Predictors of psychosocial functioning in people diagnosed with schizophrenia spectrum disorders that committed violent offences and in those that did not: Results of the Recoviwel study

,


Background
Schizophrenia Spectrum Disorders (SSD) and violent behavior share a complex association which has recently been the object of increasing scientific interest for its clinical relevance as well as for its social and public health repercussions (Buchanan et al., 2019;Caruso et al., 2021;Cho et al., 2019;Senior et al., 2020;Whiting et al., 2021).
People living with SSD are more often victims of violence rather than offenders: violent victimization of people with SSD is a frequent phenomenon that negatively affects both the quality of life and the trajectory of the illness and that is also associated with increased internalized stigma and worse rehabilitation and treatment outcomes (Barlati et al., 2021;Fazel and Sariaslan, 2021;Latalova et al., 2014;Maniglio, 2009;Rossa-Roccor et al., 2020).However, while SSD diagnosis is associated with an increased risk of becoming victims of violence, it is also associated with an increased risk of engaging in violent behavior: a recent meta-analysis including 24 studies from 15 different countries reported that individuals living with SSD show a 5.4 increased odd ratio of violence perpetration, compared to community controls (Whiting et al., 2022).Individuals diagnosed with SSD, according to meta-analytic studies, show a 6-10 % prevalence of violent crime perpetration leading to arrest or conviction (Fazel et al., 2009;Stevens et al., 2015); however, the absolute risk rate of committing violent offences is substantially superior in male subjects and increases with the duration of follow-up observations, with studies with a followup observation of 35 years showing an absolute risk of 24.7 % in men (Eriksson et al., 2011;Short et al., 2013).Beside male gender, substance use, parental violence and substance use, victimization, previous selfharm and suicide attempts, homelessness and low socio-economic status all represent well-documented risk factors for violent behavior (Bo et al., 2011;Fazel et al., 2017;Fleischman et al., 2014;Jeyagurunathan et al., 2022).Lack of insight and treatment adherence, severe positive symptoms such as hallucinations and persecutory delusions, hostile behavior and impulsivity as well as cognitive impairment represent risk factors that are specific for people diagnosed with SSD and that can be modified with appropriate pharmacological treatment and psychosocial interventions (Barlati et al., 2022;Coid et al., 2016;Reinharth et al., 2014;Witt et al., 2013).
Psychosocial functioning is a broad term describing the person's ability to perform daily tasks and to interact with others and with society in a satisfying manner: it encompasses different aspects and domains such as social functioning, occupational, school and role functioning, and physical functioning (Lam et al., 2011;Ro and Clark, 2009).It is closely related to real-world functional outcomes such as establishing meaningful interpersonal relationships or finding and maintaining a professional occupation.It represents an important determinant of overall health in both forensic and non-forensic populations: as an example, building better relationships or professional positions may act as a deterrent to offending, but may also represent a motivator improving treatment engagement in both populations (Barendregt et al., 2018;Galinari and Bazon, 2021).
Psychosocial functioning is often severely impaired in people living with SSD (Charlson et al., 2018;Galderisi et al., 2014;Harvey and Strassnig, 2012): it currently represents one of the main targets in the global treatment of SSD and a rehabilitation outcome of primary importance (Fleischhacker et al., 2014;Maj et al., 2021;Mucci et al., 2021;Vita et al., 2023b).It also represents an element of great importance in the patients' perspective which should be closely taken into account in the context of shared decision-making and also to globally improve treatment adherence (Beitinger et al., 2014;Fiorillo et al., 2020;McIntyre et al., 2022).
Several factors negatively affect psychosocial functioning in people living with SSD, including positive and negative symptoms severity and internalized stigma.In particular, neurocognitive and social cognitive deficits represent some of the strongest predictors of functional impairment (Galderisi et al., 2014(Galderisi et al., , 2018;;Harvey and Strassnig, 2012;Maj et al., 2021).Recent evidence suggest that people living with mental disorders that also committed criminal offences may show a worse functional profile compared to individuals belonging to the same population but that did not commit offences (Buchanan et al., 2021), but more research is currently needed regarding psychosocial functioning in these population.In particular, factors related to psychosocial functioning remain to be explored in people living with SSD that committed violent offences and compared to those observed in people with SSD living in the community outside of the forensic context.This could provide valuable insight to devise personalized treatment programs and more specific and targeted interventions to improve the lives of individuals diagnosed with SSD that committed violent offences and of those that did not.For instance, if cognitive performance appeared to be more strongly related to psychosocial functioning in one group than in the other, individuals belonging to that group may represent ideal targets for cognitive-oriented interventions (Deste et al., 2023;Lejeune et al., 2021;Vita et al., 2022).Likewise, if psychosis symptoms appeared to be more important in one of the groups, cognitive behavior therapy may represent a more valid approach in that context (Berendsen et al., 2024;Bighelli et al., 2018;Keepers et al., 2020).The same holds true for clinical factors that are not directly related to SSD, such as substance use, which may benefit form multidisciplinary targeted interventions (Alsuhaibani et al., 2021;Hakobyan et al., 2020;Large et al., 2014;Murthy et al., 2019;Wilson et al., 2018).

Aims
Aims of the present study were to separately assess correlates and predictors of psychosocial functioning in individuals diagnosed with SSD that committed criminal offences and in diagnosed individuals that did not commit criminal offences.A wide variety of clinical, cognitive and violence-related parameters will be taken into account.The inclusion of a control group of participants that did not commit violent offences in particular could provide valuable information on the specificity of the observed correlates and predictors.

Sample
Fifty inmates convicted for violent crimes, aged 18-75, with a diagnosis of SSD according to DSM-5 criteria (American Psychiatric Association, 2013) and enrolled in the Residences for the Execution of Security Measures in Castiglione delle Stiviere, Mantova, Italy were recruited for the present study.
Subjects were excluded from the study if they had a diagnosis of substance use disorder, a main diagnosis of intellectual disability or presence of severe or neurodegenerative organic pathologies.
Participants that committed violent offences were included in the groups defined Offenders Participants (OP).
Controls diagnosed with SSD that did not commit violent offences were recruited form patients accessing Mental Health centers of the Spedali Civili Hospital in Brescia, Italy, with the same inclusion and exclusion criteria They were matched on a 1:1 basis with participants in the OP group for age, gender, education, and diagnosis and were included in the group defined Non-OP.
Recruitment of both groups took place from June 2018 to June 2020.
Participants were informed about the study and were invited through a written and signed consent form.The study was carried out in accordance with the Code of Ethics of the World Medical Association and the Declaration of Helsinki.The protocol was approved by the Ethical Committee of Brescia (Project Identification Code NP3060).All precautions were taken for the management of sensitive data, and participants were not given monetary compensation for their involvement in the study.
More information regarding the recruitment and assessment of participants is provided in a previous publication belonging to the present study (Barlati et al., 2022).

Measures
All participants were assessed with the Brief Assessment of Cognition in Schizophrenia (BACS) (Keefe and Harvey, 2012), the Trail Making Test part A (TMT-A) and part B (TMT -B) (Reitan, 1958) and the Stroop Color-Word test (STROOP) (Stroop, 1938), in order to provide a comprehensive evaluation of neurocognitive performance in different cognitive domains (Fusar-Poli et al., 2012;Vita et al., 2021).Raw scores were corrected for age and education and equivalent scores were taken into account.
Global symptoms severity was assessed with the Clinical Global Impression Severity Scale (CGI-S) (Guy, 1976) and psychotic symptoms were assessed with the Excited Component of the Positive and Negative Syndrome Scale (PANSS-EC) (Kay et al., 1987;Montoya et al., 2011).
Impulsivity was assessed with the Barratt Impulsiveness Scale (BIS-11) (Patton et al., 1995) and the Iowa Gambling Task (IGT) (Bechara et al., 1994); aggressiveness was assessed with the Modified Overt Aggression Scale (MOAS) (Kay et al., 1988); risk of violent relapse was assessed with the Historical Clinical and Risk Management scale 20 (HCR-20) (Douglas and Webster, 1999) and the level of psychopathy with the Hare Psychopathy Checklist (PCL-R) (Hare, 2003;Vitale et al., 2002).All subscales of the HCR-20 were included in the analyses.The PCL-R was included in the analyses divided into two different factors which were identified in a factor analysis conducted by the original scale Authors (Factor 1 or "Callous" and Factor 2 or "Unstable") (Barlati et al., 2022;Hare, 2020).
Psychosocial functioning was assessed using the Personal and Social Performance (PSP) scale (Morosini et al., 2000).The PSP is a singleitem, validated (Nafees et al., 2012;Nasrallah et al., 2008), interviewbased scale assessing functioning during the last month in four areas: personal and social relationships, socially useful activities, self-care and disturbing and aggressive behavior, each one with six degrees of severity.The total score ranges from 0 to 100, with higher scores representing better functioning.
Specifically trained physicians that were not involved in the routine care of participants conducted all the assessments.

Statistical analyses
Non-parametric analyses were adopted in univariate analyses as a non-normal distribution was observed for all parameters (Shapiro-Wilk test p < 0.001).
Participants in the OP and Non-OP groups were compared regarding socio-demographic characteristics, symptoms severity and psychosocial functioning using the Mann-Whitney U test for continuous variables and χ 2 test (or Fisher exact test where appropriate) for categorical variables.
Potential predictors of psychosocial functioning measured with the PSP total score were separately identified in the OP and in the Non-OP group among the different socio-demographic, clinical, cognitive and violence-related variables with the Spearman's Rho test for continuous variables and with the Mann-Whitney U test for categorical variables.
For each group, a stepwise multivariate linear regression model was designed using the PSP total score as dependent variable and all variables emerging as significant in the selection analyses were included as potential predictors.The number of potential predictors introduced in each model was considered appropriate for the number of participants included in the study (Austin and Steyerberg, 2015;Schmidt, 1971).Multicollinearity between individual predictors was considered significant for values of Variance Inflation Factor (VIF) above 4.0 (Shrestha, 2020).

Results
The study included a total of 100 individuals diagnosed with SSD: 50 in the OP group and 50 age-, gender-and education-matched controls in the Non-OP group.Socio-demographic, clinical and functional characteristics of both groups as well as between-groups comparisons are reported in Table 1.
The two groups showed no significant difference regarding sociodemographic and clinical characteristics, with the exception of the number of school failures (U = 823.0,p = 0.002), global clinical severity (U = 872.5,p = 0.006) and excitation symptoms (U = 605.5,p < 0.001).No significant between-group difference was observed in overall psychosocial functioning levels as measured by the PSP scores (U = 1054.0,p = 0.175).
More details regarding the results of variable selection analyses are reported in Table 2 and Table 3 for continuous and categorical variables, respectively.
All variables emerging as significant in the selection analyses were included as potential predictors in the multivariate regression analyses.
A higher risk of violent relapse as measured by HCR-20 clinical subscale scores (Beta = − 0.414, p < 0.002) and greater global clinical severity as measured by CGI-S scores (Beta = − 0.301, p = 0.023) emerged as individual predictors of worse psychosocial functioning in the OP group (Adjusted R 2 = 0.341, p < 0.001).
Greater global clinical severity (Beta = − 0.619, p < 0.001), worse performance in the processing speed domain as measured by the BACS Symbol Coding (Beta = 0.248, p = 0.002) and TMT-A tests (Beta = 0.174, p = 0.016) and higher levels of non-planning impulsivity as measured by BIS-11 scores (Beta = − 0.179, p < 0.001) emerged as individual predictors of worse psychosocial functioning in the Non-OP group (Adjusted R 2 = 0.773, p < 0.001).
No significant collinearity issues between predictors were observed.More details regarding multivariate regression analyses for participants in the OP and Non-OP groups are reported in Table 4 and Table 5, respectively.
Results of supplementary correlation and regression analyses conducted on the whole sample are reported in Supplementary Materials.

Discussion
The aims of the present study were to assess individual predictors of psychosocial functioning in people living with SSD separately in a group of participants that committed violent offences and in a group of matched subjects that did not commit violent offences.According to the results of the multivariate regression analyses, some factors emerged as predictors in both groups, while others where specific for each subpopulation included in the study.
A higher risk of violent relapse, particularly related to clinical factors, and an overall greater global clinical severity emerged as predictors of worse psychosocial functioning in participants that committed violent offences.Greater global clinical severity emerged as a predictor of worse psychosocial functioning also in the control group; alongside this factor, which is shared with the OP group, worse cognitive performance, particularly in the processing speed domain, and higher levels of nonplanning impulsivity emerged as predictors of worse functional outcomes exclusively in the Non-OP group.
Of note, a correlation between worse cognitive performance and worse psychosocial functioning was also observed in the Non-OP group for most cognitive test, encompassing several different cognitive domains such as verbal memory, working memory and executive functions.Higher scores on the PCL-R "Callous" were also correlated with better psychosocial function only in the Non-OP group.Considering that these subjects did not show significant aggressive or violent behavior, G. Nibbio et al. this item could, rather than entirely reflect traits of psychopathy, represent a marker of better awareness and planning abilities: in fact, higher scores on the PCL-R "Callous" factor emerged as correlated with better cognitive performance in the sample (Barlati et al., 2022).
On the contrary, a correlation between higher levels of aggressiveness and worse functioning was observed only in the OP group.
Higher doses of antipsychotic medications emerged as negatively correlated with psychosocial functioning in both the OP and the Non-OP group.One possible interpretation of this observation is that higher doses of antipsychotic treatment may carry a significant anticholinergic burden, with a negative impact on cognitive performance (Ang et al., 2017;Eum et al., 2021;Joshi et al., 2021) and also on daily functioning (Kim et al., 2019).However, higher doses of antipsychotic treatment might also represent an indirect proxy of an overall more severe clinical condition: in the context of the present study, this interpretation appears to be more plausible, as clinical severity was correlated with worse psychosocial functioning in both investigated groups.Furthermore, clinical severity emerged as an individual predictor of worse psychosocial functioning in both regression models, while antipsychotic treatment was discarded in both cases.
These results suggest that greater clinical severity clearly impact psychosocial functioning in all patients living with SSD: this represents an expected result (Harvey and Strassnig, 2012;Maj et al., 2021;Mucci et al., 2021), and treating core symptoms represents one of the cornerstones of SSD treatment and clinical management (Bighelli et al., 2018;Keepers et al., 2020;Solmi et al., 2023).However, some relevant differences in the factors related to functional outcomes in participants that committed violent offences and in participants that did not commit them could also be observed.
The well-documented role of cognitive performance as one of the most important predictors of psychosocial functioning in people with SSD (Deste et al., 2020;Galderisi et al., 2014Galderisi et al., , 2018)), for instance, is clearly confirmed in the Non-OP group.In this regard, implementing effective treatment to improve cognitive performance such as optimizing pharmacological therapy and providing effective psychosocial interventions (Baldez et al., 2021;Barlati et al., 2024;Joshi et al., 2021) could produce clear benefits on real-world functional outcomes, particularly as the vast majority of people living with SSD that receives treatment in mental health services is represent by individuals that are not violent subjects (Fazel and Sariaslan, 2021;Whiting et al., 2021Whiting et al., , 2022)).
According to recent literature, cognitive deficits could have a relevant role in determining a worse clinical and functional condition also in patients who committed criminal offences (Barlati et al., 2022;Darmedru et al., 2017;Iozzino et al., 2021;Mervis et al., 2022;Reinharth et al., 2014;Rund, 2018).However, the results of the present study suggest that other elements may have an even greater impact on functional outcomes in this population: the risk of violent relapse and, by extension, the overall level of aggressiveness of the patient represent essential treatment targets not only in the perspective of reducing the risk of violent behavior (Bo et al., 2011;Fazel et al., 2017Fazel et al., , 2018;;Jeyagurunathan et al., 2022), but also to improve real-world functioning and to improve the odds of recovery.Different psychosocial interventions have been shown to be effective in preventing violent behavior in non-clinical samples, with physical exercise-based programs representing the intervention providing the largest benefits (Fazel et al., 2023).Considering that physical exercise is a well-recognized evidence-based intervention for SSD (Firth et al., 2015;Stubbs et al., 2018) that has also been shown to reliably provide improvements in personal and social functioning (Korman et al., 2023), it could represent a very valuable treatment in forensic settings and for people living with SSD with a history of violent offence more in general.In fact, effective rehabilitation programs for people living with SSD should be structured and personalized in day-today clinical practice not only to promote a stable remission of clinical symptoms, but also and more importantly to allow functional improvement and to promote real-world functional and personal recovery (Giordano et al., 2022;Maj et al., 2021;Vita and Barlati, 2018).
The present study has some points of strength.
The inclusion of both a group of individuals diagnosed with SSD that committed violent offences and of a control group that were separately analyzed allowed to better understand the specificity or the general nature of correlates and predictors of psychosocial functioning of participants.
The study was conducted in a real-world forensic setting and a dayto-say routine clinical setting, so both groups included in the explored sample can be considered as quite representative of the investigated populations, and the results of the study have good generalizability.
The assessment of participants included several well-validated and well-known instruments and was performed by trained raters that were independent from the usual process of care of included subjects: this further contributes to the solidity of the results.
However, some limitations have to be taken into account.One of the main limitations is represented by the small sample of the study, which did not allow to further identify and categorize the participants and to perform further comparisons.
Another limitation is the lack of dedicated measures of positive and negative symptoms severity, which could provide valuable insight into specific functional determinants of people living with SSD and of a more comprehensive functional assessment including separate domains of real-world psychosocial functioning.The complete form of the PANSS, which could also have provided more detailed information on SSD symptoms severity was also not included among the assessment measures of the present study, as also reported and discussed in a previous paper on the same study sample (Barlati et al., 2022).However, implementing a more complex and time-consuming assessment of clinical and functional parameters would have required dedicated and specific training, and would have not been representative of the observed setting.
Assessments of inter-rater reliability that were specific for the present study sample were not performed and could therefore not be included in the analyses; this represents a further limitation of the present work.
The two groups of participants were matched on gender, age and education and were similar on most other parameters, showing significant differences only regarding the number of school failures and symptoms severity.These differences may be intrinsic and related to the different settings of care, and do not present a significant impact on other results of the study, as the two samples are analyzed separately in all instances.Moreover, clinical severity emerged as a significant predictor of psychosocial functioning in both groups.However, this issue should be taken into account while considering the study results and in future research.
Finally, the cross-sectional design of the study does not allow to investigate the direction of the observed correlations and predictor effects: in the context of the presents study, the term "predictor" has to be intended more as a statistical term than as a factual one, and the relationship between the different explored factors and psychosocial functioning could well be bidirectional.In this regard, more longitudinal studies are required in violent offenders with SSD and, more broadly, in people living with SDD in forensic settings.The relationship between violent behavior and psychosocial functioning in people living with SSD could also be explored in mediation and moderation analyses: these analyses could provide further insight on this topic and represent an interesting perspective for future studies.
In conclusion, the results of the present study confirm that overall clinical severity affects psychosocial functioning in all individuals diagnosed with SSD.They also suggest that while cognitive impairment clearly represents a determinant of worse functional outcomes in most patients, the risk of violent relapse is a specific predictor of worse psychosocial functioning in people with SSD that committed violent offences.
In this perspective, dedicating increased attention to functional determinants that emerge in specific populations could be of use into routine mental health practice, particularly in the process of care of individuals that already require dedicated rehabilitation resources.Future research should focus on assessing the effectiveness of dedicated interventions and personalized treatment programs in improving functional outcomes of people living with SSD that committed violent offences and in increasing their rates of recovery.G. Nibbio et al.

Table 1
Comparison for socio-demographic characteristics, symptoms severity and psychosocial functioning.

Table 2
Variables correlated with psychosocial functioning -continuous variables.

Table 3
Variables correlated with psychosocial functioning -categorical variables.

Table 4
Predictors of psychosocial functioning in participants that committed violent offences.
CGI-S: Clinical Global Impression-Severity Scale; HCR-20: Historical Clinical and Risk Management scale 20; PSP: Personal and Social Performance Scale.

Table 5
Predictors of psychosocial functioning in participants that did not commit violent offences.: Brief Assessment of Cognition in Schizophrenia; BIS-11: Barratt Impulsiveness Scale version 11; CGI-S: Clinical Global Impression-Severity Scale; TMT: Trail Making Test. BACS