Nonsuicidal Self-Injury and Its Mediation Effect on the Association Between Posttraumatic Stress Disorder, Depression, and Suicidal Behavior in Firefighters
Article information
Abstract
Objective
This study aimed to investigate the prevalence, clinical characteristics, and the correlates of nonsuicidal self-injury (NSSI) in firefighters. We also investigated the mediating role of NSSI frequency in the association between posttraumatic stress disorder (PTSD), depression, and suicidal behavior.
Methods
A total of 51,505 Korean firefighters completed a web-based self-reported survey, including demographic and occupational characteristics, NSSI, PTSD, depression, and suicidal behavior. Multivariable logistic regression analyses and serial mediation analyses were performed.
Results
The 1-year prevalence of NSSI was 4.67% in Korean firefighters. Female gender, the presence of recent traumatic experience, and PTSD and depression symptoms were correlated with NSSI. Serial mediation analyses revealed that NSSI frequency mediated the association between PTSD, depression, and suicidal behavior; it indicates more severe PTSD was sequentially associated with more severe depression symptoms and more frequent NSSI, leading to higher risk of suicidal behavior.
Conclusion
NSSI is prevalent and may play a significant mediating role when PTSD is associated with suicidal behavior in firefighters. Our results imply the need for screening and early intervention of NSSI in firefighters.
INTRODUCTION
Nonsuicidal self-injury (NSSI) is defined as the deliberate, self-inflicted destruction or alteration of body tissue without suicidal intent that exceeds the boundaries of social norm [1,2]. The fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) considers NSSI as a distinct diagnosis in the section of “conditions for further study,” [3] which indicates that this condition is a major public health concern. Although NSSI is known to occur more often in adolescence, it also appears in adulthood [4,5]. However, studies investigating NSSI in adulthood are severely lacking when compared with those in adolescents. Considering that NSSI is a strong predictor of future suicidal behavior, including suicidal thoughts and attempts [6-8], it may be important to identify the prevalence of NSSI and its risk factors in diverse adult populations.
It is well known that firefighters are at high risk of developing mental health problems, such as posttraumatic stress disorder (PTSD) and depression [9], due simply to their frequent exposure to traumatic events than the general population. In addition, firefighters have a higher prevalence of suicidal behavior than the general population [10], and they have a higher suicidal rate compared with the line-of-duty mortality rate [11]. However, to date, the percentage of firefighters engaging in NSSI behavior remains unclear and, moreover, there is little research on the patterns and correlates of NSSI in the firefighter population. One study has shown a high rate of NSSI behavior (16.4%) in firefighters and also reported that their NSSI behavior was associated with race (Native American), physical problem (cancer), department type (volunteer), and the role of emergency medical services (EMS) [10]. This previous study employed a retrospective survey to estimate the NSSI rate among firefighters using only the following question without inquiring about the frequency and methods taken: “Since becoming a firefighter, have you ever actually engaged in nonsuicidal self-injury?” Thus, the 1-year prevalence of NSSI and its correlates in firefighters have still been uninvestigated.
In addition to investigating the prevalence and correlates of NSSI in firefighters, it may be important to examine whether and how NSSI influences suicidal behavior. Many studies on adolescents and young adults have shown that NSSI behavior has a strong association with concurrent and future suicidal behavior [8,12-18]. However, whether NSSI can be considered as a risk factor or a mediator in suicidal behavior has not been elucidated to date in the adult population.
It is well established that PTSD, a psychiatric disorder commonly found in firefighters, is a predictor of future suicidal behavior [19,20]. Previous research suggests that comorbid depression plays a critical role in the association between PTSD and suicidal behavior [21,22]. Even though depression is independently associated with suicidal behavior, it has been reported to play a mediator role when PTSD is linked to suicidal behavior [23]. The possibility of NSSI influencing the association between PTSD and suicidal behavior through the mediation of depression needs to be investigated. Empirical studies have shown that NSSI is used to regulate the negative affect, such as anxiety and depression. According to a theory, frequent NSSI increases suicidal behavior, since it could decrease the sensitivity to pain and consequently increase the acquired capability for suicide [24-26]. When considering the above, it is possible that NSSI, through depression, can induce that path from PTSD to suicidal behavior. However, to date, no studies have investigated whether the relationship among PTSD, depression, and suicidal behavior can be explained by NSSI.
The purpose of this study was twofold: 1) to investigate the 1-year prevalence, correlates, and characteristics of NSSI in firefighters and 2) to examine whether NSSI mediates the association among PTSD, depression, and suicidal behavior.
METHODS
Participants
This nationwide cross-sectional study was conducted between May and June of 2019 via a self-reported online survey, administered to all firefighting institutions across South Korea, which included 215 fire stations. Among a total of 52,759 firefighters in South Korea, 51,505 firefighters participated in the survey. They completed the self-reported questionnaire, including the number of NSSI behavior in the past year as well as the demographic and occupational characteristics, traumatic experience in the past year, suicidal behavior, PTSD symptoms, and depression symptoms. Survey respondents were anonymous and completed the survey on a voluntary basis. The anonymized dataset was provided by the National Fire Agency of South Korea and informed consent of the participants was waived by the SNUBH IRB (institutional review board of Seoul National University Bundang Hospital) (IRB No: X-1908/559-907).
Measures and assessments
Demographic and occupational characteristics
Demographic and occupational characteristics were obtained using the same self-reported questionnaire. Demographic characteristics included age, sex (male or female), and marital status (married, never married, or divorced/separated/widowed). Occupational characteristics included length of work (month), the presence of recent traumatic experiences, occupational status (full-time firefighter, social service agents, or conscripted firefighter), and current duty. The roles of firefighters include fire suppression, special investigation of the cause of fire, paramedic services providing EMS, rescuing people who are trapped or in medical emergencies, training other firefighters, and others. For analysis, the roles were categorized into the following: fire suppression, EMS (including paramedic and rescue services), and officer (including administrators, special investigators, trainers of firefighters, and communicational and informational system operators).
NSSI in the past year
The Functional Assessment of Self-Mutilation (FASM) [27] was used to assess NSSI in firefighters. The FASM was designed to assess the methods, frequency, and functions of self-reported NSSI. We used the first part of FASM, a checklist of NSSI, asking whether they purposefully engaged in each of the 11 different self-harm behavior (plus a fill-in “other” category) within the past year, and if so, how often they committed such behavior on a 7-point Likert scale (0: not at all to 6: more than 6 times). These 11 behaviors include cutting/carving, burning, self-tattooing, scraping, erasing (i.e., using an eraser to rub skin to the point of burning and bleeding) skin, hitting self, pulling hair, biting self, inserting objects under nails or skin, picking at a wound, and picking areas to draw blood.
Suicidal Behavior
Suicidal behavior was assessed using the Suicidal Behaviors Questionnaire–Revised (SBQ-R) [28], which is a 4-item self-reported questionnaire to inquire about different aspects of suicidal behavior. Item 1 explores whether the respondents have ever thought about or attempted suicide in his/her lifetime. Item 2 evaluates how often the respondents have thought about suicide over the past 12 months. Item 3 inquires about threats of suicide attempts, and item 4 explores the self-reported likelihood of suicidal behavior in the future. The first item is scored on a 4-point Likert scale from 1 (never) to 4a (I have attempted to kill myself, but did not want to die) or 4b (I have attempted to kill myself, and really hoped to die). Item 2 is measured on a 5-point Likert scale from 1 (never) to 5 (very often, 5 or more times). Item 3 is rated on a 3-point Likert scale from 1 (no) to 3a (yes, more than once, but did not really want to die) or 3b (yes, more than once, and really wanted to do it). Item 4 is rated on a 7-point Likert scale from 0 (never) to 6 (very likely). The total score of the SBQ-R ranges from 3 to 18 and higher scores reflect greater risk for suicidal behavior. The most appropriate total cutoff score for the identification of individuals with high risk of suicidal behavior based on the SBQ-R was reported as 7 for the nonclinical samples [28].
The presence of recent exposure to traumatic events
Exposure to traumatic events during the previous year was identified by using the self-reported measure, developed by Beaton et al. [29], which assesses the duty-related incident stressors. Twenty-two items were selected among the original 33 incident stressors. We excluded two stressors related to gunshots due to the generally low incidence of gunshots in South Korea. “Witness duty-related death of coworker” and “coworker firefighter fire fatality (not witnessed)” were changed to “witness duty-related death or suicide of coworker” and “coworker death or suicide (not witnessed),” respectively. Finally, three additional stressors, “remove the body of suicide victim,” “remove a severely decayed corpse,” and “involved in a safety accident that received public spotlight,” which were reported to be frequently encountered and associated with high level of stress in Korean firefighters, were added. The list of traumatic events used in this study is presented in Supplementary Table 1 (in the online-only Data Supplement). Firefighters who reported 1 or more events were coded as 1, while those who reported none were coded as 0.
PTSD symptoms
PTSD symptoms were assessed using the Korean version of PTSD Checklist–for DSM-5 (PCL-5). The PCL-5 is a 20-item self-reported measure evaluating the degree to which an individual has been bothered in the past month by DSM-5 PTSD symptoms [30]: intrusions, avoidance, negative alteration in cognition and mood, and alterations in arousal and reactivity. We instructed the participants to choose and describe the most traumatic event from a list of traumatic events and fill out the PCL-5 with this event in mind. Each item was measured on a 5-point Likert scale (0: not at all to 4: extremely). Higher scores indicated higher severity of PTSD symptoms. Items 1–5 correlated with symptoms within Cluster B (intrusions); items 6–7 with Cluster C (avoidance); items 8–14 with Cluster D (negative alteration in cognition and mood); and items 15–20 with Cluster E (alterations in arousal and reactivity). Participants were considered to have experienced the symptom when they recorded a score of 2 or higher (moderate to extreme) in each item. According to an algorithm-derived PTSD diagnosis method [3], we defined probable PTSD as having the required number of symptoms in each cluster of the DSM-5 criteria: 1 B item, 1 C item, 2 D items, and 2 E items.
Depression symptoms
Depression symptoms were assessed using the Korean version of Patient Health Quetionnaire-9 (PHQ-9) [31,32]. Respondents rated 9 items based on the DSM-IV criteria of major depressive disorder, measured on a 4-point Likert scale (0: not at all to 3: nearly every day) based on their experiences during the past two weeks. The PHQ-9 total score ranges from 0 to 27; higher scores indicated a greater severity of depressive symptoms. The total score of 10 or higher was defined as probable depression [32,33].
Statistical analysis
Descriptive statistics were used to summarize the demographic, occupational, and clinical characteristics of participants, as well as to present the 1-year prevalence of NSSI in firefighters. A multiple logistic regression analysis was used to investigate the demographic (age, sex, and marital status), occupational (work length, occupational status, current duty, and recent traumatic experience), and clinical (probable PTSD and depression) characteristics as correlates of 1-year NSSI. The dependent variable was the presence of NSSI in the previous year. The results were shown as the odds ratios (ORs) and 95% confidence intervals (CIs).
To examine linear association among PTSD and depression symptoms, NSSI frequency, and suicidal behavior, we calculated the Pearson’s correlation coefficient. The variables were significantly correlated (p<0.001); however, no associations exceeded the recommended cutoff for multicollinearity (r>0.80) [34]. Following this analysis, a serial mediation model [35] was conducted with depression (1st mediator) and NSSI frequency (2nd mediator) as mediators on the association between PTSD symptoms and suicidal behavior. For the analysis, the PCL-5 total score was used for PTSD symptoms, the PHQ-9 total score for depression symptoms, and the SBQ-R total score for suicidal behavior as continuous variables. In a serial mediation model, mediators are assumed to have a direct effect on each other [35], and the independent variable (PTSD symptoms) influences the mediators in a serial fashion, and subsequently, the dependent variable (suicidal behavior). A serial mediation model provides the results found in the simple mediation models, including the total effect (the relation between independent and dependent variables without controlling for mediators), direct effect (the relation between independent and dependent variables after controlling for mediators), and total indirect effect (the role of all mediators in the relation between independent and dependent variables). A serial mediation model with two mediators can have three specific indirect effects, providing information regarding the role of a specific mediator with respect to the relationship between the independent and dependent variables: 1) through mediator 1 alone (a1b1); 2) through mediator 2 alone (a2b2); and 3) through mediators 1 and 2, in serial fashion (a1db2) (Figure 1). Hayes’ [35] mediation analysis employs bootstrapping, which is a nonparametric resampling technique involving random and repeated subsampling of data without the need to satisfy the assumption of normally distributed data. Ten thousand bootstrap samples were used to determine the mediating effect of the 95% CI. The mediating effects were significant when the 95% CI does not contain zero.
A serial mediation model showing that the effect of PTSD on suicidal behavior is mediated by depression and NSSI frequency serially. All values are p<0.001. a1, direct effect of PTSD on depression; a2, direct effect of PTSD on NSSI frequency; b1, direct effect of depression on suicidal behavior; b2, direct effect of NSSI frequency on suicidal behavior; d, direct effect of depression on NSSI frequency; c, total effect of PTSD on suicidal behavior without accounting the effects of mediators; c’, direct effect of PTSD on suicidal behavior controlling the indirect effect through the mediators; PTSD, posttraumatic stress disorder; NSSI, nonsuicidal self-injury.
The data were analyzed using IBM SPSS statistics 22 software (IBM Corp., Armonk, NY, USA) and PROCESS macro v3.3 for SPSS. Statistical significance was defined as a two-tailed p-value of <0.001.
RESULTS
The mean age of the 51,505 participants was 40.18 (standard deviation [SD]=9.90) years. Among them, 47,321 (91.88%) were male, and 35,558 (69.04%) were married and un-separated. The average length of work was 148.35 (SD=117.75) months, and 29,921 participants (58.09%) experienced traumatic events in the past year.
The prevalence and the correlates of past-year NSSI
Among the total study population of 51,505, the 1-year prevalence of NSSI was 4.67% (n=2,403). In addition, 1.39% (n=715) reported that they engaged in NSSI more than five times in the past year. The demographic, occupational, and clinical characteristics of participants by the presence of NSSI in the past year are presented in Table 1.
Demographic, occupational, and clinical characteristics of the NSSI group and non-NSSI group (N=51,505) and the correlates of past-year NSSI (N=51,387) in firefighters
To investigate the correlates of past-year NSSI, a logistic regression was conducted on 48,991 individuals who did not commit self-harm and 2,396 participants who engaged in NSSI (118 participants with missing data were excluded). As shown in Table 1, females had 58.4% increased likelihood of past-year NSSI (OR=1.584 [95% CI=1.380–1.819]) than males. Regarding occupational status, social service significantly increased the risk of NSSI in the past year (OR=3.285 [95% CI=2.636–4.094]) when compared with full-time firefighting. With respect to current duty, EMS had 18.9% reduced odds of past-year NSSI (OR=0.811 [95% CI=0.727–0.905]) than fire suppression. The presence of recent traumatic experience (OR=1.862 [95% CI=1.685–2.057]), probable PTSD (OR=3.802 [95% CI=3.296–4.386]), and probable depression (OR=4.807 [95% CI=4.169–5.542]) were also significantly positively associated with past-year NSSI.
The characteristics of NSSI
Participants who engaged in NSSI in the past year (n=2,403) reported an average of 4.29 (SD=5.96) self-injurious behavior in the past 12 months, with an average of 1.98 NSSI methods. Twenty-seven percent (n=654) of the participants with past-year NSSI had a high risk of suicidal behavior (SBQ-R total score ≥7).
Table 2 shows the detailed NSSI methods used among the participants. Biting (34.21%) and scraping skin (34.79%) were most frequently reported. A multivariable logistic regression showed that cutting, hitting, pulling hair out, picking at a wound, biting, and scraping skin were significantly associated with high risk of suicidal behavior, whereas giving a tattoo was associated with low risk of suicidal behavior.
Logistic regression analysis on the methods of NSSI and high risk of suicidal behavior in participants engaged in NSSI in the past year (N=2,403)
Serial mediation of depression and NSSI frequency on the association between PTSD and suicidal behavior
In the results of a serial mediation analysis (Figure 1), a total effect of PTSD on suicidal behavior was significant (model R2=0.290, c=0.096 [95% CI=0.090–0.102]). The direct effect of PTSD on suicidal behavior was reduced when 2 mediators, depression and NSSI frequency, were added in the model (c’=0.024 [95% CI=0.016–0.032]), indicating significant mediating effects. The total indirect effect of PTSD on suicidal behavior and specific indirect effects (a1b1, the indirect direct effect of depression alone; a2b2, the indirect effect of NSSI frequency alone; and a1db2, the indirect effect through depression and NSSI frequency in serial fashion) were all significant (Table 3). These indicate that 1) more severe symptoms of PTSD were associated with more severe depression, and in turn, leading to higher risk of suicidal behavior; 2) more severe PTSD symptoms were associated with more frequent NSSI, and in turn, leading to higher risk of suicidal behavior; and 3) more severe PTSD symptoms was sequentially associated with more severe depression symptoms and more frequent NSSI, and in turn, leading to higher risk of suicidal behavior.
Bootstrap results for indirect effects
DISCUSSION
This study shows notable findings about NSSI in a large, national sample of firefighters. First, the 1-year prevalence of NSSI was 4.67%. Second, there were significant correlates for the past-year NSSI in firefighters, such as sex, occupational status, current duty, recent traumatic experience, as well as PTSD and depressive symptoms. Third, when the participants engage in NSSI, the methods of self-cutting or carving and self-hitting were highly correlated with increased risk of suicidal behavior. Finally, the frequency of NSSI mediates the pathway from PTSD symptoms to suicidal behavior.
The 1-year prevalence of NSSI in Korean firefighters estimated in this study (4.67%) is much higher than in adults in the United States (US) (0.9%) [5]. It is rather close to the lifetime prevalence of NSSI among nonclinical adult population (5.5%) reported in a previous meta-analysis study [36]. The 1-year prevalence of NSSI in Korean firefighters in this study is also higher than that in US military personnel; the 1-year prevalence of NSSI was 3% in 335 US military personnel and veterans [37] and 1.2% in 21,449 US soldiers [38]. In addition to a high 1-year prevalence of NSSI in Korean firefighters, it is noteworthy that the rate of NSSI does not decrease significantly with increasing age in the current study. Although NSSI is known to have the highest prevalence in adolescence and decrease with age [4], aging did not have a significant effect on lowering the prevalence of NSSI in firefighters. In this study, the 1-year prevalence of NSSI was the highest in those in their 20s (6.32%); however, the increase in age was not associated with a significant decrease in the 1-year prevalence of NSSI among firefighters (4.56% in 30s; 3.86% in 40s; and 3.80% in 50s). This result suggests that firefighters may have cumulative occupational risk factors for NSSI.
Among the demographic characteristics, female gender was found to be significantly associated with the 1-year prevalence of NSSI in the participants; the rate of past-year NSSI in female participants was 6.91%, while that in male participants was 4.34%. This finding is consistent with the result in a previous study, which used meta-analysis to examine the presence of gender difference in the prevalence of NSSI; it reported a higher prevalence of NSSI in females than in males [39]. Among occupational factors, it was found that social service agents showed a higher association with NSSI than regular firefighters. Social service agents in fire stations are men on alternative military service with firefighters, since they are unfit to participate in mandatory military service in South Korea, due to their physical and/or mental health disabilities. Thus, among them, there might be more individuals with mental problems compared with regular firefighters. In addition, this study showed a lower association between EMS and past-year NSSI compared with fire suppression. Even though the 1-year prevalence of NSSI in EMS (4.8%) was higher than that in fire suppression (4.3%), EMS became less likely to be associated with the past-year NSSI according to a logistic regression analysis after controlling for other variables, such as age, sex, and recent traumatic experience. This result suggests the need to explore reasons why firefighters with a duty of fire suppression are more likely to engage in NSSI behaviors than those with a duty of EMS. A follow-up study is needed to explore how stress, related to current duty, in firefighters may influence NSSI behavior and mental health problems.
This study also shows the characteristics of NSSI in firefighters. Compared with the well-known characteristics in adolescent NSSI, those in adult NSSI have been less investigated. This study not only shows the characteristics of NSSI in firefighters, but also broadens the understanding of NSSI in the adult population in general. Specifically, when engaging in NSSI behavior, an average of 1.9 methods are used, and the most commonly used methods of NSSI in this study were biting (34.21%), scraping skin (34.79%), pulling hair out (23.39%), tattooing (21.56%), and self-hitting (20.22%). Serious methods of NSSI [27], such as cutting or carving (8.16%) and burning skin (4.91%), were not often used by Korean firefighters. Also, among the 11 methods, cutting or carving, hitting, pulling hair out, wound picking, biting, and skin scraping were significantly associated with high risk of suicidal behavior, while tattooing was associated with low risk of suicidal behavior. Even though tattooing is one of the self-injury behaviors and has functioned to regulate negative affect [40], it also reflects self-care and, to some degree, protection against greater self-harm [41]. According to cultures, tattooing is a way to achieve self-empowerment or take control of the body [42,43]. However, a recent study found that adolescents engaging in NSSI via tattooing were more likely to meet the DSM-5 criteria, practice more than one method of self-injury behaviors, and have more social dysfunction compared with adolescents engaging in NSSI without tattooing [44]. It supports the necessity to assess tattooing as a type of NSSI. Further studies are needed to investigate the characteristics of adult tattoo users and the association among tattooing, psychological distress, and suicidal behavior in adults.
Finally, this study found that NSSI partially mediated the association between PTSD symptoms and the risk of suicidal behavior. In the results of a serial mediation analysis, PTSD symptoms were associated with suicidal behavior by the four pathways: 1) direct effect; 2) indirect effect via higher levels of depression symptoms; 3) indirect effect via higher levels of depression, and sequentially, higher frequency of NSSI; and 4) indirect effect via NSSI frequency. It is well known that psychiatric symptoms, such as depression and anxiety, may cause NSSI [45-47]. In addition, previous studies showed that frequent NSSI is associated with frequent suicidal thoughts and attempts [48,49]. Taken together, it is suggested that the frequency of NSSI may mediate the association between psychiatric symptoms, including anxiety and depression, and suicidal behavior. However, the role of PTSD, depression, and NSSI on suicidal behavior has been studied independently without investigating the relationship among them. The current study showed that the frequency of NSSI have an indirect effect on the association of PTSD with suicidal behavior in two ways: 1) frequent NSSI related to PTSD symptoms is associated with higher risk of suicidal behavior in firefighters and 2) PTSD symptoms are related, serially, to more severe depression symptoms, and in turn, leading to more frequent NSSI behaviors and higher risk of suicidal behavior.
The primary purpose of engaging in NSSI appears to be alleviation from emotional pain, such as anxiety or depression [50]. By definition, NSSI is not directly related to patients’ suicide attempts; however, risk of suicidal behavior increases once NSSI loses its coping effectiveness, leading to the perception that suicide is the only option [50]. A growing body of research has reported increased pain threshold and tolerance among individuals who engaged in NSSI [51]. According to the interpersonal-psychology theory of suicide, it is hypothesized that increased pain tolerance related to NSSI could increase the acquired capability to commit suicide [25,26]. The previous studies have supported the hypothesis by showing that individuals with a history of frequent NSSI showed increased suicidal capability and decreased pain sensitivity [24,52]. Also, NSSI is one of the sequelae of prolonged exposure to trauma and associated with posttraumatic stress symptoms [53]. Taken together, our findings could suggest the need for early detection of NSSI as a mediator and thereby prevention of suicidal behavior in firefighters suffering from PTSD.
To the best of our knowledge, this study is the largest to evaluate NSSI with the inclusion of the largest and most representative sample of firefighters to date. However, it has several limitations to consider when interpreting and generalizing the results. First, since this study was conducted by cross-sectional analyses on the association among psychiatric symptoms, NSSI, and suicidal behavior, it cannot provide information about the casual relationship among the variables. Second, this study employed a retrospective survey, thus, it is subject to recall biases of the participants’ perspective on their past behaviors. Third, this study adopted a self-reported questionnaire. Although the results of the questionnaires did not lead to any disadvantages for the participants, psychiatric symptoms may have been under-reported due to concerns about confidentiality and social stigma. Further research using more objective methods, such as structured interviews and longitudinal designs, is needed in the near future.
In conclusion, this study found a relatively high 1-year prevalence of NSSI in firefighters. Female gender, occupational status of social service, current duty of fire suppression, recent exposure to traumatic events, and high levels of PTSD and depression symptoms were correlates of NSSI behavior. Furthermore, there was an association between PTSD and suicidal behavior, partially mediated by depression and NSSI frequency, in addition to the direct effect of PTSD on suicidal behavior. The present findings have an important clinical implication for proposing the screening and early intervention of NSSI in firefighters with PTSD to prevent suicidal behavior.
Supplementary Materials
The online-only Data Supplement is available with this article at https://doi.org/10.30773/pi.2023.0006.
The list of traumatic events
Notes
Availability of Data and Material
The dataset used in this study belongs to the National Fire Agency, and it cannot be used or shared without the agency’s permission. If necessary, please contact the corresponding author.
Conflicts of Interest
The authors have no potential conflicts of interest to disclose.
Author Contributions
Conceptualization: Heyeon Park, Beomjun Min, Jeong-Hyun Kim. Formal analysis: Heyeon Park, Sohee Oh, Jonhanna Inhyang Kim. Writing—original draf: Heyeon Park, Jeong-Hyun Kim. Writing—review & editing: Beomjun Min, Hankaram Jeon, Jeong-Hyun Kim.
Funding Statement
None