We evaluated the clinical characteristics of children and adolescent psychiatric patients who had played with fire or had committed arson. The hypothesis that they are more likely to be boys, have a history of abuse, have a diagnosis of ADHD, and antisocial behavior was partly supported.

Among the case groups, the age at the time of hospital consultation following the incident ranged from 6 to 18 years, with a mean of 12.4 years. Kolko investigated 307 children (aged 6–13 years) from nonpatient, outpatient, and inpatient samples who were classified into fire-setters, match-players, and non-fire-setter groups [2]. This cited study showed that children and adolescent psychiatric patients were more likely to be boys, even at an older age. Moreover, in another study, female juvenile arsonists were reported to be more likely to be experienceing family stress, while male juvenile arsonists were more likely to be involved with delinquent groups [23]. These studies showcase that more in-depth findings could have been obtained in the current research had we examined the case group in terms of family stress and involvement with delinquent groups. Future researchers are suggested to conduct such examinations.

Fireplay and arson were strongly associated with ADHD in this study. The association between ADHD and fire-related behavior was shown not only in community populations, as shown in previous studies [15, 16], but also in the clinical population of children and adolescents in a psychiatry department. It is possible that the impulsivity specific to ADHD is more associated with fire-related behaviors than the restricted interests specific to ASD. Previous studies in community populations have shown that family support, such as living with both parents, decreases the probability of engaging in fire-setting behavior [15]. Therefore, examining the environment of patients with ADHD in the clinical population under scrutiny in the current study might lead to important insights as to ways to improve fire-related behaviors. Since the impulsivity and vigilance deficits that characterize ADHD might place these children at higher risk for specific types of burn injuries (vs. children without ADHD) [24], it might be clinically useful to examine children and adolescents with the characteristics similar to those of the case group of this study for burn injuries.

There was no statistically significant difference in the percentage of abuse history between the case and control groups after multivariate logistic regression analysis. Previous studies have shown a relationship between fire-related behaviors and a history of abuse in community populations and healthy youth [14]. These differences in the results of the cited study and the current study may be attributed to the fact that both the case and control groups in this study were children and adolescent psychiatric patients. Stress from parents due to abuse was reported to motivate fire-setting [17], albeit the children and adolescent psychiatric patients in this study were participants who went to consultations in the hospital because of stress at school, at home, or in relationships with parents or friends. Moreover, both the case and control groups had a certain percentage of abuse history, regardless of the presence of fire-related behaviors. Therefore, the proportion of patients with a history of abuse did not differ significantly between the case and control groups.

The association between antisocial behavior and fire-setting behavior was shown not only in the community populations of previous studies [18, 19], but also in the clinical settings shown in the present study. Previous studies have shown a link between fire-setting behaviors and conduct disorder, underpinning that participants in this study may have a likelihood of developing conduct disorder as they age. Because fire setting and other antisocial behaviors increase the risk of lifetime and current psychiatric disorders, even in the absence of a DSM-IV diagnosis of antisocial personality disorder [25], it might be necessary to follow children with antisocial behaviors to prevent the development of antisocial personality disorder.

Several psychiatric studies have been conducted on adults involved in arson. Psychiatric diagnoses commonly associated with arson include schizophrenia, bipolar disorder, substance use disorders, mood disorders, anxiety disorders, intermittent explosive disorder, pervasive developmental disorders, attention deficit hyperactivity disorder, and intellectual disability [26, 27]. However, it is important to note that not all adults with psychiatric disorders commit arson, and other factors such as personality disorders and environmental causes may also play a role [28, 29]. Some studies have suggested that there is a pathway for children with ADHD through oppositional defiant disorder and conduct disorder to antisocial personality disorder [30, 31]. ADHD and antisocial behavior were also identified as significant factors associated with fire-related behaviors for the case group in this study, even after multivariate analysis. Therefore, the case group of this study, and children and adolescents with similar characteristics, needs to be monitored to ensure that these children and adolescents do not develop psychiatric disorders, including the use of illegal drugs that lead to arson in adulthood. There is limited research that verifies the association between arson committed by adult psychiatric patients and childhood psychiatric disorders. However, this study provides some insights regarding the link between patients diagnosed with ADHD and antisocial behaviors in childhood and the likelihood of it leading to their committing arson as adult psychiatric patients later in life.

This study has several limitations that must be considered. First, a measurement bias may have existed for several reasons. Ascertainment bias affected this study, as it could not completely define fireplay or arson. Moreover, reporting bias was present in this study. Not all psychologists and psychiatrists in charge at the hospital actively asked all patients about engagement in fire-related behaviors, and since most data on fireplay or arson were based on the system used to search for words in all medical records, this might have led to missing information or patients who were involved in fireplay or arson. Furthermore, most data of the control group were registered after the initial consultation. Thus, it is possible that additional information regarding substance and child abuse could have been obtained in case further examinations were possible. Second, this study cannot be considered representative of children and adolescent psychiatric patients involved in fireplay or arson because it was conducted in a single hospital. Moreover, and as mentioned in the Methods section, some psychiatric disorders, such as drug-induced psychiatric disease, were excluded from the study prior to the initial consultation; this introduced selection bias to the sample and hindered generalizability. Finally, the results of this study confirmed the associations between fire-related behaviors and the clinical characteristics of children and adolescent psychiatric patients, but could not determine causal relationships.