Rate of psychiatric disorders and associations with quality of life among community members following the Kaohsiung gas explosion: an 18-month cross-sectional follow-up study

A major petrochemical gas explosion occurred in the Lingya District, Kaohsiung City, Taiwan. That event is known as the “81 petrochemical gas explosion of August 1, 2014.” [1] The incident caused traffic disruptions, casualties, and environmental damage [1, 2]. Three hundred seventy-two casualties were acknowledged as of October 6, 2014, including 32 deaths and 340 serious injuries [1]. This highly stressful event had a severe impact on victims, especially in terms of their mental health.

Many studies [3‐7] have shown evidence of psychiatric impairment in disaster survivors. The most common disaster-related psychiatric diagnoses are major depressive episode(MDE) and posttraumatic stress disorder(PTSD), which are strongly associated [3, 8]—an issue that continues to gain attention in trauma outcome research. [3] The prevalence of psychiatric diseases, especially PTSD and MDE, is highest at the beginning of a disaster and gradually declines after mental rehabilitation. In addition, rescue workers such as nurses, fire fighters, and soldiers incur a high prevalence of psychiatric disorders after disaster rescue.

Physical and psychological trauma after various disasters has been shown to significantly impact the quality of life of survivors of various disasters, and physical damage is strongly related to psychological trauma [1, 2, 9‐16]. However, three symptoms of PTSD—namely, re-experiencing, avoidance and numbing, and hyperarousal—are common and may be dismissed as unimportant behavioral shifts, or some physiologic reactivity at exposure to cues that symbolize trauma was under-reported [15, 17, 18]. For example, when a person hears the sound of firecrackers, he or she might be too scared to go out the door. Another might have “hypervigilance”, which is a symptom of hyperarousal syndromes. Although the prevalence of PTSD tends to decrease over time among burn survivors, rates of PTSD have been found to increase over time from injuries due to delayed onset or repercussions from the traumatic event (i.e., job loss, health problems) [15, 19]. Therefore, regardless of the mental affliction involved, early detection and longer periods of effective treatment are most important after a disaster.

We know that physical and mental health are interrelated [10] and that physical injuries and material losses are also strongly associated with PTSD [13, 14]. Long-term financial problems are one of the many important factors affecting quality of life (QoL) [8, 12, 20, 21]. Health-related quality of life (HRQoL) includes both physical and mental dimensions [22]. A composite measure of physical and psychological health is HRQoL [19], as assessed by short form-12 or short form-36. Following a disaster, psychological problems have a greater impact on quality of life than injuries do [23]. We know that physical and mental healthcare are interrelated [10] and that physical injuries and material losses are also strongly associated with PTSD [13, 14]. These two dimensions are included in health-related quality of life (HRQoL) [22]. A composite measure of physical and psychological health is HRQoL, as assessed, for example, by short form-12 or − 36 [23]. Experiencing long-term financial problems is one of several important factors affecting quality of life [10, 12, 20, 21]. Through this continuous cycle, trauma may become a personal problem that evolves into a family problem. Therefore, early detection and assistance can reduce the incidence of such problems and improve the QoL of participants and individuals. Furthermore, long-term follow-up is necessary for mental rehabilitation. In the face of a catastrophic disaster, many people may experience psychiatric impairment and poorer quality of life, making it important to give them support, such as mental health rehabilitation, to decrease later sequalae. Mental health rehabilitation involves provision of both economic and health information, lifestyle modification, exercise, emotional support, psychiatric services (including, inter alia, drug treatment and psychotherapy), and stress/adversity management, among other factors. The health organization and study team carried out disaster mental health rehabilitation after the 81 petrochemical gas explosion and simultaneously established a disaster psychological assistance group with the support of the Kaohsiung City government. We performed “walk around” care (a mental health term that denotes active community outreach instead of traditional OPD to offer psychiatric services and mental health information) to provide the victims of the event with emergency mental health services.

A support system is known to be particularly important after a disaster [13]. Accordingly, we conducted a series of studies to investigate the rate of and associated factors for probable disaster-related psychiatric disorders such as MDE, PTSD, and quality of life (QoL) after six [1] and 12 months [2]. In total, 7.4–13.9% (6 months) [1] and 2.9–10.3% (12 months) [2] of individuals with probable PTSD and/or probable MDE showed a reduced quality of life. The reconstruction of one’s life after a disaster can be a challenging process. Mental rehabilitation is a part of this reconstruction process and requires a planned and comprehensive approach [16]. Our research hypotheses are: (1) disasters would increase psychiatric impairment and decrease quality of life in disaster survivors, highlighting the importance of mental rehabilitation, which consists of both short-term intervention and long-term follow-up and referral of cases to regular psychiatric treatment; and (2) that individuals would also benefit from mental rehabilitation [24, 25]. After a disaster, PTSD can have severe and painful effects and must be addressed with appropriate treatment and assistance to enable rapid screening so that victims can receive an early diagnosis and long-term complications can be prevented. Persistent follow-up is necessary to understand the status of the affected after a disaster and to strengthen their support systems. Subsequently, it is important to continue to track the needs of those affected after a disaster to help provide relevant information so that traumatized victims can return to society as quickly as possible. The purpose of this study was to investigate the prevalence of psychiatric disorders (PTSD and MDEs) and associations with QoL in individuals who were exposed to the petrochemical gas explosion, either directly or indirectly.

After 18 months, only community dwellers were included. Ultimately, 388 participants were included (average age: 43.27 ± 15.98 years; males: 203, average age: 41.44 ± 15.74 years; females: 185; average age: 45.27 ± 16.03 years). Overall, 331 of the 388 respondents (85.3%) were assigned to the non-P or non-M group. For a comparison of the rates of PTSD and MDE at the three time points, we also used demographic data obtained in two previous studies as a representative sample of 502 participants (average age: 42.90 ± 16.61 years; M: 270, average age: 40.89 ± 16.40 years; F: 232; average age: 45.25 ± 16.58 years) 6 months after the event and a representative sample of 486 participants (average age: 42.89 ± 16.05 years; M: 255, average age: 40.68 ± 15.92 years; F: 231; average age: 45.32 ± 16.20 years) 12 months after the event (please see others in Table 1).

Fifty-seven of the 388 respondents (14.7%) had at least one probable psychiatric diagnosis and were assigned to the following groups: probable PTSD (n = 34) [8.8%], probable MDE (n = 14) [3.6%], and probable PTSD and MDE (n = 9) [2.3%]. We found that respondents who survived the postgas explosion differed significantly from those who did not experience the event in terms of financial problems. Additionally, financial problems were significantly more frequent in the probable PTSD and MDE group than in the non-P or non-M group (p < 0.001) (Table 2).

The only factor found to be significantly associated with probable psychiatric disorder in respondents was financial problems, which showed a significant positive association with probable psychiatric disorders. The association of probable psychiatric disorders was greater in those with financial problems (OR = 6.344; p < 0.001) than in those without financial problems (Table 3).

The QoL scores for the RP, BP, VT, RE, MH and MCS in the non-P or non-M group were significantly higher than the scores of the different probable psychiatric impairment groups (p < 0.001). In addition, the QoL scores for the PF, RP, GH and SF in the non-P or non-M group were significantly higher than the scores of the probable PTSD and the probable PTSD and MDE groups (p < 0.001). Furthermore, the QoL scores for the PCS in the non-P or non-M group were significantly higher than those in the probable PTSD group (p < 0.001). Finally, the QoL scores for the PF, RE, MH and MCS in the probable MDE group were significantly higher than the scores of the probable PTSD and MDE group (p < 0.001) (Table 4).

As shown in Table 5, most subscales were worse for older victims than for younger ones. In addition, QoL scores on the PF (p = 0.003), RP (p = 0.028), GH (p = 0.001) and PCS (p < 0.001) subscales among victims with mild or moderate physical injuries were worse than those among victims with no physical injuries. Additionally, most subscales among victims with severe physical injuries were worse than the scores among those with no physical injuries. However, QoL scores on the RE (p = 0.018) subscales among victims who had an occupation were better than those among victims who had no occupation. Moreover, QoL scores on the PF (p = 0.022), BP (p = 0.006) and PCS (p = 0.019) subscales among victims with a college or higher educational level were worse than those among victims with a junior high school or below educational level, although QoL scores on the GH (p = 0.008) and PCS (p = 0.023) subscales were better than those among victims with a junior high school educational level. QoL scores were worse among victims with financial problems on the most subscales. Additionally, QoL scores (all p < 0.001) at the PF, RP, BP, GH, VT, SF, RE, MH, PCS and MCS subscales were worse among victims with probable MDE. Finally, among victims with probable PTSD, QoL scores were worse on the most subscales.

After the gas explosion, the Kaohsiung City government immediately organized a psychiatric team to offer mental health services to residents and victims, including psychiatric services for people with psychiatric impairments (including, inter alia, drug treatment and psychotherapy) and a mental health promotion program that included lifestyle modification and stress/adversity management, among other services, for community residents. Comparing the study outcomes at six [1] and 12 [2] months reveals a slightly decreasing trend in the rate of PTSD. The probable MDE prevalence was 7.4, 2.9, and 3.6%. A decreasing trend of MDE prevalence was observed at 1 year and beyond. The probable prevalence of MDE and PTSD was 13.9, 2.9, and 2.3%. The trend is similar to that of probable MDE. A comparison with previous studies [1, 2] and other disasters [16] in Taiwan showed a high rate of psychiatric diseases within 6–12 months and a slight decrease after 12–36 months. The current research identified that those exposed to the incident reported similar levels of psychological trauma and the presence of associated financial problems, similar to levels reported at 6 and 12 months after the event. [1, 2]. However, there are some associated factors for probable PTSD or MDE, including being female and having experienced a physical injury, at 6 months [1] and 12 months [2]. After a disaster, financial difficulty is highly associated with psychiatric morbidity; in turn, psychiatric morbidity might be the source of financial issues, a finding that is consistent with previous studies on different disasters [1, 2, 11, 13, 21]. Most victims were able to recover from their physical injuries, and their physical functioning was restored. However, the recovery period for psychological trauma is longer than that for physical injuries, causing financial problems and affecting QoL.

When we compared the QoL of each of the different groups among the postgas explosion respondents (Table 3) with that of the respondents after 6 and 12 months in two previous studies [1, 2], we obtained the following results: first, non-P or non-M QoL scores were still higher than those associated with a psychological trauma diagnosis. Second, the PCS scores associated with the non-P or non-M QoL scores were higher than those associated with the probable PTSD scores and probable PTSD and MDE. Third, PF, RE, MH and MCS subscale scores in non-P or non-M individuals were higher than those in probable PTSD and MDE individuals. Among them, PF affected the 6- and 18-month QoL scores. Disasters can cause physical and psychological harm. Extremely stressful events can have a significant effect on mental health, and PTSD and MD are common sequelae [1, 2, 9, 10, 16, 23, 24]. Wu et al. [10] confirmed that physical and mental health in postearthquake victims are highly interrelated. Concurrent psychopathology symptoms, physical illnesses and long-term financial problems are associated with QoL [1, 2, 10, 12, 13]. Thus, we infer that two or more psychiatric diagnoses result in family dysfunction, financial problems, and the emergence of poor QoL scores.

As shown in Table 4, we found that age, physical injuries, financial problems, probable MDE and probable PTSD were significantly negatively correlated with QoL. However, occupation and educational level were significantly positively correlated with QoL. Compared with the results of [1, 2], there were significant negative correlations between MDE and QoL. Next, age had no significant correlation with QoL. Finally, there was no significant correlation between current marital status and religious belief for either sex at any time point. Many studies have found that after a disaster, QoL is negatively correlated with aging, physical illness, financial problems and PTSD/MDE [10, 12, 28]. Furthermore, the persistence of long-term financial problems was one of many important factors affecting QoL [1, 2, 10]. Thus, a lower education level and no occupation will affect disaster victims in terms of financial support to a greater extent than it will affect others [11]. Interestingly, material losses and physical injuries were positively associated with the later development of psychiatric impairments [13]. However, Wu et al. [10] found that older victims’ QoL scores were lower, especially in the physical dimension. Therefore, these findings suggest that aging, severe physical injuries, and a lower educational level in the physical dimension are important negative factors. No employment is correlated with financial problems. Thus, levels of physical injuries and financial problems are correlated with the development of psychiatric impairment. Once again, it is confirmed that regardless of the time postdisaster, physical and psychological factors affect QoL, and early detection of problems and longer provision of assistance are necessary.

Our research hypotheses were that disasters highlighted the importance of mental rehabilitation, which consists of both short-term intervention and long-term follow-up and referral of cases to regular psychiatric treatment, and that affected individuals would benefit from mental rehabilitation. [24, 25] We believe that this study has the following strengths: 1. We provided mental rehabilitation after a major disaster, allowing professionals to offer early intervention and provide continuous follow-up and care or medical treatment. 2. We provided disaster epidemiological data, including associated factors, for public health. However, we found that although the number of psychiatric diseases had decreased compared to that 12 months after the disaster [2], the number of probable MDE cases did not change. Additionally, this study has the following limitations: Those who have experienced a gas explosion may be reluctant to consciously recall the situation at that time without any problems and refuse to be visited. This issue may result in an underestimation of the severity of participants’ psychiatric conditions or an underestimation of the proportion of participants suffering from mental illness. We did not collect information about the participants’ prior mental health histories, which might be a confounder for this study. We also used screening scales instead of employing psychiatrists to survey the participants’ psychiatric disorders, which require more comprehensive mental health assessments that were not able to be undertaken. Because the disaster studied is a man-made disaster (which made it difficult to conduct a well-designed study), the authors acknowledge the potential impact of multicollinearity in the current study, we did not use an adjusted p value due to the important clinical implications of the research. Furthermore, we simply used a global “distress” variable instead of PTSD or MDE because of the lack of an adequate number of psychiatrists to provide diagnoses.

Early intervention was performed in the acute postdisaster phase, resulting in a mild decrease in the number of cases of mental illness 18 months later. Therefore, mental rehabilitation should continue for a longer period. There are still many participants who were profoundly affected by the event, and their QoL was also affected, especially among those with probable PTSD and MDE. However, we also used a limited study design and found a sufficient number of people impacted by the disaster who were willing to participate in our study. In the future, researchers could compare subgroups, focusing on exposure and on those who developed psychiatric morbidity versus those who did not. In addition, we believe that long-term follow-up and regular psychiatric management are needed to help participants successfully reconstruct their lives.