Risk perception and Emotion evaluation of Health care workers varied during different periods of COVID-19: a repeated cross-sectional research

Background Health care workers (HCWs) on the front line of the COVID-19 were facing greater risks than usual, and their perception of these risks could also be an important part of their anti-epidemic work. This study aimed to examine risk perception and negative emotions of during two periods of the COVID-19 and emphasized emotional intervention management of HCWs impacting the risk perceptions and self-protection and would provide plausible intervention points for the psychological aid under a stressful condition. Methods The current study was adapted a repeated cross-sectional research participated by 504 Chinese HCWs who were assigned to work at the current disease resistance line. The demographic information about gender, profession, and location were collected in the demographic questionnaire. Risk perception questionnaire was adapted for COVID-19 to assess risk perception and the Chinese version of emotional self-rating scale (PANAS) was used to evaluate HCWs’ negative emotions. Results Findings revealed the risk perception and negative emotions of HCWs varied in different gender, profession and the location, as well as in different periods of COVID-19. Besides, the levels of tension, fear, worry and risk perception were higher during Period 1 than they were during Periods 2 of COVID-19. Over the different periods, the predominated negative emotions of HCWs presented varied, but the positive relations to risk perception were consisted and could be a signicant predictor of risk perception. Worry was found to be closely related to and a signicant predictor of high-level risk perception in the period 1; whereas in the period 2 the predictive type of negative emotion was tension. signicance this research lies in its examination of risk perception and negative emotions of HCWs combating the COVID-19 during two periods of the pandemic. The ndings showed both risk perception and HCWs by COVID-19 of a signicant for risk of HCWs the great pandemic. of the

Worry was found to be closely related to and a signi cant predictor of high-level risk perception in the period 1; whereas in the period 2 the predictive type of negative emotion was tension.
Conclusion The signi cance of this research lies in its examination of risk perception and negative emotions of HCWs combating the COVID-19 during two periods of the pandemic. The ndings showed both risk perception and negative emotions of HCWs were affected by the COVID-19 and underscored the importance of negative emotions as a signi cant factor for risk perception of Chinese HCWs enduring the great challenge of pandemic. To ensure safety and prevent the return of the pandemics, it would be better to monitor the risk perception and negative emotions of HCWs and emphasize the risk protection behaviors.

Background
On February 12, 2020, the novel coronavirus was o cially named by the world health organization as COVID-19. Due to its rapid spread, the outbreak of covid-19 has not only caused widespread public health concern, but also caused great psychological stress to the public. The researchers assessed the burden of mental health brought by COVID-19, analytically showed that in Chinese samples, the medical personnel would bear the brunt of the risk of mental health problems, which affected their attention, cognitive function and clinical decision, and even endangered the patient due to medical negligence [1]. In addition, while saving lives, frontline health care workers (HCWs) were also faced with increasing workload and risk of infection. According to the report of National Health Commission of China, 1,716 HCWs were infected up to February 11, 2020, and 11 of them died up to February 24, 2020. Hence, HCWs on the front line of the epidemic resistance were facing greater risks than usual, and their perception of these risks could also be an important part of their anti-epidemic work.
Risk perception, a psychological term, refers to the individual's perception and understanding of various objective risks existing in the outside world, and emphasizes the in uence of the experience acquired by individuals from intuitive judgment and subjective feelings on cognition [2]. Notably, individuals' ability to perceive risks associated with the virus is important for rapid changes and adaptation of population behaviors as for self-protective measures [3]. Given the importance of human psychological and behavioral factors in staving off pandemics, it is crucial to assess psychological and behavioral responses to the situation and determine how perceived risk is impacted by the pandemic [4]. Combined with studies in other elds and relevant literature analysis [5][6][7], HCWs' risk perception refers to their knowledge, feelings and understanding of risk factors and risk characteristics in the health care profession. However, a clear de nition is still discussed and explored. Recently, a study on the in uencing factors of risk perception was participated by 608 nurses and concluded that risk perception of HCWs could be clustered into ve dimension: personal safety; physical function; occupational exposures; psychosocial concerns; organization safety; timing pressure [8]. The key constructs of HCW's risk perception included personal health risks, health risks to others, social isolation and acceptance of risks [9]. A high perception of risk can in uence the retention of HCWs within the workforce and their willingness to care for infected patients [10], particularly if they are concerned about their own and their family's health and safety from the risk [11]. In contrast, HCWs with very low risk perceptions may be noncompliant to protective behaviors such as vaccinations and personal protective equipment (PPE), increasing both their own risk as well as increasing the risk of propagation of nosocomial transmission within the hospital and community [12]. As we hypothesized, with regard to the particularity of health care profession and the uncertainty of the disease in patients, HCWs combating against the pandemic are undertaken the risk of occupational exposure combined with the busy working status and shift work, which undoubtedly render their body in a state of fatigue and busy for a long time, hence, HCWs' risk perception would be signi cant in uenced by the outbreak of COVID-19.
Against the backdrop of epidemics, a paucity of previous studies has assessed the risk perception of HCWs. Koh et al. [13] conducted a qualitative interview schedule and attained that suffering from the infectious diseases, such as, would signi cantly impact the nurses' risk perception. The latest study of Diego et al. conducted in Latin American HCWs showed HCWs perceived limited access to protective equipment and supports, which indicated the organizational safety were also added to the epidemicrelated risk for HCWs [14]. However, the mechanism of the effect of COVID-10 on HCW's risk perception were scarce. Many factors play a role in one's perceptions and these may be varied between individuals. Related studies showed that the cognitive and decision-making abilities of health care workers are in uenced by their own emotions [15,16]. This knowledge can inform us the psychological status re ected by HCWs' emotion could associate with the perceptions of personal or organization risks.
Previous studies found that HCW's distress and fears were generated after their exposure to infectious diseases [17,18]. In 2003, HCW faced the earliest exposures of SARS and then showed some psychological problems signaling with negative emotions such as fear, anxiety, anger and depression [19].
Presently, researches during COVID − 19 epidemic using online questionnaire to assess the respondents' depression and anxiety found that symptoms of depression and anxiety were presented in a majority of HCWs, indicating the salient existence of negative emotion [20][21][22]. Therefore, the in-depth understanding of negative emotions of HCWs defending against covid-19 and their association with risk perception will help to reduce the adverse risk judgment of HCWs due to negative emotions. Importantly, it could also improve HCWs self-protection and clinical treatment level.
To explore the potential mechanism of pandemic effects on risk perception of HCW's after the outbreak of COVID-19, the current study was adapted a repeated cross-sectional research participated by Chinese HCWs who were assigned to work at the current disease resistance line, assessing risk perception and evaluating their emotional state. Two cross-sectional assessments conducted in a same group sample.
The rst cross-sectional assessment was conducted immediately after the outbreak of COVID-19, when the HCWs were dispatched to the most-hit areas in China; the other were at the period of resumption when the new con rmed cases of COVID-19 came to none and HCWs were back to their regular positions. Two main hypotheses were proposed: 1. The level of HCWs' risk perceptions would decrease given the knowledges and protections against the pandemic were gradually improved. 2. After the peak of COVID-19, HCWs' overall negative emotions would be eased, while the remarkable or major subtype of negative emotions could be different in the two period.
3. HCWs' negative emotions could be signi cantly related to risk perception in both periods and there would be differences in the pattern of the association between the subtypes of negative emotions and risk perception across two periods. Based on the literature in this eld, we would expect that negative emotions would become more meaningful in the explanation of risk perception.
We aimed at exploring in a public health emergency HCWs' negative emotional association with the level of perceived risk from emotional and cognitive behavior perspective, which emphasized emotional intervention management of HCWs impacting the risk perceptions and self-protection and would provide plausible intervention points for the psychological aid under a stressful condition.

Participants
The current study was conducted during two periods of the course of COVID-19.
"Period 1" was 3rd to 5th of February 2020, when the HCWs had been assigned to medical defense of COVID-19 for 10 to 15 days responding to the national rst level emergency response. "Period 2" was 3rd to 5th of May 2020, when all the dispatched HCWs were returned to their previous location (Shanghai) and the resumption of their work. HCWs were recruited from hospitals of Shanghai and using online questionnaires. For Period 1, there were 220 questionnaires quali ed with the criteria of the jobs of subjects (only involving HCWs) and the standard time for completing the questionnaires (set at 500-1500 ms). The average age of the subjects was 31.91 ± 7.0. For Period 2, there were 304 quali ed questionnaires collected from HCWs with a mean age of 30.75 ± 9.28. There was no age difference of the two stage samples from the same group (t = 1.56, p = 0.120) and a detailed description of the demographic differences of the two samples was presented in Table 1.

Risk perception questionnaire
The questionnaire of HCWs' risk perception for the COVID-19 was referred to the risk perception questionnaire of nursing staff [8]. The questionnaire was self-rated including 14 questions pertaining to ve dimensions: personal safety risk (questions 1-2), physical function risk (questions 3-4), occupational exposure risk (questions 5-7), psychosocial evaluation risk (questions 8-10), organizational risk (questions [11][12], and time pressure (questions [13][14]. The rating potentiality of risk was divided into ve grades from "never" to "almost always", and was assigned 1-5 points in turn.
The higher the score for each dimension, the higher HCWs' awareness of the risk. The total score represents the general level of risk perception. The self-compiled questionnaire had good internal consistency, and its Clonbach Alpha was 0.905.
The self-assessment questionnaire for negative emotions The Chinese version of Watson and Tellegen (1988) emotional self-rating scale (PANAS-Positive and negative emotions scale) was adopted in the experiment of emotional self-rating scale, which was veri ed by Chinese scholars to be of cross-cultural consistency [23][24][25].The internal consistency of this scale was with a good reliability of 0.87 and widely used. The scale contained 20 words describing emotions, including 10 positive words and 10 negative words. The participants were asked to evaluate the emotional intensity they experienced on current state on a scale of 5, among which 1 meant "very slight or no"; 2 for "a little"; 3 for "moderate"; 4 for "relatively strong", and 5 for "extremely strong". Adapting to the needs of the survey, seven negative emotions were extracted in the emotional assessment to form a reduced version of the emotional scale, which was composed of seven emotions: impatience, sadness, upset, tension, guilt, fear and worry.

Procedure
Prior to administration of the questionnaires, the research was evaluated and approved by the ethics committee of the Navy Medical University. In accordance with the requirements of the ethics committee, the written informed consent was obtained from the managers of the hospital and the HCWs from the emails. All participants were informed that the researchers were interested in their experiences during COVID-19; participation was voluntary and anonymity was emphasized. HCWs were recruited from the city of Shanghai, China through our contacting with their hospital managers who helped us to distribute the online questionnaires and emphasize the anonymity of this study. All the quali ed questionnaires were scrutinized by our researches according to selection criteria-being HCW and nishing the questionnaires within the standard time.
Statistical methods SPSS24.0 statistical software was used for data processing and analysis. The subtypes of negative emotions were represented by the mean scores of the sub-dimensions, while the risk perception was the sum of ve risk perception dimensions. The Mean and standard deviation (SD) of the statistical score data were reported. The t-test was used to analyze the differences in emotion and risk perception with the control of the demographic factors as these calculated variables were normal distributed in a large sample. In addition, a MANOVA was performed with seven subtypes of negative emotion (i.e., Impatience, Sadness, Upset, Tension, Guilt, Fear, Caution) to explore the differences in two periods. In this analysis, disparate demographic variables of the two samples was treated as covariates. Then, we computed zeroorder correlations between negative emotions and risk perception among the participants for each of the two periods (each period was evaluated separately). Fisher's test was used to examine the signi cance of the difference between each two correlation coe cients. Finally, strati ed regression analyses were employed to examine the differences in the contribution of negative emotions in predicting levels of risk perception across the two periods under the control of the demographical factors. All the test levels were at = 0.05.

Results
Demographic differences in negative emotions and risk perception across the two periods In light of the composition of the two samples, a higher percentage of females than males, a higher percentage of nurses than doctors (this issue was particularly acute for the Period-2 sample) and a small proportion of dispatched HCWs were consisted in the two samples, therefore, demographic differences in the study variables for each period were examined. The ndings of this analysis are presented in Table 2. This analysis revealed gender differences was not signi cant in the overall negative emotions, except that in the period 2 females show more worry then males with slight signi cance (t=-2.55, p = 0.017). Notably, professional difference in the two periods were not consistent. Only in period 1, doctors rated higher than nurses in risk perception (t = 2.23, p = 0.027) and negative emotions including impatience(t = 3.64, p < 0.001), sadness(t = 3.22, p = 0.002), upset(t = 2.60, p = 0.01), and tension(t = 2.55, p = 0.012). In Period 2, no signi cantly different risk perception was discovered in the location of HCWs (t = 1.26, p = 0.209), however, compared to HCWs dispatched, those not dispatched to Wuhan presented more negative emotions including impatience( t = 2.39, p = 0.018), sadness(t = 2.44, p = 0.016), tension(t = 3.08, p = 0.003), fear(t = 2.91, p = 0.038) and worry(t = 2.57, p = 0.011). Due to the demographic differences found for the study variables, in subsequent analyses, these differences were treated as a covariate.  Table 3). The negative emotions, including tension, fear, and worry, and risk perception were all signi cantly higher during Period 1 than they were during Periods 2 (mean difference tension =-0.38, p < 0.001, mean difference fear =-0.30, p < 0.001; mean difference worry =-0.16, p = 0.024; mean difference risk perception = − 2.76, p = 0.002). No signi cant differences were found between the levels of impatience, sadness, upset, and guilt during the two periods.  The association of negative emotions to the prediction of stress reactions across the two periods As showed in Table 3, the correlation between all types of negative emotions and risk perception were all signi cantly (all the P values were less than 0.01) and the strength of their associations were remained the same across two period ( z1.96> ). To closely examine the complex associations between negative emotions and risk perception among HCWs, two regression analyses were employed for the two periods. Levels of risk perception served as outcomes in these regressions. In Block 1, the predictors were dummycoded variables representing gender (0 = female), profession (0 = doctor), and location (0 = not dispatched). In Block 2, main effects of different types of negative emotions were examined (see Table 5). The results indicated that under the control of the covariates of demographic difference, in period 1, the negative emotion of worry could contribute most to risk perception(B = 2.67, p = 0.005), while in period 2, the negative emotion of tension turned to be the signi cant contributor (B = 2.95, p = 0.006).

Discussion
Our ndings revealed the risk perception and negative emotions of HCWs varied in different gender, profession and the location, as well as in different periods of COVID-19. Our ndings also revealed differences in the association between risk perception and negative emotions across the two periods. The levels of tension, fear, worry and risk perception were higher during Period 1 than they were during Periods 2 of COVID-19. Over the different periods, worry was found to be closely related to and a signi cant predictor of high-level risk perception in the period 1; whereas the predictive type of negative emotion changed to tension in the period 2.
The current study examined differences in risk perception and negative emotions and the varied associations between those variables among HCWs across two period of COVID-19. The rst hypothesis that risk perception and negative emotions decreased during two successive periods of COVID-19 were supported by our ndings, which were convergent to the development of the pandemic. In fact, our study was consisted with few longitudinal or cohort studies conducted after the outbreak of SARS which showed during the initial phases of the SARS outbreak a steady increase showed in risk perception and then leveled off in later phases [17,26]. As a infectious disease, the increase of risk perception could be neutralized by the effective vaccines or behavioral containment measures, so that it is likely the notion of risk will be less concerned. As regards the negative emotion components, signi cant decreases were found in tension, fear, and worry across the two studied periods. In period 1, which was the peak of the pandemic, an initial knowledge about the origin or the life course of a newly emerging virus was lack and visible effects are delayed, therefore, unknown risks and threats to public health were produced [9]. HCWs, as the rst-line warriors for the public health, were dispatched to the hardest-hit areas or mandated to work under the threat of infection. Hence, tension, fear and worry were dominated in the dimension of HCWs negative emotions. Whereas in period 2, when the pandemic was contained with protective measures, tension, fear and worry were gradually relieved and the perception of the pandemic was reduced, however, the risk was not nulli ed as none speci c treatment and vaccine were developed at this moment. From the practical views of the results, in the acute stage of COVID-19, tension, fear and worry are likely to be imminent and therefore hamper the rational perception about gains and losses related to protective behaviors of HCWs not only for themselves but also their patients. In such cases, the proper emotion copying ways could be avail for fending off the invisible losses and long-term consequences. In the later stage, risks were downplayed due to the emotional copying strategy, which would have double effects on the protective behaviors. To ensure the safety, it would be better to monitor the risk perception and negative emotions of HCWs for the sake of preventing the return of the pandemics.
Notably, some demographical differences were detected in the analysis. In China's COVID-19 responses, female nurses and community health workers were the rst line of defense against the outbreak, which could explain why the two samples were dominated by female HCWs, most of who were nurses.
Therefore, the study might be less convinced for the interpretation of gender difference, but we specially found doctors were more negative feelings, like impatience, sadness, upset and tension in the period 1. It may be re ected the hardship of doctors since there were the overwhelm stress from saving the lifethreaten patience and numerous patience waiting for them. Consequently, the risk perception was also higher in the doctors with more negative emotions. While in period 2, nurses showed higher risk perception than doctors, which could be explained by that in the stage nurses had to take more care supports for the patience and intense contacts. More importantly, in the late stage of COVID-19, not dispatched HCWs rated higher in the subscale of impatience, sadness, tension and worry compared to HCWs dispatched to Wuhan, while their risk perception was not signi cantly different. It could be the potential reasons that on the one hand dispatched HCWs received more psychological and substantial helps and on the other, they could have improvement in coping emotions after confronting with the worse stress. Alternatively, it could be a caveat for the sampling bias. Thereby, more studies focused on the difference between dispatched HCWs and not dispatched to clear the pandemic's impacts on HCWs mental health.
After controlled the effect of gender, profession and location, the further exploration for association between negative emotions and risk perception not only provided the supports for our third hypothesis, but also the emotion-based risk perception. Damasio [27] has thus suggested that automated incentive or alarm signals linked to pleasant or unpleasant "gut feelings" often precede cognitive reasoning. In a similar vein, processing theories have emphasized the importance of an emotion-cognition pathway [28,29]. In an acute threat situation like COVID-19, emotional aspects might gain more immediate importance, especially in the early stages of an outbreak when experts are unable to make more than tentative statements and provide partly contradictory prognoses and recommendations. Under these circumstances cognitive risk assessments might be severely hampered by lack of evidence-based information. Consequently, individuals might have little choice but to rely on experiential judgement. Our results showed the consistent positive relation between negative emotions and risk perception in the entire course of the pandemic, similar to the results of Goodwin and Sun [30] using public samples and a lot researches about the emotion and risk perception for vaccine. This study was novel not only the background and subjects but also the repeated cross-sectional method which could give light on the emotional effect on risk perception. As the analytical data showed, worry could be prominent contributor to the higher level of risk perception immediately after the outbreak of pandemic, while gradually the prominent components of negative emotions varied with the course of pandemic, and thus in the our second studied period of COVID-19, tension, re ecting the local situation and the severity of pandemic, played a key role in the risk perception. However, what role exactly an emotion-based judgement plays in interaction with risk-related cognition is still an issue in need of clari cation. Besides, only a minority of the subgroup of studies which investigated the possible predictive role of risk perception for protective behaviors could actually be considered model-based. The resulting lack of opportunity for studying the complexity of the decision-making process is bound to also affect the extent to which ndings for the relationship between risk perception and protective behaviors can be interpreted. Thus, a more systematic application of multifactor models, including the emotion-based models, would allow for far more complex insights into the workings of risk perceptions in shaping behavior.
The current study has some limitations that should be addressed. First, the data were collected in the midst of COVID-19. Therefore, some degree of potential sample bias should be taken into account. For example, the two samples each included a higher percentage of females than males, nurses than doctors (This issue was particularly acute for the Period 2 sample). Therefore, future studies should recruit larger samples of male and female HCWs. Moreover, the current study employed a cross-sectional repeated design to examine differences in the levels of negative emotions and risk perception. Further studies should adopt a longitudinal design, to follow the same sample of HCWs across different periods. That type of study would be the best way to examine changes in the levels of different variables and the associations between those variables among the same individuals over time, which could also provide more valuable ndings to the emotion-cognition-behavior model of risk perception. Moreover, the followup researches could achieve more convinced results about the effect of being dispatched for HCWs defensing against the pandemics and contribute to the targeted help services for HCWs.

Conclusion
The signi cance of this research lies in its examination of risk perception and negative emotions of HCWs combating the COVID-19 during two periods of the pandemic. The ndings showed both risk perception and negative emotions of HCWs were affected by the COVID-19. In addition, the predominated negative emotions of HCWs presented varied with time course, but the positive relations to risk perception were constant and could be a signi cant predictor of risk perception. These ndings underscore the importance of negative emotions as a signi cant factor for risk perception of HCWs enduring the challenge of pandemic within Chinese population. More importantly, our results underscore studying the change of pandemic risk perceptions and their "true" in uence over time remains an important objective, requiring not only studies with longitudinal designs, starting after potential future outbreaks, but also long-term surveillance studies which are initiated before an actual outbreak occurs.

Declarations
Ethics approval and consent to participate This study was approved by the ethics committees of the Navy Military University. A complete survey description was rst presented to the participants. Informed written consent, together with oral approvals, was obtained before the testing session according to the Declaration of Helsinki.

Consent for publication
Not appliable.
Availability of data and materials