Two major aims of forensic psychiatry and psychology are (1) to assess and treat mental health issues in people at risk of criminal behavior and (2) to identify risk and protective factors that increase or reduce the risk of further delinquency. These aims become of specific importance when working with criminal adolescents or young adults because effective intervention may increase their chances to develop into a healthy adulthood desisting from future crime.

Mental health issues are common among young offenders. International studies have reported rates of psychiatric problems of up to 93% among young offenders, including externalizing–i.e., conduct disorder, oppositional defiant disorder, attention deficit hyperactivity disorder (ADHD)–but also internalizing (i.e., mood disorders, anxiety) problems, substance use disorders, and personality disorders (1–12). Rates differed with regard to the setting young people were assessed in, e.g., community-based treatment settings vs. incarceration, with the latter usually showing higher frequencies of mental health issues. However, it has been criticized that it often remains unclear to what extent mental health problems have existed before and thus may have led to criminal behavior, and to what extent placement circumstances may have influenced mental health [e.g., (1)].

There is a vast amount of research that has examined the predictive value of mental health problems in young offender samples with respect to future crime. In a recent meta-analysis including data of 5,737 juveniles, Wibbelink et al. (2) found small to moderate predictive effects of externalizing but not internalizing problems on criminal recidivism. Other studies reported similar results. Higher rates of criminal recidivism were found in young offenders with ADHD (3–5), conduct disorder (6), oppositional defiant disorder (7) as well as substance use disorders (8) and personality disorders [especially DSM Cluster B disorders (10)]. However, findings remain inconsistent across studies due to differences in definitions and assessment of mental health issues (e.g., self-reported vs. clinician-administered), recidivism (e.g., reconviction vs. reincarceration, self-reported vs. officially recorded), and crime concepts (e.g., in terms of severity and type of criminal acts). For example, mental health may relate differently to criminal recidivism when differentiating violent from non-violent crime. Bessler et al. (8), for instance, found that young offenders' mental health problems and substance use disorders in particular, were associated with risk of violent but not general (non-violent) criminal re-offending. Plattner et al. (10) concluded that substance use disorders were predictive of future non-violent, drug-related crime, but problematic alcohol use in particular was associated with violent criminal recidivism. Conversely, Mulder et al. (11) found a negative predictive relationship of psychopathology on violent reoffending. In addition, conclusive empirical evidence is lacking due to different follow-up periods among studies [e.g., adolescence vs. adulthood (11)].

However, considering differences in follow-up periods is of specific importance as delinquency has been claimed to be a common phenomenon during adolescence indicating that young people with repeated crime only during this developmental period may not be as burdened by mental health problems as those who continue criminal behaviors until adulthood (2, 12). A well-known perspective on young peoples' courses of delinquency is Moffit's developmental taxonomy on adolescent-limited and life-course-persistent antisocial behavior (13): According to this theory, most juveniles may engage in (non-pathological) antisocial behavior during “a contemporary maturity gap” (p. 674) but desist from crime after this period, whereas a smaller proportion showing early conduct problems and higher psychosocial burden continues repeated and more severe (pathological) criminal behaviors beyond adolescence. Moffitt (14) recently described evidence from 25 years of research on this taxonomy and emphasized that more research is needed, e.g., concerning associations of delinquent pathways with mental health.

Considering the developmental courses of delinquency as well as the potential individual and societal consequences, it appears of major importance to identify those young offenders who are at risk of serious, violent, and chronic (SVC) offending. SVC offenders have been suggested to be a rather small group but “responsible for a disproportionate amount of serious crime” (15). Following a cohort of more than 27.000 individuals over 16 years, Kempf-Leonard et al. (16) found that among young offenders with serious, violent, and chronic delinquency, those who had shown a combination of these three crime characteristics had the highest rates of adult crime perpetration. Baglivio et al. (17) examined the prevalence as well as risk and protective factors of SVC offending among more than 363.000 juveniles referred to the juvenile justice system in the US over a 5-year period. They reported a proportion of SVC offenders of 8.9%, with SVC status defined as having shown a history of four or more official referrals with at least one felony offense against a person or a weapon/firearm charge. Compared to non-SVC offenders, SVC offenders were younger at first referral and had more risk but less protective factors regarding criminal recidivism after 1 year follow-up. Although SVC offenders showed higher scores on history of mental health problems, current mental health did not differ between SVC and non-SVC offenders. Current substance use predicted future SVC rearrest. Among more than 64.000 young delinquents, Perez et al. (18) stated a proportion of 16.66% SVC offenders (defined as having committed three or more serious felony offenses with at least one violent offense) and found that a predictive effect of adverse childhood experiences (ACEs) on SVC offending was partially mediated by maladaptive personality traits (e.g., impulsivity) and adolescent problem behavior, including substance use and mental health problems.

In summary, the scientific foundation on the relations of mental health and SVC offending among young delinquents is still scarce. More long-term investigations are needed to shed light into the dynamics of mental health and other risk and protective factors with perpetration of and desistance from SVC offending in order to identify those young people at risk of continuous, severe crime involvement. Further empirical evidence may serve to elaborate adequate treatment and prevention approaches to increase young offenders' chances to develop into a healthy, crime-free adulthood and, thus, also contribute to the protection of society.

Considering abovementioned findings but also limitations of previous research, the present study aimed at examining the course of SVC offending among young detainees up to 15 years after release from incarceration and respective associations with mental health. We hypothesized that compared to SVC desisters, SVC offenders would show higher rates of mental health issues, especially externalizing problems. Baseline SVC status was suggested to be positively associated with future SVC status. We also expected that current externalizing problems, substance use problems and cluster B personality disorders would increase the risk of being a future SVC offender.

Mental health problems are common among young offender samples but their role in predicting criminal recidivism is still not clear. Early identification and treatment of young offenders at risk of SVC offending is of major importance to increase their chances to develop into a healthy and non-criminal future and protect society from further crime. The present study aimed at contributing to and expanding the current knowledge on the dynamics of mental health and SVC offending by examining mental health status and long-term courses of delinquency in a high-risk sample of young detainees.

Consistent with previous research [e.g., (1)], we found a high prevalence of mental health issues in the present sample, especially in terms of externalizing problems. Internalizing behaviors including problems with anxiety and depression as well as rule-breaking behaviors appeared to be higher with shorter incarceration time until assessment, whereas social withdrawal and internalizing problems increased with longer time remaining until release. Although effects were rather small, they might reflect a particular dependency of mental health issues on incarceration time, especially at the beginning when young offenders need to adapt to the circumstances of incarceration, but also when facing rather long-lasting imprisonment. Whereas the initial phase of incarceration may thus be associated with feelings of loneliness, fear and uncertainty on the one hand and rule-breaking, oppositional behavior on the other hand, extended imprisonment may evoke thoughts and feelings of hopelessness and pointlessness in young offenders (37). These findings emphasize the need of an adequate monitoring of young detainees' mental health not only at the beginning but over the course of incarceration, especially in those facing long-term imprisonment.

One third of our sample met the criteria of being a SVC offender until assessment but more than half of the young offenders were identified as SVC offenders after release. Compared to previous studies (17, 18), SVC offending prevalence was rather high, which may be due to the fact that we focused on a high-risk incarcerated sample instead of somewhat broader and more heterogeneous juvenile justice samples. Although about 68% of the young offenders who identified as SVC offenders before assessment also showed SVC offending after release, SVC-status before assessment and after release were not significantly associated. This finding is not consistent with our initial hypothesis but suggests that also young offenders with a history of severe offending may still be able to desist from SVC offending. On the other hand, more than half of the SVC-post offenders had not shown SVC offending before, highlighting the need of effective early identification to reduce young people's risk of engaging in serious, violent, chronic delinquent careers.

Early identification is challenging due to the multifactorial etiology of criminal behavior. In the present study, differences between young people with SVC offending before assessment and those without were only found in terms of their prior criminal involvement, with SVC-pre offenders showing higher rates of previous convictions and incarcerations, which was expected based on the definition of SVC offending. More interestingly, no differences emerged regarding mental health. However, young offenders without SVC offending after release differed from those with SVC-post offending in several ways. First, SVC desisters were less likely to have shown early involvement in crime (i.e., delinquency before the age of criminal responsibility) and held higher academic qualifications. These findings corroborate previous research that pointed to more disadvantageous social conditions in young individuals engaging in continuous and severe criminal conduct (17). Early onset of criminal behavior and low academic achievement may both display indicators for deficient social integration and control early in life, thus highlighting the need to implement adequate support offers, e.g., in terms of youth welfare measures or family-based treatment approaches such as multisystemic therapy (38). Regarding mental health, SVC desisters reported fewer mental health problems in general and especially fewer externalizing behaviors, attention problems, alarming alcohol use, and personality disorders (cluster B personality disorders in particular). Statistically controlling for the influence of age, higher level of school education, less mental health issues as well as absence of alarming alcohol use and absence of cluster-B personality disorders predicted desistance from SVC offending in univariate analyses, although solely alcohol consumption remained a significant predictor in multiple regression. These findings are in line with previous research stating that criminal recidivism in young offenders is associated with mental health issues, substance use problems, and cluster-B personality disorders [e.g., (2, 9, 10)]. Substance use problems were found to be associated with increased risk of violence perpetration and to predict future violent and also SVC offending (8, 17, 39). However, the present results stress that not substance use problems in general may contribute to increased risk of future SVC offending, but alarming alcohol consumption in particular. Similar findings with regard to future violent offending were reported by previous research [e.g., (10)]. Alcohol use problems among young offenders are concerning in different ways. First, research has repeatedly emphasized that alcohol can affect an individual's emotional and behavioral regulation capacity and lower the threshold to engage in aggressive and violent acts. A recent meta-analysis stated a causal relationship between alcohol (but not stimulant drug) intoxication and aggression (40). Parrot and Eckhardt (41) introduced the alcohol-aggression link within I3 [e.g., (42)] and Alcohol Myopia Theory (43). According to the authors, I3 theory stresses that behavior is influenced by instigating, impelling, and inhibitory factors. Aggressive behavior may, thus, be probable when self-regulation is inhibited by the influence of alcohol in case a person is provoked and does show traits or attitudes in favor of aggressive (violent) behavior. Alcohol Myopia Theory highlights distorted attention processes due to alcohol influences with focus on short-term situational goals (e.g., lowering frustration) while neglecting long-term (legal) consequences. Second, alcohol is easily accessible, in Germany even legally as early as at the age of 16 years. The availability of and easy access to alcohol may contribute to the development of problematic alcohol use patterns, especially in those young people who suffer from early psychosocial burden and societal problems. Thus, prevention and intervention approaches addressing alcohol use in young people appear beneficial in order to prevent further dysfunctional outcomes, e.g., in terms of continuous criminal careers [e.g., (44)].

The interpretation of the present results requires the consideration of several strengths and limitations. First, we assessed a multitude of indicators for mental health including internalizing and externalizing problems, personality disorders, and alcohol and other (illegal) substance use problems. We combined both self-rated and clinician-administered measures and relied on well-established instruments. The long-term observation of criminal careers up to 15 years after release from incarceration allowed a more sophisticated insight into pathways of criminal offending beyond adolescence, which is of major importance in light of age-dependent crime prevalence [e.g., (5, 13)]. In the same regard, focus on continuous SVC offending is crucial to identify those young individuals who are in greatest need of prevention and treatment in order to reduce maladaptive personal but also societal consequences. On the other hand, the present sample represented a high-risk sample of young detainees, thus generalization to and implication for somewhat more heterogeneous juvenile offender samples is limited. Moreover, although sample size appeared satisfactory for long-term forensic examination, it was still rather small compared to general population studies. Because our sample size was predetermined by data availability, we did not perform a priori power calculations. However, post-hoc power analyses have been criticized as well (45). Yet, we conducted sensitivity analyses in G^*Power Version 3.1.9.7 (46) that indicated, for example, that group differences between post-SVC offenders and desisters would have required at least an effect of d = 0.55 to be detected with a power of.80. Thus, the limited sample size (and statistical power) available in the present study may bear the risk of leaving some more subtle effects undetected due to statistical insignificance. Similarly, future research may benefit from examining female offenders, too, because gender influences have been discussed both in the dynamics between mental health and criminal recidivism as well as in the field of SVC offending (2, 16, 47). Second, although self-reports of mental health issues have been used in offender samples before, there is a risk of biased estimates due to under- or over-reporting [e.g., (27)]. Besides, there could be a possible bias in the self-rating instruments, as participants could have answered in a socially desirable manner, which is common in different settings, however, especially in offender samples. Likewise, despite the common scientific procedure in forensic psychology and psychiatry research of relying on officially registered crime, not all offenses may come to the attention of law enforcement agencies. Eventually, the consideration of other influencing factors underlying the effects of mental health on criminal behavior was beyond the scope of the present study. For instance, a vast amount of research has focused on ACEs as potential exploratory factors in the context of mental health and adolescent and adult (SVC) criminal behavior (9, 18, 22, 27, 48). More research is needed to broaden the knowledge on the associations between maladaptive developmental factors including mental health and perpetration but also desistance from criminal behavior in order to derive early and effective prevention and treatment approaches aimed at reducing young people's risk of engaging in continuous (SVC) criminal careers and, thus, support their development into a healthy, non-delinquent future.

The datasets presented in this article are not readily available because of the specific confidentiality of the assessed clinical and forensic information. Scientists wishing to use them for non-commercial purposes are kindly asked to contact the present authors in order to frame individual agreements. Requests to access the datasets should be directed to steffen.barra@uks.eu.

The studies involving human participants were reviewed and approved by Ethics Committee of the Medical Chamber of Saarland, Saarbrücken, Germany. Written informed consent to participate in this study was provided by the participants themselves in case they were at least 18 years old at time of assessment or participants' legal guardian when younger.

SB, WR, DT, and MR: conceptualization. SB and WR: methodology and data curation. SB: formal analysis, writing—original draft preparation, and visualization. WR, MR, and PR-J: investigation. WR and MR: resources and project administration. SB, WR, DT, PR-J, and PH: writing—review and editing. All authors have read and agreed to the published version of the manuscript.

We acknowledge support by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) and Saarland University within the Open Access Publication Funding programme.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.