Introduction

The insanity defense is a legal construct that, in specific circumstances, exempts defendants affected by mental illness from being held accountable for their criminal behavior [1]. It is a defense that dates to ancient times, and some traces can be found in ancient Greek and Roman texts [2]. The Napoleonic code was the main point of reference for the whole continental European codification of the nineteenth century. The legal nature of the criteria underlying insanity has sources that are also historical and that in Italy has led to giving greater weight to the epistemic and inhibitory control components. The most acknowledged standards in the Anglo-American systems are the M’Naghten RuleFootnote 1, and The Model Penal Code’s test, also known as the American Law Institute (ALI) standard [3] (see Table 1).

Table 1 Insanity defense standards.
Full size table

Nowadays, insanity definition and the threshold for satisfying its legal criteria tend to vary depending on the jurisdiction. Yet, in Western countries, the legal standards for insanity often rely on the presence of defendants’ cognitive and/or volitional impairment at the time of the crime [2].

In Italy, forensic experts have three possible conclusions regarding defendants’ criminal responsibility: nonresponsible, with substantially diminished responsibility, and responsibleFootnote 2 [4]. The forensic psychiatric evaluation leading to criminal irresponsibility or substantially diminished responsibility must also detect a cause-effect association between infirmity and specific criminal behaviour. Expert consultants can be appointed by the judge if he/she decides to avail himself/herself of their help or at the same time they can be summoned by the prosecution or by the defense with the aim of assessing the criminal responsibility and social dangerousness of the offender. However, the judge can also decide not to rely upon experts’ opinions nor is obliged to follow their suggestions.

Despite the presence of established legal criteria, the reliability and objectivity of insanity evaluations have been widely questioned, because of the frequent disagreement among forensic experts regarding the same case [5, 6], the lack of standardized procedures to guide the forensic evaluation [7], and the presence of unintentional or cognitive biases [7,8,9,10]. In addition, several factors such as money, prestige, and the amount of public attention attracted by the case have been shown to have an influence on forensic experts’ decisional processes [11]. Another source or limit could be represented by the intrinsic limits of psychiatric diagnosis and the longitudinal variability of psychiatric symptoms [12]. Moreover, discrepancies between forensic reports may be accounted for also by the absence of biological markers available and the relative scarcity of reliable diagnostic tools to guide forensic evaluations. Finally, the dialectic of the criminal trial itself may play a role, with the possibility of the presence of different opinions from different parties which plead their case, which the court or jury weighs.

To overcome these limits, theoretical models, some practice guidelines and forensic assessment instruments have been developed to assist the criminal responsibility evaluation [1, 13,14,15,16,17,18,19,20,21]. Among the tools used to perform insanity evaluations, are included the “Mental State at the Time of the Offense Screening Evaluation” (MSE) [22], the “Rogers Criminal Responsibility Assessment Scales” (R-CRAS) [23, 24], the “Rating scale of criminal responsibility for mentally disordered offenders” (RSCRs) [25], and the “Criminal Responsibility Scale” (CRS) [26]. In addition to those previously mentioned, there is the “Defendant’s Insanity Assessment Support Scale” (DIASS) [21]. Differently from the MSE, which is structured to screen out defendants whose criminal conduct clearly was not caused by significant mental abnormality, and the RSCRs, which evaluates defendants’ cognitive and volitional dimension through a hypothetical scenario based on clinical vignettes, the R-CRAS, RSCRs, and DIASS evaluate defendants’ criminal responsibility associated with the alleged crime.

The aim of the present study was to validate the DIASS, to study its psychometric properties and its factorial structure.

Materials and methods

The present study was conducted from February 2020 to June 2020 and involved the development of 10 hypothetical forensic cases, based on real forensic cases modified to ensure nonrecognizability. Each forensic case included a measure of the defendant’s psychiatric symptomatology at the time of the forensic evaluation through the Brief Psychiatric Rating Scale (BPRS, Expanded Version) [27]. It was chosen to describe the psychiatric symptoms at the time of the assessment, since a detailed description of the symptoms is not available at the time of the crime, at least not such as to be able to fill in a BPRS. Usually, psychiatric symptoms at the time of the crime must be deduced retrospectively by the forensic evaluator on the basis of several factors such as the defendant’s account of the offense, past and present behavior, the characteristics of the alleged crime, and the psychiatric history of the accused, the witness report, or by video recordings, where available. Consequently, we provided hypothetical forensic cases with information related to the state of mind at the time of the crime, which was inferable from the methods used and data available, reported in the historical descriptions. The BPRS scores were based on the information available including what the experts had reported in their forensic reports of the real forensic cases. We decided to use this methodology, despite its possible limitations, because the reports we used to develop the hypothetical forensic cases neither include a BPRS evaluation nor other psychiatric rating scales. By doing so we assumed that the psychopathological elements that the expert had decided to describe in the report and therefore to convey to the judge, were those that the expert had taken into account in the decision-making process. This method has already been used in another study by our group [28]. Five BPRS factor scores were calculated, i.e., manic/excitement, depression/suicidality, hostility, positive symptoms, and negative symptoms [29].

We then sent the 10 hypothetical forensic cases by email to 50 experts in forensic psychiatry and legal psychology from Northern, Central and Southern Italy, asking them to rate just 5 of them (chosen by us) with the help of the DIASS and to apply its decision-making model to reach their final judgement (nonresponsible, substantially diminished responsibility, responsible), while evaluating the remaining 5 cases as they used to.

An ad-hoc form was developed to collect data about the forensic evaluator, such as name and surname, age, sex, years of forensic experience, and specialty.

Forensic cases were emailed to each examiner together with a cover letter explaining the aim of the study. Each evaluator was told to imagine being the judge’s expert consultant and was reminded of the study up to a maximum of three times, 1 month, 2 months, and 3 months after the first email, respectively. Those evaluators who still failed to give feedback were considered nonresponders.

The Defendant’s Insanity Assessment Support Scale (DIASS)

The DIASS is a tool that has been developed on the theoretical model of competent decision-making and reflects core issues relevant to legal insanity in many jurisdictions in Western Countries [21]. It is composed of 9 binary items (present/absent) grouped into 4 dimensions: “Knowledge/understanding of the crime” (3 items), “Appreciating of the crime” (1 item), “Reasoning” (3 items), and “Control of voluntary motor activity” (2 items). The first two dimensions refer to the “Epistemic component”, while the third and fourth dimensions refer to the “Control component”. A box at the end of the scale refers to the final judgements on the Epistemic and Control components, which are scored on a 3-point scale (Intact, partially compromised and compromised). All the items refer to the state of mind of the accused, reconstructed at the time of the crime. The DIASS is an instrument designed to support and guide the insanity evaluation and should be used only after having examined all the legal and health documentation of the defendant, as a framework during the clinical evaluation. The instrument does not produce a compute score, but assists forensic evaluators in identifying which capacities relevant to mental criminal responsibility were present at the time of the offense. For a deeper description of the instrument and its development see Parmigiani et al., 2019 [21].

Forensic cases

Each forensic case described the defendant’s personal, medical, social, and employment history, family relationships, arrest behavior, and current mental status. A summary of the 10 hypothetical forensic cases is shown in Table 2 (The full text is available on the Supplementary Material data file). The severity of psychiatric symptoms was shown through the BPRS.

Table 2 Hypothetical forensic cases summary.
Full size table

Statistical analysis

Statistical analyses were performed with the Statistical Software for Social Sciences v. 20.0 (SPSS). The alpha value was set to 0.05, all tests were two-tailed.

We performed an exploratory factor analysis by principal axis factoring of the DIASS items through Eigenvalue method >1 in order to extract the factors, also observing the scree plot. We rotated the factors by an oblique rotation (direct oblimin), and analyzed the DIASS internal consistency through Cronbach’s alpha.

To investigate Inter-Rater Reliability (IRR) Cohen’s kappa statistics [30] were used; IRR was investigated by comparing every possible combination of pairs of raters, scoring the same forensic case. Landis and Koch [31], have arbitrarily defined intervals in Cohen’s kappa for inter-rater agreement, where 0–0.20 is considered as slight, 0.21–0.40 as fair, 0.41–0.60 as moderate, 0.61–0.80 as substantial, and 0.81–1 as almost perfect agreement.

Spearman’s correlation coefficient was used to investigate associations among the items of the DIASS and the psychopathological dimensions investigated through the BPRS.

Results

The 10 hypothetical forensic cases were sent to a sample composed of 50 forensic experts, of which 80% (N = 40) replied. The characteristics of the sample are shown in Table 3.

Table 3 Forensic experts’ sample characteristics.
Full size table

The exploratory factor analysis by principal axis factoring disclosed a 2-factor solution with a different distribution of the items compared to the original scale (Table 4), explaining 57.6% of the total variance. The first DIASS factor, which we named “epistemic component” comprised five items referring to the knowledge /understanding of the crime context (A1), appreciation of the criminal behavior through the subjective moral standard (B1), and reasoning about the possibility of non-acting/alternative choices (C1), about consequences -pros and cons (C2) and integration of relevant information (C3). The second DIASS factor, named “control component” comprised three items dealing with the knowledge/understanding of the criminality of the act and moral standard (A3), the control of voluntary motor activity through the ability to program, organize, finalize the action (D2), and the ability to inhibit one’s own behavior (D1). Item A2 (knowledge/understanding of the nature of the criminal act) was dropped from the scale because it loaded on both factors. In Appendix A is reported the final structure of the DIASS.

Table 4 DIASS principal component analysis.
Full size table

The DIASS showed good internal consistency (Cronbach’s alpha = 0.86), and a good concurrent validity as shown through the highly significant association between forensic experts’ judgment reached using the DIASS and the court ruling on the real cases (rho = 0.674; p < 0.001).

Cohen’s kappa disclosed substantial DIASS internal consistency among different raters varying from 0.44 to 1, with a mean value of 0.72.

No differences emerged between forensic experts’ final judgement reached with the use of the DIASS and without it, as disclosed by Fisher’s exact test (Table 5).

Table 5 Differences in forensic cases evaluation performed with and without the use of the DIASS.
Full size table

Correlations among the capacities analyzed through the DIASS and the severity of psychiatric symptoms in nonresponsible, with substantially diminished responsibility and responsible defendants are shown in Table 6.

Table 6 Correlations between the items of the DIASS and psychopathological dimension in defendant judged nonresponsible, with substantially diminished responsibility and responsible.
Full size table

Discussion

Our study on a sample of 40 forensic experts confirmed the bifactorial structure of the DIASS, although exploratory factor analysis revealed a different distribution of the items than the original construct/scale. Specifically, item A3, (Knowledge/Understanding of the criminality of the act and moral standard), which was originally conceived as part of the epistemic component, after the exploratory factor analysis, loaded on the control component. In interpreting this result, we must consider that this item deals with an aspect that forensic experts analyze through the observation of external behavior, which is associated with the notion of inhibitory control. For example, the absence of efforts to avoid detection or apprehension may suggest a lack of knowledge/understanding of wrongfulness.

The DIASS showed good internal consistency, concurrent validity and substantial inter-rater reliability. These results underline the good psychometric characteristics of the DIASS as well as the importance for criminal responsibility evaluations of both the cognitive and control components, as already is the case in Italy and in those countries, which use the ALI standard, compared to those that use the M’Naghten [2]. In addition, the role of inhibitory control and executive functions is supported by neuroscientific studies, showing the presence of cerebral alterations in violent offenders, specifically prefrontal dysfunctions [32,33,34,35,36,37,38,39].

The absence of significant differences between final judgements of forensic cases evaluated with and without the use of the DIASS, suggests that our instrument does not alter the decisional process typically used in forensic experts’ practice, despite improving the transparency and verifiability of judgment. In this regard, it is worth underlining that the scale may be apparently of little use for those forensic experts with many years of experience. However, we deem that the DIASS could be useful also for experienced forensic experts, as it would allow them to make explicit in the forensic reports the patterns of reasoning that are generally implicit, making them more objective. In addition, the DIASS could be especially useful to those who are at the beginning of the forensic profession as it could act as support and guidance for the evaluation of the defendant’s accountability as well as the training of forensic specialists.

From the analysis of the correlations between the defendants’ capacities analyzed through the DIASS and the psychopathological dimension measured by the BPRS, three different profiles of forensic experts’ decision-making emerged. While regarding defendants judged nonresponsible the impairment of the capacities analyzed through the items of the DIASS (such as the ability to control one’s own behavior, etc.) were mainly associated with the mania/excitement dimension, for defendants with substantially diminished responsibility the most affected psychopathological dimension was represented by positive symptoms on the BPRS. This is in line with some studies reporting positive, negative, and manic symptoms as to be more frequently associated with judgments of nonresponsibility and could be considered as generally related to the impact of the “psychotic dimension” on responsibility [25, 40, 41]. Finally, regarding responsible defendants, the psychopathological dimensions most involved in the impairment of the cognitive-volitional components were those of hostility and negative symptoms, probably reflecting the fact that our responsible sample was mainly composed of defendants affected by personality disorders such as antisocial, borderline or paranoid, with high levels of anger and hostility. Differently from defendants judged nonresponsible and with substantially diminished responsibility, where the presence of psychotic symptoms plays a role in forensic experts’ decisional processes, regarding responsible defendants, the psychiatric dimension of hostility and negative symptoms (composed of motor retardation, blunted affect, disorientation, and self-neglect) are not associated with a judgement of nonresponsibility, but merely present. In interpreting these results, we must consider that these correlations refer to the psychopathological dimensions present in our forensic hypothetical case (based on real forensic cases modified to make them not recognizable) and that the BPRS referred to the state of mind at the time of the forensic evaluation, as it has been reported in the forensic report, and not at the time of the crime. A possible limitation of this study is represented by the limited number and typology of hypothetical forensic cases (although based on real cases modified to make them unrecognizable) and the reduced number of forensic experts (N = 40) who evaluated the cases.

Despite the DIASS having been validated on forensic cases based on the Italian penal code, it can also be used in those countries where insanity criteria follow the M’Naghten Rule or the ALI standard. In such cases, the evaluators should focus more on those components of the DIASS that refer to the criterion deemed most relevant within a specific jurisdiction. Furthermore, it could be useful in those states such as the Netherlands where no specific legal standards have been formulated to establish when a person is criminally responsible for their actions. In addition, there is a need for cross-cultural validation within both continental European and Common Law legislation.

Moreover, it represents the first step toward a reduction of the heterogeneity of forensic evaluations and toward standardization of their methodology, thus leading to an improvement in the whole criminal justice process.