Prevalence and Predictors of PTSD during the Containment Stage of COVID-19 Epidemic among College Female Students in China

Background College students are a uniquely vulnerable group and may experience high stress levels due to COVID-19, especially for girls. This study aims to identify the post-traumatic stress disorder (PTSD) symptoms and related factors among the target population during the initial phases of the COVID-19 pandemic. Methods A cross-sectional online survey was conducted during the initial phases of the COVID-19 pandemic in China. A total of 2205 college female students from six provinces enrolled in this study and completed the questions about cognitive status of COVID-19, the Impact of Event Scale-6 (IES-6), the Multidimensional Perceived Social Support Scale (MPSSS) and a self-developed 10-item Perceived threat scale. Univariate and multivariate logistic regression were performed by SPSS software to explore the determinants of PTSD symptoms. Results PTSD symptoms were prevalent in this sample of college female students, and 34.20% met the cut-off for PTSD. Self-reported fair or poor health (AOR=1.78, 95%CI: 1.22-2.59), high concern about COVID-19 (AOR=1.66, 95%CI: beliefs that 'COVID-19 can cause a global outbreak' (AOR=1.26, 95%CI: 1.02-1.56), the perception of ‘risk of infection’ (AOR=2.46, 95%CI: 2.16-2.81), beliefs that ‘closed management’ and ‘COVID-19 as a public health emergency of international concern’ would have an impact, and the fear of ‘impact on life planning’ were all positively associated with PTSD (AOR=1.37, 1.22 and 1.29, respectively), whereas perceived social support from family (AOR=0.81, 95%CI: 0.70-0.93) was negatively associated with PTSD. Among the signicant variables at the bivariate level, multivariate logistic regression revealed that the greatest protector for PTSD was the high knowledge score (AOR=0.73, 95%CI: 0.60-0.90), while had conrmed cases among relatives and on female college students and identify the PTSD symptoms and related factors of this vulnerable group during the initial phases of the COVID-19 pandemic through a multi-center survey. The results of this study may assist government and healthcare agencies in developing interventions by better understanding the causes of adverse psychological outcomes to alleviate stress in college students, especially for female students in the event of future outbreaks of infectious diseases.

The COVID-19 pandemic has caused enormous psychological impact worldwide [4][5][6]. Globally, relatively high rates of symptoms of post-traumatic stress disorder (PTSD) (7-53.8%) were reported in the general population during the COVID-19 pandemic, especially in college students [6][7][8][9][10]. Exposure to such stressful events can lead to the development of acute stress disorder and nally PTSD, if symptoms persist. At the end of February, college students in China were noti ed that the start of school would be postponed. They could not resume school at the normally scheduled time, their regular routines of study were disrupted, and at the same time they were youths who were still developing self-discipline and emotional control [11]. In a study among college students in mainland China during the beginning of COVID-19 outbreak, 30.8% (95%CI: 28.8 to 32.8%) of the participants completing the survey presented clinically relevant PTSD symptoms [9].
Females are usually more likely to have mental health problems than males when they are facing stressful events. Due to sex hormones such as oestradiol and progesterone and the life trajectory they experience, they always worry about interpersonal stressors, gender-based violence, lack of gender equality and discrimination compare to the similar-age male [12,13]. Studies showed that the home isolation due to COVID-19 was associated with a worse psychological status with a stronger association in women in comparison to men [14]. PTSD was especially one of the most prevalent long-term psychiatric disorders for females when they were facing traumatic events [15][16][17]. Previous studies indicated that acute psychological disorders and PTSD symptoms were more prevalent in females than males after traumatic events [18][19][20]. A study conducted in South Lebanon revealed that females were twice in the prevalence of PTSD as likely as males after experienced the 2006 war (24.3 vs. 10.4%) [19]. Similarly, among Indian females and males affected by ash oods in 2010, 22.3% and 18.1% met the diagnostic criteria for PTSD, respectively [20].
Given the dispositional traits of female and situational in uences of COVI-19 to the university students, this study aims to focus on female college students and identify the PTSD symptoms and related factors of this vulnerable group during the initial phases of the COVID-19 pandemic through a multi-center survey. The results of this study may assist government and healthcare agencies in developing interventions by better understanding the causes of adverse psychological outcomes to alleviate stress in college students, especially for female students in the event of future outbreaks of infectious diseases.

Settings and participants
We conducted a cross-sectional survey from February 23 to March 5, 2020, when the COVID-19 outbreak showed a downward trend of new cases in China and an upward trend overseas. Due to the isolation and social distancing measures recommended by the Chinese government, the questionnaire survey and corresponding data were collected electronically via the internet. In mainland China, potential participants from six universities (University of Jinan, Sichuan University, Sun Yat-sen University, Shanxi University of Chinese Medicine, Henan University of Engineering and Dalian University of Technology) located in different geographical territories were electronically invited by partner teachers in each university.
The study was approved by the Ethics Committee of Jining Medical University of Shandong Province on February 12, 2020. To ensure the authenticity and representativeness of the survey, our research was based on voluntary participation without any compensation measures, and potential participants provided electronic informed consent if they were willing to participate in the questionnaire survey voluntarily. The participants have to meet the following criteria: (1) college female students; (2) aged ≥ 18 years; and (3) pro ciency in Chinese. After passing the quali cation review and answering yes to the participation question, they were directed to complete the self-reported questionnaire.

Measurements
The structured questionnaire was designed to encompass the following areas: (1) demographic characteristics such as age, major and health status,etc. (2) variables related to COVID-19 (e.g. awareness of COVID-19, levels of concern to the outbreak), (3) daily perceived social support, (4) perceived threat associated with COVID-19, (5) PTSD. PTSD, perceived threat, and perceived social support were measured as outlined in the following.

Impact of Event Scale-6
The Impact of Event Scale-6 (IES-6) is an abbreviated 6-item version of The Impact of Event Scale-revised, which has been widely accepted as a core outcome measure for PTSD in critical illness survivorship research [21,22]. Respondents are asked to report the psychological impact after exposure to a crisis situation within 7 days of exposure using a 5-point Likert scale (0 = not at all; 4 = extremely), and PTSD is measured when the score of IES-6 is greater than or equal to 10 [21]. In the present study, the Cronbach's α was 0.84.

The Perceived threat scale
To measure the extent of threat to which participants perceived from COVID-19 pandemic, the 10-Item perceived threat scale was developed. The items were based on earlier studies [23][24][25], each item can be rated on a scale from 1 (Strongly disagree) to 5 (Strongly agree). Exploratory factor analysis on the 10 items of perceived threat yielded three dimensions: Risk of infection, Impact on life planning, and Impact of PHEIC. The three dimensions of the perceived threat scale were proved to have satis ed content and concurrent validity. Sample items includes: 'I feel anxious and scared when thinking of COVID-19'. In the present study, the Cronbach's α for the three dimensions were 0.83, 0.84 and 0.88, respectively (See Appendix 1 and 2).

The Multidimensional Scale of Perceived Social Support (MSPSS)
The 12-item perceived social support scale (MSPSS) assesses a respondent's perceived support on a 7point Likert scale (1 = strongly disagree; 7 = strongly agree) [26]. Total possible scores range from 12 to 84, with higher scores indicating higher levels of perceived support obtained from family, friends and signi cant others. The Chinese version of the MSPSS has adequate internal consistency in eariler study and in present study (Cronbach's α = 0.89 and 0.94, respectively) [27]. The Cronbach's α for family, friends and signi cant others were 0.88, 0.92 and 0.87, respectively.

Statistical Analysis
The data were analyzed using SPSS 23.0. A descriptive analysis of the demographic data was conducted to describe the sample characteristics. Categorical variables were expressed as frequencies and percent distributions, while continuous variables were presented as means ± standard deviations (SD). The prevalence of PTSD was derived according to the cut-off values routinely used in previous studies.
Bivariate generalized linear modeling with a binary logistic distribution was conducted to examine background and other variables related to PTSD. Variables that were signi cant at the bivariate level (P < 0.1) were adjusted in further multivariate analysis for factors related to PTSD. A value of P < 0.05 was considered statistically signi cant in multivariate analysis. The results were reported as odds ratios (ORs) and adjusted odds ratios (AORs).

Background characteristics
A total of 2205 individuals participated in the questionnaire survey with a mean age of 20.85 years (SD, 1.50 years; range, 18-26 years). Among participants, 96.78% were Han Chinese, 60.09% were rural residents, 45.90% were majored in medicine, 1.86% ever had at least one of the chronic diseases (e.g. diabetes, cardiovascular diseases) and 93.15% reported they have good health status. The mean score of perceived support from family, friends and signi cant others was 5.65, 5.60, and 5.49, respectively.
Among the participants, 41.54% had been quarantined or isolated due to the COVID-19 pandemic, 3.76% had close relatives and friends in key epidemic areas. 85.62% reported there were con rmed cases in their current living city, 3.81% in their current community or village, and 0.41% among their relatives and friends (Table 1).   Table 3 presents the knowledge score, awareness and perceived threat of COVID-19 among participants. 51.80% had a knowledge score ≥ 11 (Mean score of participants). 85.80% of the respondents were aware of the news that the WHO de ned COVID-19 as a PHEIC on January 30, 42.31% were highly concerned about the epidemic, and 31.66% believed that COVID-19 would cause a global outbreak. In addition, 50.88% thought the closed-off management had impacted their daily lives. Regarding perceived threat of COVID-19, the mean score of risk of infection, impact on life planning and impact of PHEIC was 1.88, 3.10, and 3.26, respectively.

Levels of concern
Less concern 1272 57.69 High concern 933 42.31 Those who majored in non-medical (OR=1.23, P=0.02) and self reported in fair or poor health (OR=2.10, P=0.00) were positively associated with PTSD, whereas perceived social support from family was negatively associated with PTSD (OR = 0.85, P = 0.00). 'Ever been quarantined or isolated' was positively associated with PTSD (OR=1.21, P=0.04). Having close relatives and friends in key epidemic areas (OR=1.49, P=0.08) and having con rmed cases among relatives and friends (OR=7.10, P=0.02) were positively associated PTSD, whereas having high knowledge of COVID-19 was negatively associated PTSD (OR=0.76, P=0.00). High concern about the epidemic was positively correlated with PTSD (OR=1.57, P=0.00), and those who believed COVID-19 would cause a global outbreak was signi cantly associated with PTSD (OR=1.79, P=0.00). In addition, beliefs that 'Impact of closed management on life' and perceived threat of COVID-19 in the three sub-dimensions were all positively associated with PTSD (Table 4).

Multivariate correlates of PTSD
A multivariate analysis controlling for all signi cant variables in the bivariate analysis was also undertaken ( Table 5). Self-reported fair or poor health (AOR=1.78, P=0.00) and had con rmed cases among relatives and friends (AOR=7.70, P=0.03) were positively associated with PTSD, whereas perceived social support from family (AOR=0.81, P=0.00) and had high knowledge score (AOR=0.73, P=0.00) were negatively associated with PTSD.
Concerns that 'risk of infection', 'impact on life planning' and 'impact of PHEIC' were all positively associated with PTSD (AOR=2.46, 1.29 and 1.22, P=0.00). Major in school, ever been quarantined or isolated and had close relatives and friends in key epidemic areas were not related to PTSD at the multivariate level.

Levels of concern
Less concern 1 High concern 1.66 1. 35 overseas studies, the prevalence of PTSD symptoms in this study was relatively high, nearly one-third of participants met the cut-off for PTSD [9,10,28,29]. This might be partly due to the timing of the study, which was carried out at the early stages of the COVID-19 outbreak. At that time, the overall understanding to the disease, and the corresponding measures of prevention and control to the epidemic was still unclear. In addition, females are more prone to acute stress during major public health emergencies, such as SARS and earthquakes [30,31]. Furthermore, the international in uence of the COVID-19 pandemic and its impact on people's lives are unprecedented, hence the short-term response of females is so overwhelming that they were likely to develop ASD, and eventually PTSD symptoms if it persists for a period of time. Suffering from PTSD will undoubtedly affect female college students' physical and mental health, academic performance and work plan, and even their future lives. Therefore, identifying risk factors at the beginning of the epidemic, offering targeted psychological counseling, providing appropriate psychological support and theory-based interventions during the rehabilitation process is particularly important.
People with fair or bad health were not only susceptible to the COVID-19 infection (especially with the lack of proper protection) but also had a poor prognosis once diagnosed. The study results were in line with the previous study that PTSD symptoms were remarkably high among those who perceived themselves not very healthy [32]. During the closed-off period, unhealthy participants might have the need to seek medical treatment or other health services, however, it may become a little di cult due to the tra c control and shortage of medical care in this extraordinary situation [33,34] With the absence of medical care and medical supply during the epidemic, these unhealthy population were at an increased risk of PTSD symptoms [35,36]. Therefore, in the COVID-19 era, health o cials should carefully weigh risks and bene ts when planning medical needs of vulnerable population and chronic patients during the containment period. Basic medical services for these people should be ensured and thus releive their mental stress to avoid the extra burden of other diseases apart from the epidemic.
Consistent with the González-Sanguino C's study, our results revealed that those who had con rmed cases among relatives and friends were prone to have PTSD symptoms, which was found to be one of the strongest predictors of adverse psychological impacts during pandemic of infectious diseases [17].
When something stressful and terrifying happens around someone, they may fell frightened and helpless since they are themselves positively affected by the event [37,38]. Therefore, they are likely to go into the 'crisis phase' which is characterized by crying, stressful, insomnia and nightmares, and eventually develop symptoms of PTSD if conditions persist [30,39,40]. This reminds us that we should identify and focus on female college students who have COVID-19 patients among close relatives and friends, guide them to adjust their mentality, divert attention, and provide social support to encourage them to re-plan their lives under existing conditions [41].
Additionally, our self-developed COVID-19-related perceived threat scale uniquely predicted PTSD symptoms among young females, underscoring how the speci c features of this pandemic give rise to acute and post-traumatic stress. Participants were not only worried about accidentally getting infected by 2019-nCoV, but also worried that their work and life arrangements would be affected by the growing epidemic. Long-term accumulation of adverse emotions would nally lead to PTSD symptoms including intrusion, avoidance and hyper-arousal, if symptoms persist [32,40]. There is an urgent need to understand the mechanisms behind their potential PTSD symptoms and formulate speci c interventions accordingly. First, the o cial media, schools and teachers should strengthen the publicity of preventive measures to reduce their perceived infection risks; second, the network structure of online teaching and job recruitment should be improved to alleviate the target population's concerns about future plans, and to ensure their normal paces of study and work; nally, the government should show con dence to the public in coping with the economic downturn due to the pandemic and provide appropriate subsidies to minimize the impact of the epidemic on individuals, such as measures the UK and Australia adopted [42,43].
In current study, most of the respondents had high levels of social support, and the support from family, but not from partner or peers was negatively associated with PTSD symptoms, the same trend had been found among U.S. young adults during the COVID-19 pandemic [28]. College students were in home quarantine during the study period, substantial support and security might be important in maintaining the mental health of young adults because many were faced with acute stress and broken rhythm of life, such as holiday extension, nancial pressure, undergraduate and unemployment. Emotional support coupled with material security provided by family would be an important factor in protecting them against the risk of PTSD. In contrast, peers may have di cult to provide material and living security considering that they are also young adults experiencing similar struggles. Therefore, family support should be addressed in the whole society to guanrantee psychological wellbeing of college female students during the pandemic.
Another protective factor for PTSD symptoms found from this study was high level of knowledge on COVID-19. With easy access to the internet, the majority of respondents can get instant updates on news related to the COVID-19 outbreak, and these real-time effective information can help keeping students informed on the epidemic, the role of information seemed to be fundamental to cope with the epidemic mentally and physically. A negative relationship between su cient knowledge and adverse psychological impact of the pandemic was also found in researches conducted by Wang C and González-Sanguino C [15,16,44]. However, it should be cautious on the infodemic [45,46]. Social platforms play an important role in information acquisition and dissemination during the epidemic. Though the information from various sources and platforms are updated in real time, it lacks quality control sometimes. When we are surrounded by information noise, which may have impact on our emotions and behavior, such as the looting of Shuanghuanglian in the early stage of the epidemic in China, worries, fear and panic driven by false information may cause people to develop PTSD symptoms. Therefore, it is necessary to ensure that up-to-date, accurate and scienti c information can be obtained from reliable sources to avoid panic caused by insu cient and inaccurate information and knowledge.
This study has several limitations to acknowledge. First, the generalizability of our conclusions is impacted given the results were possibly due in part to nonrandom selection of the sample, the possibility of sampling bias should be considered. Second, the uncontrolled online observational study of college students and the cross-sectional nature of the data warrant further longitudinal studies to determine the predictors of PTSD symptoms during the outbreak. Furthermore, the different psychological scales used in this studies may lead to different measurement outcomes compare to other similar studies. Finally, this study relied on anonymous network responses and self-reported answers regarding experience during home-quarantine stay which may not align with clinical psychological diagnosis.
In summary, the ndings from this study provide a basis for implementing measures to improve the mental health among college female students during the outbreak of COVID-19 and have a reference value for identifying students who may have a high risk of adverse psychological problems. Our study has a number of strengths, including a large sample size and robust statistical analysis. Additionally, in order to verify the relationship between knowledge of COVID-19 and PTSD, we have designed a 14-item COVID-19-related knowledge and behavior questionnaire, including basic information about COVID-19, main symptoms, routes of transmission, and health behaviors to deal with COVID-19 infection. The results of the study demonsrated that the high-levle knowledge on COVID-19 is of most importance in protecting against the PTSD among female college students. Finally, taking the characteristics of college female students into account, we have developed the 10-item perceived threat scale with three dimensions, and its validity has also been con rmed in this research.

Conclusions
During the containment stage of the COVID-19 pandemic, this observational cross-sectional study revealed that more than one-third of participants suffered from PTSD symptoms. College female students who were unhealthy, highly concerned about the outbreak, had con rmed cases among relatives and friends, held an pessimistic attitude towards the epidemic and perceived threat of COVID-19 were all associated with PTSD symptoms. While proper knowledge of COVID-19 and perceived support from family were protectors for PTSD symptoms. Our ndings can provide a basis for further exploration and designing of psychological intervention approaches to reduce the risk of obtaining PTSD and to improve mental health of college female students during epidemics of infectious disease.