Relationships Between Patient-Level Factors and Criteria for Fitness to Stand Trial

In Western criminal justice systems, proceedings may be halted if an individual is deemed mentally unfit to stand trial. As a prerequisite to adjudication fitness to stand trial can be evaluated through structured or unstructured assessments. Previous studies suggest limited use of structured assessments in clinical practice. Few studies have looked at the success of unstructured measures of psycholegal abilities, and fewer still have investigated the influence of individual variables on criteria for fitness to stand trial. The purpose of the present study was to examine the relationship between variables relevant to opining fitness as determined by previous research and the criteria for fitness to stand trial. The study yielded significant correlations between the three criteria for fitness to stand trial and the following variables: impaired mental status during assessment, presence of intellectual disability, nature of index offence, socioeconomic status, and all unstructured measures of psycholegal abilities. These results suggest that unstructured clinician assessment of fitness to stand trial can be successful at determining fitness and fulfillment of the three underlying criteria, and further clarify the role of specific symptoms on opinions of unfitness. Future directions for research in the areas of structured professional judgment and fitness restoration are discussed.


Introduction
In Canada, court proceedings may be paused if there is sufficient evidence to indicate that the accused may be mentally unfit to stand trial. Under section 672.23 of the Criminal Code of Canada (herein referred to as the Criminal Code), an individual is presumed to be Fit to Stand Trial (FST) unless the issue of fitness is raised by the Crown Attorney (prosecution), the Judge, or the Defence lawyer [1]. The accused person is considered to be Unfit to Stand Trial (UST) if he or she is unable to meet the criteria outlined in section 2 of the Criminal Code [1], which specifies that: s. 2: "unfit to stand trial" means unable on account of mental disorder to conduct a defence at any stage of the proceedings before a verdict is rendered or to instruct counsel to do so, and, in particular, unable on account of mental disorder to (a) understand the nature or object of the proceedings, (b) understand the possible consequences of the proceedings, or (c) communicate with counsel.
An individual's state of fitness to stand trial is thus related to the individual's mental status and the impact of his or her mental functioning on his or her ability to fulfill the criteria outlined above. This does not mean that a mental disorder is indicative of unfitness; it must be that the individual is impaired on one or more of the above criteria as a result of the mental disorder. Consequently, the presence of a mental disorder is necessary but not sufficient to determine unfitness. However, if the individual does meet one or more of the criteria outlined in s. 2 of the Criminal Code, the Judge may conclude that he or she is UST or Incompetent to Stand Trial [IST] for the American standard [2], and his or her legal proceedings will be suspended until fitness is restored.
However, the Criminal Code is unclear on how to evaluate these criteria so as to determine fitness. Over the last thirty years, numerous studies have attempted to assess the relationship between patientlevel factors (demographic information, psychiatric information, criminal history information, and psycholegal information) and findings of fitness to understand how better to determine fitness.
The majority of studies on demographic factors have left the field unclear, as have the studies on criminal history. The most consensus in the field has been in research related to psychiatric factors, where it has long been clear that those found UST are more likely to have a history of psychiatric symptoms or diagnoses [3]. Part of the reasons for the lack of consensus is that these studies assess fitness using structured assessment measures, though it is clear from the research that clinicians use unstructured assessment measures, thus leaving a gap between empirical research and clinical practice [3,4]. However, of the limitations posed by the literature, the most striking is that very minimal research (with no research occurring in Canada) has been conducted on the influence of the accused's psycholegal abilities as measured by clinicians in an unstructured manner on fitness determinations. An extensive literature review was conducted by the authors of this paper prior to commencing the study detailed below [3].
The aim of the present study is to expand upon the limited research on how specific variables impact determinations of fitness to stand trial, and to resolve many of the above inconsistencies, particularly in addressing the variance of how psycholegal abilities are measured. This is done by focusing on what information clinicians actually rely upon in their evaluations: questions based on their own clinical and diagnostic experiences and knowledge, not on standardized measures or checklists. Few studies have looked at the success of unstructured measures of psycholegal abilities, and fewer still have investigated the influence of relevant variables on individual prongs of fitness to stand trial. The goal of the present study is to examine the relationship between relevant variables and each criterion of fitness to stand trial in a real-life sample in order to better understand how the clinician's interpretation of the accused's answers leads to a determination of fitness. Secondly, this study attempts to address some of the previous methodological issues such as sample bias and statistical limitations.

Methods
The study consisted of a retrospective file review that included 51 patients consecutively referred for an assessment of fitness to stand trial at a psychiatric facility in a large Canadian city in 2014. This sample represented all fitness assessments for the forensic catchment area, thus minimizing sample bias or referral bias from a particular court or jurisdiction. The following study received ethics approval from the institutional research ethics board in July 2015. The study included files from both inpatient and outpatient assessments. Files were coded by two trained raters under the supervision of a co-author and clinical forensic psychologist. The coders met regularly to ensure coding was completed accurately.

Materials
Materials reviewed included discharge summaries from all previous hospital admissions, the most recent criminal record provided by the local police department, previous and current fitness report(s), detention centre records (both previous and at the time of assessment), any notes provided from other consultations, notes from the index fitness assessment, and any email correspondence that was also included in the files, such as from lawyers or mental health providers. Data from the following sections of the report were coded: Identifying Data, Sources of Information, Background History, Mental Status Examination, Fitness to Stand Trial Assessment, and Opinions and Recommendations.

Variables
Variables were coded within the following categories: Demographic Information, Psychiatric Information, Fitness Assessment, and Criminal History. With respect to demographic factors, this study expanded on the current literature by coding for variables that represented changes in home configuration and financial status, as well as variables about immigration status and if the accused has any children. In accordance with other studies, the present study included variables for gender, age, race, level of education, employment status, home configuration, financial/income status, and family history. With respect to psychiatric factors, this study included a wide range of variables, including psychiatric history, number of previous psychiatric admissions, comorbidity, and codes for primary, secondary, and tertiary diagnoses. Information was also gathered about all aspects of the Mental Status Examination including mood, affect, thought content, presence of nonsensical phrasing, thoughts of harm, paranoia, presence of hallucinations, and level of aggressiveness during the assessment. With respect to criminal factors, the present study coded for prior criminal history, including information about convictions, charges, and prior incarceration. It expanded upon the index offence, coding not only the nature of the offence, but also details about the victim, intoxication of the accused, level of injury in the offence, and whether or not a weapon was used. For the purposes of this study, psycholegal abilities were coded based on the psychiatrist's summary of the individual's answers to the questions posed during the assessment. There are seven psycholegal abilities that mimic those assessed by the standardized measures but are far more open-ended when assessed in an unstructured interview setting [3].
Ability on each of these questions was coded as Yes or No, as per the fitness report outlined by the psychiatrist. The other psycholegal variable of interest was location of the assessment. Fitness assessments were conducted either on an inpatient admission basis, on an in-person basis through the outpatient Fitness Clinic, through a secure video assessment, or on an outpatient basis at a detention centre.

Results
Fifty-one assessments of fitness to stand trial were ordered by the court to the forensic psychiatry program in 2014. Four files were removed due to an inconclusive finding during the assessment (e.g. neither Fit nor Unfit). Five files were removed because of their limitation to the administration of a Treatment Order, and as such did not include much of the required data. Of the remaining forty-two files, five files were reviewed by both raters to establish and evaluate interrater reliability (IRR). IRR was excellent, with kappa values ranging from .84 to .93. The mean estimated kappa was .90. Data were analysed and treated for missing values, outliers, and normality assumptions.
With respect to analysis, the type of correlation statistic calculated depended on the nature of the data. For continuous variables, the Pearson r correlation was calculated, as well as for categorical, dichotomous variables as r approximates Phi when both of the variables each have two possible values. For correlations that involved categorical variables that have an unequal number of possible values-for example, when one is dichotomous and one is trichotomous-the Cramer's V correlation was calculated, a statistic that is used to measure the strength of association between two categorical variables.
A standardized fitness assessment test was used for only one file, consistent with the rationale of the study and the previous research. The other forty-five cases used an unstructured assessment interview to evaluate the aforementioned psycholegal questions. In addition, the group distribution between those found FST or UST and those able to fulfill or not able to fulfill the respective criteria is almost identical, which is consistent with the rationale underlying the hypotheses.

Sample Description
The total number of files used for descriptive purposes was forty-two (N = 42). Thirty-six subjects (85.7%) were judged to be FST and six subjects (14.3%) were judged to be UST. There were thirty-five males and seven females, which is representative of the larger Canadian forensic psychiatric population. The mean age of the total sample was 36.86 years, with a standard deviation of 12.58. A plurality of the sample was Caucasian/White (47.8%) and a majority of the sample was single (76.1%). Seventy-four percent (73.8%) of the sample was unemployed (n = 31), 2.4% were students (n = 1), 4.8% were on disability (n = 2), and 7.1% were retired (n = 3). Forty-three percent of the sample lived alone (n = 18, 42.9%) at the time of the assessment order, and received some form of welfare or disability funding (n = 26, 61.9%). Four files were missing information on home configuration, and ten files were missing information on finances. Just under half of the sample (n = 20, 47.6%) was previously employed and receiving income from employment prior to the Fitness to Stand Trial Assessment.
A minority of the sample had a family history of mental illness, substance abuse, or criminal activity. Schizophrenia was the most prevalent primary psychiatric diagnosis (40%). Thirty-eight percent of the sample had a comorbid diagnosis, the most prevalent being Substance Abuse Not Otherwise Specified (NOS). In only 28.6% of the cases was a secondary psychiatric diagnosis made, followed by only 7.1% having a listed tertiary psychiatric diagnosis.
For the three criteria outlined in s. 2 of the Criminal Code, 85.7% of the sample (n = 36) was able to understand the nature and object of the proceedings, 83.3% of the sample (n = 35) was able to understand the consequences of the proceedings, and 85.7% of the sample (n = 36) was able to communicate with counsel. These percentages are consistent with the percentages of people found FST versus UST stated above.
Over half (57.1%, n = 24) of the sample had previous general convictions and 33.3% of the sample (n = 14) had previous general charges. Fifty percent of the sample (n = 21) had previous violent convictions and 21.4% of the sample (n = 9) had previous violent charges. Less than 10% of the sample (9.5%, n = 4) had previous sexual convictions and 2.4% of the sample (n = 1) had previous sexual charges. Table 1 shows the significant Cramer's V correlations between demographic variables and all three criteria for fitness to stand trial. There were no significant correlations found for the following variables: age, gender, race/ethnicity, marital status, level of education, home configuration at the time of order, home configuration before the order, financial/income status at the time of the order, changes in home configuration, or changes in financial/income status, family history of mental illness, or family history of substance abuse.   Table 2 shows the significant Cramer's V correlations and the significant Pearson's r correlations between psychiatric variables and all three criteria. Presence of nonsensical phrases was significantly negatively correlated with the fulfillment of all three criteria. Mood and the three criteria showed strong Cramer's V correlations, such that those with a normal or stable mood were able to fulfill the criteria. With respect to specific criteria, affect was significantly positively correlated with under-standing the nature and object of the proceedings and communicating with counsel, such that appropriate affect made it more likely to be able to fulfill the aforementioned criteria. Thought content and understanding the consequences of the proceedings showed a strong correlation as well, suggesting that normal thought content was more likely to be indicative of someone's ability to understand the consequences. Finally, intellectual disability was very strongly negatively correlated with understanding the consequences of the proceedings such that a presence of intellectual disability makes it more unlike-ly for someone to understand the consequences of the proceedings. There were no significant correlations found for the following variables: number of previous psychiatric admissions, history of child abuse, previous psychiatric admissions, substance abuse, primary psychiatric diagnosis, secondary psychiatric diagnosis, tertiary psychiatric diagnosis, capability to consent to medication, comorbidity, presence of hallucinations during assessment, presence of thoughts of harm to self or others, presence of paranoid feelings, or aggression.

Criminal Factors
For criminal variables, the nature of the index offence was strongly correlated with understanding the consequences of the proceedings (V = .42, p < 0.05), such that those who were accused of committing a violent offence were able to understand the consequences of the proceedings. Previous sexual convictions were found to be strongly negatively correlated with understanding the nature and object of the proceedings, and communicating with counsel (r = -.33, p < 0.05). The presence of previous sexual charges was negatively correlated with understanding the nature and object of the proceedings and with communicating with counsel (r = -.38, p < 0.05), but positively correlated with understanding consequences of the proceedings (r = .35, p < 0.05). However, these results should be interpreted with caution as there was an outlier effect with the vari-ables that were found to be significant in these analyses. When the data was analyzed, it appeared that small cell numbers could have skewed the results we obtained, as there was only one individual with a previous sexual charge, and only four with previous sexual convictions. There were no significant correlations found for the following variables: previous general convictions, previous general charges, previous violent convictions, previous violent charges, previous incarceration, victim gender, level of injury, or relationship to the victim. Table 3 displays the significant Pearson's r correlations between the psycholegal abilities, the three criteria, and the final opinion of fitness. There were no significant correlations found for location of assessment. Table 3. Relationship between the psycholegal abilities, the three criteria and the final opinion of fitness.

Discussion
The goal of the present study was to evaluate the use of unstructured measures of psycholegal ability as a valid measure of fitness to stand trial, and to investigate the relationships between specific variables and the three criteria outlined in the Canadian Criminal Code. These two distinctions are crucial because while existing research has looked at the role of various factors on final opinions of fitness to stand trial, few studies have analyzed the role of those factors on specific prongs of fitness or looked at measures of psycholegal ability beyond standardized assessment. The results of the present study support the main hypothesis regarding the use of unstructured measures of psycholegal ability as tools for fitness assessment. All seven psycholegal abilities outlined above showed significant positive correlations with the respective criteria for fitness, which supports the notion that these unstructured questions successfully assess the ability to understand the nature and object of the proceedings, the ability to understand the consequences of the proceedings, and the ability to communicate with counsel. The results further support the role of specific symptoms and variables in opinions of fitness to stand trial and the three prongs.

Demographic Factors
The results indicate significant correlations between employment status and all three criteria, as well as between financial status prior to the time of assessment and all three criteria. That the other demographic variables were not correlated is desirable, as otherwise it would indicate a bias if demographic information was used to determine someone's ability to fulfill a given criterion. The demographic relationships could be due to a number of things, one of which could be a confounding variable. Many individuals who were unable to fulfill the relevant criteria suffer from a psychiatric disorder that may increase the likelihood that someone is unable to find employment, thus also leading them to a lower financial status prior to assessment. In the current study the majority of the group that was able to fulfill the three criteria was unemployed at the time of their assessment, suggesting that something (perhaps a psychiatric disorder or a lengthy pre-trial detention) led to their inability to find employment. Furthermore, even though a third of the data were missing on the financial status variable, the majority was receiving income from employment at the time before their assessment, suggesting that those who were previously employed may have suffered circumstances that led to their loss of employment and subsequently, their involvement with the criminal justice system. Thus, these results should be interpreted with caution.

Psychiatric Factors
No significant correlation was found between presence of psychotic disorder (or any specific disorder) and any of the three psycholegal criteria and this reflects the important distinction made in the Criminal Code (1). The presence of a mental disorder is necessary to be opined UST, but it is not sufficient on its own. The mental disorder must impact those abilities outlined in the Criminal Code. Thus, the fact that the specific symptoms of mental illness (e.g. impaired mood, nonsensical phrases) are significant in determining ability, and not the disorder itself, is commensurate with the standard outlined in s. 2 of the Criminal Code. Significant results were found for mood, affect, thought content and nonsensical phrasing during assessment, suggesting that indicators of impaired mental status are associated with the likelihood of someone being found unable to fulfill the relevant criteria. These findings support the distinction outlined above.
With respect to disorders, though a specific psychiatric diagnosis was not found to be correlated with any of the criteria, presence of an intellectual disability was found to be negatively correlated with ability to understand the consequences of the proceedings. This result regarding intellectual disability is consistent with the findings of Gay et al., in that intellectual disability was related to unfitness [5]. However, the specific psycholegal criterion it impacts upon is different. Gay et al. found that understanding the nature and object of the proceedings (referred to as factual understanding in the American fitness standard) was influenced by the presence of an intellectual disability, whereas the present study found that understanding the consequences of the proceedings was influenced by this disability (5). It may be that understanding the consequences of the proceedings requires an ability to understand the nature and the object of those proceedings. Therefore, intellectual disability impacts one's ability to appreciate the consequences, and inability to appreciate one's own legal situation results in findings of unfitness.

Criminal Factors
As with previous research, there was a significant correlation between the index offence and understanding the consequences of the proceedings, suggesting a link between the severity of crime and appreciating possible outcomes, in particular when the index offence is violent [6]. This could be a result of underestimating the sanctions associated with the crime or even malingering by the accused. An accused individual who is charged with a violent offence may believe it is easier to feign unfitness, and future research could analyze the prevalence of malingering in fitness assessment, as little research has been done in this area. However, it is promising that few criminal variables are correlated with determinations of fitness under s. 2 of the Criminal Code, as this suggests that bias is not playing a role in such determinations.

Psycholegal Factors
With respect to the psycholegal abilities correlated with understanding the nature and object of the proceedings, the strongest correlations existed for Identification of Roles, Description of Roles, and Understanding Outcomes. The nature of the proceedings is represented by the identification and understanding of key roles in the courtroom, whereas understanding the outcomes represents the object of the proceedings. The other psycholegal abilities all showed strong, positive correlations with this prong as well, likely because the knowledge of one's charges, the ability to describe said charges, and the ability to define pleas and legal terms all contribute greatly to the understanding of the nature of the proceedings.
Concerning the psycholegal abilities that measure understanding the consequences of the proceedings, the strongest correlations exist for Description of Roles, Understanding Outcomes, and Definition of Legal Terms. The Understanding Outcomes ability exhibits a perfect correlation, which supports the notion that understanding the consequences of pleas such as guilty and not guilty indicates an understanding of the consequences of the entire proceedings. The Description of Roles ability also exhibits a perfect correlation, and this is in line with what this ability tests, which is an understanding of how the system and its key players work for and against the accused, which is related to the potential consequences of the proceedings. Finally, the Definition of Legal Terms is strongly positively correlated with this criterion as well, and this may be because terms such as Oath and Perjury come with their own outcomes and consequences (particularly perjury), and this relates to the criterion of understanding the consequences of the proceedings as a whole.
With regard to the psycholegal abilities that measure communication with counsel, the strongest correlations exist for Identification of Roles, Description of Roles, and Understanding Outcomes, but it is important to note that all of the abilities are correlated strongly with this prong because ability to communicate pertinent information to counsel encompasses all of these abilities. However, the two Roles abilities may be strongly correlated because this prong measures the accused's communication with his or her counsel, and these abilities assess the identification and understanding of who represents which side and how defence and prosecution will conduct their case. It is important to note that the correlations for communication with counsel are identical to those for understanding the nature and object of the proceedings. This could perhaps be due to the variables measuring the exact same ability from a theoretical perspective. However, it may also be because the ability to understand the nature and object of the proceedings is an explicit ability, requiring the accused to communicate with counsel about the actions of the individuals in the courtroom, whereas understanding the consequences of the proceedings is more implicit and can be fulfilled without being able to expressly communicate that knowledge.

Limitations
Though the results of this study generally support previous findings, and some suggest new developments with respect to fitness assessments, several limitations exist and caution should be exercised when forming conclusions. The study was limited to a small sample size and the groups used in analyses were unequal. Studies in this field usually collect files across a range of a few years to amass a large enough sample and equal enough groups (e.g. 100 files) as there are generally more FST individuals than UST individuals. Although steps were taken to correct for these issues statistically, this could have affected the results. However, the current sample is comprised of all referrals across multiple jurisdictions for an entire year, and in this way, is representative of the population.
The results are also limited by the variation in data collected in the files. Files varied in the information provided with respect to questions asked in the assessment, the legibility of the rough notes from the fitness assessment, the detail in the individual's background history, as well as how detailed the individual's criminal history was. The manner in which some of the variables were coded also limited the ability to make conclusions. The dimensions of mood, affect, thought content, nonsensical phrasing, and demographics were all coded categorically, sometimes simply as Yes or No or Present or Absent. Some of these limitations were in part due to the lack of specificity, or comprehensive historical information. While some would suggest that the assessment should not focus on the accused's background, it can be argued that knowing which factors an individual brings with them to an assessment helps to inform possible restoration. Finally, the results of the study support the use of unstructured measures of psycholegal ability, however, there was no comparison group of accused assessed with standardized instruments, which would be an interesting future study.

Conclusion
Patient-level factors and fitness to stand trial IJRR 2018;1 (3) i The present study has provided the field with interesting developments regarding factors that affect specific prongs of fitness, as well as factors that affect the overall opinion of fitness. In particular, this study is among the first of its kind to look at the relationship between the criteria for fitness and individual variables, especially in Canada where the codified fitness standard is relatively new in comparison to other jurisdictions. In sum, the study supports the use of unstructured measures of psycholegal ability when assessing fitness and provides support for the criteria outlined in the Criminal Code and the factors that clinicians use to inform their decision. Factors include impairment in mental status, relevant psychiatric history, nature of index offence, and all unstructured measures of psycholegal abilities. To our knowledge, this study is among the first to evaluate unstructured measures of psycholegal ability as well as among the first to assess the relationship of specific variables to specific criteria for fitness to stand trial.
Conflict of Interest: none