent victimization and PTSD among men and women with serious mental illness. J. Trauma. Stress. 1999;12:587–599. doi: 10.1023/A:1024708916143. [DOI] [PubMed] [Google Scholar]
  19. Havassy BE, Mericle AA. Recent violence among persons entering short-term residential mental health and substance abuse treatment. J. Dual Diagn. 2013;9:222–227. doi: 10.1080/15504263.2013.804773. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hiday VA, Swanson JW, Swartz MS, Borum R, Wagner HR. Victimization: A link between mental illness and violence? Int. J. Law Psychiat. 2001;24:559–572. doi: 10.1016/s0160-2527(01)00091-7. [DOI] [PubMed] [Google Scholar]
  21. Hiday VA, Swartz MS, Swanson JW, Borum R, Wagner HR. Criminal victimization of persons with severe mental illness. Psychiatr. Serv. 1999;50:62–68. doi: 10.1176/ps.50.1.62. [DOI] [PubMed] [Google Scholar]
  22. Hillbrand M. Homicide–suicide and other forms of co-occurring aggression against self and against others. Prof. Psychol. Res. Pr. 2001;32:626–635. [Google Scholar]
  23. Huizinga D, Elliott DS. Reassessing the reliability and validity of self-report delinquency measures. J. Quant. Criminol. 1986;2:293–327. [Google Scholar]
  24. Jennings WG, Piquero AR, Reingle JM. On the overlap between victimization and offending: A review of the literature. Aggress. Violent Beh. 2012;17:16–26. [Google Scholar]
  25. Johnson KL, Desmarais SL, Grimm KJ, Tueller SJ, Swartz MS, Van Dorn RA. Proximal risk factors for short-term community violence among adults with mental illnesses. Psychiatr. Serv. 2016 doi: 10.1176/appi.ps.201500259. Advance online publication. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Johnson KL, Desmarais SL, Van Dorn RA, Grimm KJ. A typology of community violence perpetration and victimization among adults with mental illnesses. J. Interpers. Violence. 2015;30:522–540. doi: 10.1177/0886260514535102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kay SR, Flszbein A, Opfer LA. The Positive and Negative Syndrome Scale (PANSS) for schizophrenia. Schizophrenia Bull. 1987;13:261–276. doi: 10.1093/schbul/13.2.261. [DOI] [PubMed] [Google Scholar]
  28. Khalifeh H, Johnson S, Howard LM, Borschmann R, Osborn D, Dean K, Hart J, Hogg P, Moran P. Violent and non-violent crime against adults with severe mental illness. Br. J. Psychiatry. 2015;206:275–282. doi: 10.1192/bjp.bp.114.147843. [DOI] [PubMed] [Google Scholar]
  29. Lieberman JA, Stroup TS, McEvoy JP, Swartz MS, Rosenheck RA, Perkins DO, Hsiao JK. Effectiveness of antipsychotic drugs in patients with chronic schizophrenia. N. Engl. J. Med. 2005;353:1209–1223. doi: 10.1056/NEJMoa051688. [DOI] [PubMed] [Google Scholar]
  30. Lindenmayer J, Liu-Seifert H, Kulkami PM, Kinon BJ, Stauffer V, Edwards SE, Chen L, Adams DH, Ascher-Svanum H, Buckley PF. Medication nonadherence and treatment outcome in patients with schizophrenia or schizoaffective disorder with suboptimal prior response. J. Clin. Psychiatr. 2009;70:990–996. doi: 10.4088/JCP.08m04221. [DOI] [PubMed] [Google Scholar]
  31. Mayfield D, McLeod G, Hall P. The CAGE questionnaire: Validation of a new alcoholism instrument. Am. J. Psychiatry. 1974;131:1121–1123. doi: 10.1176/ajp.131.10.1121. [DOI] [PubMed] [Google Scholar]
  32. McNiel DE, Gregory AL, Lam JN, Binder RL, Sullivan GR. Utility of decision support tools for assessing acute risk of violence. J. Consult. Clin. Psychol. 2003;71:945–953. doi: 10.1037/0022-006X.71.5.945. [DOI] [PubMed] [Google Scholar]
  33. Michie C, Cooke DJ. The structure of violent behavior: A hierarchical model. Crim. Just. Beh. 2006;33:706–737. [Google Scholar]
  34. Monahan J, Redlich A, Swanson JW, Robbins PC, Appelbaum PS, Petrila J, McNiel DE. Use of leverage to improve adherence to psychiatric treatment in the community. Psychiatr. Serv. 2005;56:37–44. doi: 10.1176/appi.ps.56.1.37. [DOI] [PubMed] [Google Scholar]
  35. Muthén LK, Muthén BO. Mplus User’s Guide. Los Angeles, CA: Muthén and Muthén; 1998–2014. [Google Scholar]
  36. Overall JE, Gorham DR. The Brief Psychiatric Rating Scale. Psychol. Rep. 1962;10:799–812. [Google Scholar]
  37. Piquero AR, MacDonald J, Dobrin A, Daigle LE, Cullen FT. Self-control, violent offending, and homicide victimization: Assessing the general theory of crime. J. Quant. Crim. 2005;21:55–71. [Google Scholar]
  38. Pratt TC, Turanovic JJ, Fox KA, Wright KA. Self-control and victimization: A meta-analysis. Crim. 2014;52:87–116. [Google Scholar]
  39. Roaldset JO, Bjørkly S. Comparison of patients who were violent, victimized and violent–victimized during the first year after discharge from emergency psychiatry. Psychiatr. Res. 2015;230:978–981. doi: 10.1016/j.psychres.2015.11.019. [DOI] [PubMed] [Google Scholar]
  40. Sadeh N, Binder RL, McNiel DE. Recent victimization increases risk for violence in justice-involved persons with mental illness. Law Hum. Behav. 2014;38:119–125. doi: 10.1037/lhb0000043. [DOI] [PubMed] [Google Scholar]
  41. Silver E. Mental disorder and violent victimization: The mediating role of involvement in conflicted social relationships. Crim. 2002;40:191–212. [Google Scholar]
  42. Silver E, Piquero AR, Jennings WG, Piquero N, Lieber M. Assessing the violent offending and violent victimization overlap among discharged psychiatric patients. Law Hum. Behav. 2011;35:49–59. doi: 10.1007/s10979-009-9206-8. [DOI] [PubMed] [Google Scholar]
  43. Steadman HJ, Mulvey EP, Monahan J, Robbins PC, Appelbaum PS, Grisso T, Silver E. Violence by people discharged from acute psychiatric inpatient facilities and by others in the same neighborhoods. Arch. Gen. Psychiatry. 1998;55:393–401. doi: 10.1001/archpsyc.55.5.393. [DOI] [PubMed] [Google Scholar]
  44. Straus MA, Hamby SL, Boney-McCoy S, Sugarman DB. The Revised Conflict Tactics Scales (CTS2): Development and preliminary psychometric data. J. Fam. Issues. 1996;17:283–316. [Google Scholar]
  45. Swanson JW, Holzer CE, Ganju VK, Jono RT. Violence and psychiatric disorders in the community: Evidence from the Epidemiologic Catchment Area Surveys. Hosp. Community Psychiatry. 1990;41:761–770. doi: 10.1176/ps.41.7.761. [DOI] [PubMed] [Google Scholar]
  46. Swanson AJ, Pantalon MV, Cohen KR. Motivational interviewing and treatment adherence among psychiatric and dually diagnosed patients. J. Nerv. Ment. Dis. 1999;187:630–635. doi: 10.1097/00005053-199910000-00007. [DOI] [PubMed] [Google Scholar]
  47. Swanson JW, Swartz MS, Elbogen EB. Effectiveness of atypical antipsychotic medications in reducing violent behavior among persons with schizophrenia in community-based treatment. Schizophrenia Bull. 2004;30:3–20. doi: 10.1093/oxfordjournals.schbul.a007065. [DOI] [PubMed] [Google Scholar]
  48. Swanson JW, Swartz MS, Elbogen EB, Van Dorn RA, Ferron J, Wagner HR, Kim M. Facilitated psychiatric advance directives: A randomized trial of an intervention to foster advance treatment planning among persons with severe mental illness. Am. J. Psychiatry. 2006;163:1943–1951. doi: 10.1176/appi.ajp.163.11.1943. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Swanson JW, Swartz MS, Essock SM, Osher FC, Wagner HR, Goodman LA, Meador KG. The social–environmental context of violent behavior in persons treated for severe mental illness. Am. J. Public Health. 2002;92:1523–1531. doi: 10.2105/ajph.92.9.1523. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Teasdale B. Mental disorder and violent victimization. Crim. Just. Beh. 2009;36:513–535. [Google Scholar]
  51. Teplin LA, McClelland GM, Abram KM, Weiner DA. Crime victimization in adults with severe mental illness: Comparison with the National Crime Victimization Survey. Arch. Gen. Psychiatry. 2005;62:911–921. doi: 10.1001/archpsyc.62.8.911. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Van Dorn RA, Desmarais SL, Grimm K, Tueller S, Johnson KL, Sellers B, Swartz M. The latent structure of psychiatric symptoms across mental disorders as measured with the PANSS and BPRS-18. Psychiatr. Res. 2016;245:83–90. doi: 10.1016/j.psychres.2016.08.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Van Dorn RA, Kosterman R, Williams JH, Chandler K, Young MS, Catalano RF, Hawkins JD. The relationship between outpatient mental health treatment and subsequent mental health symptoms and disorders in young adults. Adm. Policy Ment. Health. 2010;37:484–496. doi: 10.1007/s10488-010-0291-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Van Dorn RA, Volavka J, Johnson N. Mental disorder and violence: Is there a relationship beyond substance use? Soc. Psychiatry Psychiatr. Epidemiol. 2012;47:487–503. doi: 10.1007/s00127-011-0356-x. [DOI] [PubMed] [Google Scholar]
  55. Volavka J, Czobor P, Citrome L, Van Dorn RA. Effectiveness of antipsychotic drugs against hostility in patients with schizophrenia in the Clinical Antipsychotic Trials of Intervention Effectiveness study. CNS Spectr. 2014;19:374–381. doi: 10.1017/S1092852913000849. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Volavka J, Van Dorn RA, Citrome L, Kahn RS, Fleischhacker WW, Czobor P. Hostility in schizophrenia: An integrated analysis of the combined Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) and the European First Episode Schizophrenia Trial (EUFEST) studies. Eur. Psychiatry. 2016;31:13–19. doi: 10.1016/j.eurpsy.2015.10.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Webster CD, Martin ML, Brink J, Nicholls TL, Desmarais SL. Manual for the Short-Term Assessment of Risk and Treatability (START) (Version 1.1) Coquitlam, Canada: British Columbia Mental Health & Addiction Services; 2009. [Google Scholar]
  58. Witt K, Hawton K, Fazel S. The relationship between suicide and violence in schizophrenia: Analysis of the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) dataset. Schizophrenia Res. 2014;154:61–67. doi: 10.1016/j.schres.2014.02.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Witt K, Van Dorn RA, Fazel S. Risk factors for violence in psychosis: Systematic review and meta-regression analyses of 110 studies. PLoS One. 2013 doi: 10.1371/journal.pone.0055942. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

NIHMS821763-supplement.docx (83.2KB, docx)

RESOURCES