Introduction
People with mental illness are overrepresented throughout the criminal justice system.Reference Fazel and Danesh 1 Mental illness as a concept holds no exact definition, being grounded in psychiatry and representing psychiatric disorders that are considered both medical and social problems, while definitions of crime and violent offenders are offered by criminal justice institutions. Explanations of the two concepts originate from quite opposite fields and come with different goals: psychiatry provides treatments and law provides justice and social rehabilitation.Reference Forrester and Hopkin 2 Additionally, taking into account criminal responsibility in psychiatric offenders’ cases, this is regularly assessed and an individual’s classification as dangerous can be renewed or removed by the courts. If the latter, the person goes back into the community and standard mental health care is provided by community mental health services.Reference Peloso, D’Alema and Fioritti 3
A model of care for psychiatric patients that is partly different from the rest of Europe has been developed in Italy from 1978. Psychiatric hospitals were closed, and greater emphasis was placed on social interventions, supporting the hypothesis that modifying certain environmental factors would reduce relapse.Reference Rossetto, Clerici and Franconi 4 In the last three decades, there were several measures which moved to a deinstitutionalization of psychiatric facilities and designed new methods for the management of psychiatric patients with criminal behaviors. A significant step was made in 2008 with the shift of psychiatric administration from the Ministry of Justice to the Ministry of Health. It followed the total closure of forensic psychiatric hospitals (OPGs) in Italy and the conversion to a care model based on residential units in the community (REMS), fully integrated in public mental health services which were prescribed by law (L. 9/2012). 5 Law 81/2014 also stated that a patient cannot stay in a REMS for a period longer than the prison sentence for the same index offense (L. 81/2014). 6 Similarly, in Croatia and Portugal, psychiatric detention is limited to the amount of time the patients would have spent if they had not been mentally ill and had been given a jail sentence.Reference Völlm, Clarke and Herrando 7 The other EU countries allow the imprisonment of mentally disordered offenders for longer periods of time than the regular sentence.Reference Sampson, Edworthy and Völlm 8
Therefore, when patients end their duty for criminal behaviors, their clinical management moves back to the outpatient psychiatric centers. High rates of psychiatric disorders in correctional facilities have fueled widespread concerns about the “criminalization of mental illness.” However, elevated risks of violent behavior are not equally shared across the spectrum of psychiatric disorders. In the past years, multiple studies in the field of forensic psychiatry confirmed a close relationship between violent offenders and comorbid substance use. There is consistent evidence that, particularly in combination with a comorbid substance use disorder (SUD), mental disorders may contribute to the likelihood of violence and offending for part of the population.Reference Fazel, Gulati and Linsell 9 Recent literature has observed that acute substance use may be influential on behavior by disinhibiting controls, increasing antisocial behaviors and violence and that psychiatric or psychological exacerbation due to SUD, intoxication, withdrawal, and dependence may increase the likelihood of offending behavior.Reference Ogloff, Talevski and Lemphers 10
In 2013, 110 international studies reporting factors associated with violence were analyzed: violence was very strongly associated with a history of polysubstance misuse (OR = 10.3) and strongly associated with a diagnosis of comorbid SUD (OR = 3.1).Reference Witt, van Dorn and Fazel 11 Research also showed that patients who suffer from schizophrenia and concomitant substance use are not only more likely to experience a variety of psychosocial difficulties, such as violence, victimization, incarceration, homelessness, and family difficulties, but also are highly prone to adverse consequences, including poor treatment response, relapse, hospitalization, HIV infection, hepatitis C infection, and suicide.Reference Drake, McHugo and Xie 12
Moreover, patients with a psychiatric diagnosis and no abstinence during the follow-up or with a mental and behavioral disorder due to psychoactive substance use showed more often criminal recidivism than patients without such a disorder.Reference Probst, Bezzel and Hochstadt 13
As regards the Italian forensic psychiatric system, few studies were conducted in Italian centers. One of the variables associated with violent behavior in patients in Italian public and private acute psychiatric inpatient facilities was SUD.Reference Biancosino, Delmonte and Grassi 14 This was also confirmed by another study showing that aggressive patients were more likely to have used substances in the past (43.0% vs. 31.6% in nonaggressive group). This study also reported that hostile and violent patients were more frequently hospitalized in public versus private facilities.Reference Amore, Menchetti and Tonti 15
Furthermore, in a study on female patients who were discharged from a REMS in Castiglione Delle Stiviere (Italy) before 2008 and readmitted in the same place from 2008 to 2018, the readmission was positively associated with the presence of SUD and a primary diagnosis of personality disorder.Reference Rossetto, Clerici and Franconi 4 Furthermore, on a sample including all patients admitted in Volterra (Italy) REMS from 2015 to 2017, the most frequent lifetime comorbid psychiatric diagnosis was substance-related disorder (54.1%). Almost two-thirds of those patients were already followed up by mental health services before committing the crime. In that sample, SUDs were the main psychiatric comorbidity and resulted more frequent in bipolar patients than in other patients.Reference Lombardi, Veltri and Montanelli 16
In the still growing field of forensic psychiatry, we aimed at characterizing a sample of psychiatric patients who were also offenders, focusing on the role of substance use in the social and psychopathological picture.
Methods
Patients with a history of crimes, previous security measures, and/or with ongoing investigation of either gender or any age attending one community mental health center of the Azienda Socio Sanitaria Territoriale Fatebenefratelli Sacco based in Milan, Italy, were recruited. The inclusion criteria were the following: a confirmed psychiatric diagnosis based on the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR, DSM-5, or DSM-5-TR, according to the manual in use at the time of the diagnosis) and an age between 18 and 75 years at the time of recruitment; no exclusion criteria were applied.
All medical records of recruited patients were retrospectively reviewed, anonymized, and held in a secure database according to the local data protection policies. Patients gave their written informed consent to participate in this study and to have their personal, clinical, and demographic data used for research purposes. The present study was conducted according to the principles expressed in the Declaration of Helsinki (PMC2566407).
Outcome measures
Main clinical and sociodemographic variables were collected reviewing patients’ medical records. Sociodemographic variables included sex, presence of a partner, education level, and employment; clinical variables were diagnosis (psychosis, personality disorder, depressive disorder, bipolar disorder, anxiety disorder, obsessive–compulsive disorder, cognitive impairment, cognitive decay, pathological gambling, eating disorder, attention deficit hyperactivity disorder, lifetime and current alcohol use disorder [AUD], lifetime and current SUD, and presence of a psychiatric comorbidity), prescribed drugs (mood stabilizers [valproate, lithium, lamotrigine, gabapentin, and pregabalin], first-generation antipsychotics [FGA], and/or second-generation antipsychotics [SGA]), prescription of long-acting injection (LAI) therapy, number of hospitalizations (sample was stratified by less than 4 or 4 or more hospitalizations), and other medical comorbidities; we collected the following forensic variables: type of committed crimes (economic crime and/or violence against others), presence of past history of crimes, previous security measures, previous incarceration, and if the current measure was a confinement in REMS. The sample was divided into 2 main groups: patients with and without a history of SUD (present and/or lifetime). AUD was considered separately given its high frequency in psychiatric patients with psychosis spectrum syndromeReference Archibald 17 and the underlining different neurocircuits involved compared to other SUDs.Reference Karoly, YorkWilliams and Hutchison 18 Those subgroups were then compared in order to find potential differences.
Statistical analyses
Patients’ sociodemographic, clinical, and forensic characteristics are presented using descriptive statistics (Table 1). χ 2 test for dichotomous variables were performed to compare patients with and without a lifetime substance use (Tables 2–6). Nonparametric Mann–Whitney U test was used for continuous variables, comparing patients with and without lifetime SUD. We used logistic regression to analyze lifetime SUD as an independent variable. Dependent variables analyzed were sex, presence of a partner, education level, and employment. Furthermore, we investigated the following clinical features: diagnosis (psychosis, personality disorder, depressive disorder, bipolar disorder, anxiety disorder, obsessive–compulsive disorder, cognitive impairment, cognitive decline, pathological gambling, eating disorder, attention deficit hyperactivity disorder, lifetime and current AUD, lifetime and current SUD, and presence of a psychiatric comorbidity), prescribed drugs (mood stabilizers [valproate, lithium, lamotrigine, gabapentin, and pregabalin], FGA, and/or SGA), prescription of LAI therapy, number of hospitalizations (sample was stratified by less than 4 or 4 or more hospitalizations), and other medical comorbidities; other variables analyzed were type of committed crimes (economic crime and/or violence against others), presence of past history of crimes, previous security measures, previous incarceration, and if the current measure was a confinement in REMS. A p-value ≤0.05 was considered statistically significant. Statistical analyses were performed using IBM SPSS Statistics V26.0 (IBM Corporation, Armonk, NY, USA).
Descriptive Statistics: Main Sociodemographic Variables
Sociodemographic Variables
* p-value ≤ 0.05.
Clinical Variables
Abbreviation: ADHD, Attention-deficit/hyperactivity disorder.
* p-value ≤ 0.05.
Hospitalizations
* p-value ≤ 0.05.
Prescription-Related Variables
Abbreviations: FGA, first-generation antipsychotics; LAI, long-acting injection; SGA, second-generation antipsychotics.
* p-value ≤ 0.05.
Forensic Variables
Note: For binary variables, p-values were calculated by chi-square test.
Abbreviation: REMS, residenza per l’esecuzione delle misure di sicurezza.
* p-value ≤ 0.05.
Results
Main sociodemographic and clinical data are summarized in Table 1.
A total of 76 subjects were considered for this study. The mean age was 48.7 ± 14.6 years, the mean age of illness onset was 27.9 ± 13.8 years, and the mean illness duration was 20.4 ± 13.3 years. Nine (11.8%) individuals were females, 17.1% had a partner, and 22.4% had a job or were retired. The most represented diagnoses were psychosis (48.7%) and personality disorders (47.4%); 71.1% of subjects had at least one psychiatric comorbidity, 51.3% subjects used substances of abuse throughout life and 31.6% alcohol, 25.0% had AUD when they committed the crime, and 31.6% had current SUD. The most frequent offense was violence against others (60.5%), and the second most represented was economic crime (27.6%). The most prescribed drugs were antipsychotics (85.5%), being 36.9% FGA and 63.1% SGA. More than 20% of the total sample was receiving antipsychotics via LAI formulation.
Fifty-four patients had clear notation regarding the presence of a lifetime SUD. We found statistically significant differences between patients with lifetime SUD versus patients without lifetime SUD (Tables 2–6 and Figure 1) in terms of absence of a partner (87.2% vs. 77.6%; p < 0.05), psychiatric comorbidities (84.6% vs. 60.0%; p = 0.05), presence of LAI prescription (28.9% vs. 0%; p < 0.05), 4 or more hospitalizations (64.1% vs. 33.3%; p < 0.05), charge with economic crimes (40.5% vs. 6,7%; p < 0.05), previous crimes (43.6% vs. 13.3%; p < 0.05), and previous incarceration (35.9% vs. 6.7%; p < 0.05) (Figures 2 and 3). We also found, in the lifetime SUD population, an association with age (p < 0.05) and hospitalizations number (p < 0.05). Mann–Whitney U test confirmed a significant difference on the number of hospitalizations between patients with lifetime SUD versus patients without lifetime SUD (28.4 vs. 18.1; p < 0.05). Logistic regression confirmed that lifetime substance use may be a risk factor for economic crimes (OR = 9.5; p < 0.05). We did not find statistically significant differences between patients with current SUD versus patients without current SUD.
Significant differences between patients with Lifetime SUD vs Absence of Lifetime SUD.
Forensic variables regarding the whole sample.
Percentage of psychiatric comorbidities regarding the whole sample.
Discussion
In the present report, sociodemographic, clinical, and forensic variables were examined with particular emphasis on patients with SUD; thus, we focused on the role that substance use plays in this population.
Most of our sample was represented by male patients, with a mean age of 49 years, partnerless, and unemployed. These results were also found in numerous studies on psychiatric patients with a crime history trying to predict violent offense,Reference Fazel, Wolf and Larsson 19 examining arrest records of 13,816 individuals receiving services from a regional Department of Mental Health,Reference Fisher, Roy-Bujnowski and Grudzinskas 20 describing main characteristics of forensic psychiatric inpatients,Reference Gu, Guo and Zhou 21 –Reference Streb, Lutz and Dudeck 23 analyzing hypothetical correlations between violence and psychosis,Reference Witt, van Dorn and Fazel 11 and between criminality and bipolar disorder.Reference Webb, Lichtenstein and Larsson 24 The only variable in contrast with most of the current literature was the level of education, which turned out to be higher in our sample: 40%, in fact, achieved a diploma or degree versus 9%–13% found in the current literature.Reference Fazel, Wolf and Larsson 19 , Reference Gu, Guo and Zhou 21 , Reference Streb, Lutz and Dudeck 23 A plausible explanation could be that our sample was represented by patients on voluntary treatment on psychiatric service, whereas, in the other studies, the sample consisted of the entire psychiatric population on criminal records.
Consistently with previous findings,Reference Elbogen and Johnson 25 –Reference Pagerols, Valero and Dueñas 27 most of our sample had a psychiatric comorbidity (71.1%). The most represented disorder in our sample was psychotic disorder, which affected almost a half of the sample; specifically, the diagnosis mainly found was schizophrenia, followed by substance induced psychosis. Psychotic disorders are, in fact, the only illnesses still considered as independent risk factors which would increase, albeit slightly, the likelihood of committing crimes, regardless of concomitant substance use.Reference McCabe, Christopher and Druhn 28 –Reference Falconer, El-Hay and Alevras 32 Other psychiatric illnesses found in our sample included bipolar disorder and major depressive disorder; previously, other studies showed that these diagnoses, in absence of concomitant psychotic symptoms, did not play a determinant role in the commission of crimes.Reference García 33 –Reference López, Laviana and Saavedra 35 Considering the nature of psychotic symptoms, impaired decision-making ability and increased sensitivity to the environment and external events could explain this correlation.Reference Leclerc, Regenbogen and Hamilton 36 –Reference Yee, Matheson and Korobanova 38
Almost 50% of the present sample suffered from a personality disorder, mostly borderline personality disorder (BPD), followed by antisocial personality disorder (APD). Both BPD and APD DSM-5 criteria include among clinical features, “intense anger or difficulty controlling anger” and “irritability and aggressiveness, as indicated by repeated physical fights or assaults,” respectively. 39 Those features have also been highlighted by the current literature; some authors focused on the differences between outpatients and the incarcerated ones,Reference Esposito, Ceresa and Auxilia 40 some on the impact of SUD and personality disorders on criminal behavior, 41 , Reference Howard, Hasin and Stohl 42 while others studied lifetime risk and correlates of incarceration excluding substance-related mental illness.Reference Nakic, Stefanovics and Rhee 43 These findings emerge also in larger descriptive literature.Reference Völlm, Edworthy and Huband 44
The absence of a partner was prevalent in the population with lifetime substance use compared to the population of nonusers (p < 0.05). Some authors have studied the role of rejection sensitivity and risk behavior, finding higher rates in patients with a history of substance abuse.Reference Woerner, Kopetz and Lechner 45 Risk behavior and higher rejection sensitivity could be related to greater difficulty in establishing stable relationships.
Borderline statistical significance was found regarding the prevalence of psychiatric comorbidities (p = 0.05) in the lifetime substance users group compared to the nonusers group: this is partly intrinsically related to the definition of the sample, consisting in patients who have experienced several stressful events and have multiple risk factors for psychiatric diseases.Reference Zijlmans, van Duin and Jorink 46
The present study, moreover, highlighted the use of LAI therapy exclusively in the substance user population, compared to nonusers (p < 0.05). The reason may lie in several causes: first of all, these are complex, comorbid, chronic patients who need multispecialty treatment; injectable therapies are likely to be better tolerated by the patient; moreover, recent studies have shown that the earlier an LAI is prescribed, the more quickly the patient is stabilized, reducing the risk of relapse.Reference Hsu, Kao and Lu 47 –Reference Marcus, Zummo and Pettit 49
Substance users also showed a higher frequency of hospitalization compared to nonusers (more than 3 times; p < 0.05). This finding is supported by the present literature, which also identifies past hospitalizations as a risk factor for recurrence.Reference Olivares, Sermon and Hemels 50 –Reference Karlsson and Håkansson 53 Recidivism of dysfunctional or criminal behavior has been related to a difficulty of institutions to best support these individuals.Reference Karlsson and Håkansson 53
Consistently with several other studies,Reference Pagerols, Valero and Dueñas 27 , 41 , Reference Zgoba, Reeves and Tamburello 54 substance users more frequently committed economic crimes (p < 0.05) probably related to the need to obtain the drug of abuse. This also generates a vicious cycle driven by craving: users spend all the money they have to obtain the substance and are willing to commit crimes against property in order to afford more of it, in ever-increasing quantities, as their tolerance requires.
In contrast to other studies,Reference Mundt and Baranyi 52 , Reference Zgoba, Reeves and Tamburello 54 , Reference James and Glaze 55 we did not find a prevalence of drug dealing crimes in substance users. The possible interpretation could be that as stated above, our sample included people voluntarily attending an outpatient facility, while drug dealing is a crime often related to a refusal of institutional health care.Reference Hepburn, Barker and Nguyen 56 , Reference Small, Maher and Lawlor 57
As already stated by other authors,Reference Pagerols, Valero and Dueñas 27 , Reference Mundt and Baranyi 52 , Reference Karlsson and Håkansson 53 , Reference James and Glaze 55 , Reference Baillargeon, Penn and Knight 58 –Reference Smith 60 previous offenses and previous incarceration were found to be significantly higher (p < 0.05) in the substance users’ population. This finding affects the risk of recidivism of psychiatric patients with SUD.Reference Zgoba, Reeves and Tamburello 54
In this study emerges that psychiatric patients with a history of crimes and lifetime SUD need a comprehensive consideration, they could benefit from an approach that further integrates biological, psychological, and social factors; those elements should be seen not only as complementary but as “facets of the same dynamical system.”Reference Gómez-Carrillo, Paquin and Dumas 61 A novel precision approach and the adoption of individualized treatments could break the vicious cycle that leads to recidivism.Reference Wilkinson, Luján and Hales 62
The abovementioned results should be interpreted considering some methodological limitations. First, the cross-sectional nature of the study allowed only a 1-time assessment. Second, variables were obtained retrospectively, being susceptible to recall bias. This was a monocentric study; thus, it may lack external validity. Moreover, the absence of variables such as severity of addiction, abstinence duration, and the relationship between the substance consumed and psychopathological variables could have affected the considerations drawn from the results obtained. Due to the nature of this study, we were not able to retrace some variables like antidepressant prescriptions and the personality disorder cluster diagnosed. Diagnoses have not been explored in dept through psychometric scales, leaving out some patients’ specifics. Finally, sample size should be increased.
Conclusion
The results of the present research highlight that individuals with a history of crime records and lifetime SUD tend to have higher rates of hospitalization, a higher amount of criminal reoffending, and an increased number of incarcerations, particularly for economic crimes.
The complexity of those patients lies not only in medical and psychiatric reasons but also represents a social and economic challenge for the whole community. As future perspectives, a bigger sample could be examined, possibly involving other community mental health centers; it could be helpful to evaluate variations in time and in terms of age ranges, adopting a longitudinal approach.
Author contribution
Conceptualization: B.M.D., B.B., D.G., E.P., F.A.; Writing – review & editing: B.M.D., B.B.; Data curation: A.F., D.G., E.P., L.M., M.C., M.L., F.A.; Formal analysis: B.B., F.A.; Writing – original draft: E.L., M.C., M.L., S.L., F.A.; Investigation: S.L., F.A.
Financial support
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for profit sectors.
Disclosure
B.M.D.O. has received lecture honoraria from Angelini, Janssen, Lundbeck, Livanova, Arcapharma, and Neuraxpharm. The other co-authors do not have any disclosures.
Open access
