|
Won Mo; Health Insurance Review and Assessment Service, Health Review and Assessment Committee Kim, Un-Na; Ministry of Health and Welfare, Complete List of Authors: Jang, Won Mo; Health Insurance Review and Assessment Service, Health Review and Assessment Committee Kim, Un-Na; Ministry of Health and Welfare, Complete List of Authors: Jang, Won Mo
Abstract
Objectives: This study aimed to assess the risk perception in the general population of Middle East respiratory syndrome (MERS) during the 2015 MERS coronavirus (MERS-CoV) outbreak in South Korea and the influencing factors. Design: Serial cross-sectional design with five consecutive surveys. Setting: Nationwide general population in South Korea. Participants: These included 5,015 respondents (aged 19 years) from the general population during the MERS-CoV epidemic. Primary and secondary outcome measures: The main outcome measures were 1) personallevel risk perception and 2) societal-level risk perception. Multivariate logistic regression models were used to identify the factors associated with risk perceptions.
Results: During the MERS-CoV epidemic, overall risk perception gradually decreased with the level of risk perception. Proportions of personal-level risk perception were nearly doubled and slowly declined compared to societal-level perception. Females (adjusted OR [aOR] 1.72-2.00; 95% confidence interval [CI] 1.14-2.86) and older adults (aOR 2.84-3.29; 95% CI 1.27-6.66) were more likely to perceive the personal-level risk, while younger adults were more likely to perceive the societal-level risk. The respondents who had low trust in the president or the ruling party had a higher risk perception at both levels.
Risk perceptions appear to be noticeably different between the personal-and societal-level, during outbreak. Gender, age, presidential approval rating, and party identification were significantly associated with risk perception, while the direction and intensity of involvement varied according to the level of risk perception. There is need of further efforts to understand the mechanism regarding the general public's risk perception for effective risk communication.
This is the first study to evaluate the difference in risk perception between the personal-and societal-level during MERS-CoV outbreak in South Korea.
We used five consecutive cross-sectional surveys using nationwide representative samples.
The validity of the questionnaire used in the survey was not evaluated because of the urgency of the outbreak.
This study could not confirm causal relationship between personal characteristics and risk perception due to the limitation of the cross-sectional study design.
The risk perception of disease can contribute to behaviors related to disease. 1 During contagious disease epidemics, perceived risk can have a significant impact on precautionary behaviors that might affect disease transmission. 2 3 Therefore, understanding characteristics of risk perception and factors relating to how people perceive the risk is important in terms of minimizing the impact of spread of infectious disease.
According to the Health Belief Model and Impersonal Impact Hypothesis, risk perception can be separated into personal and societal levels objectively. 2 4 5 Personal-level risk perception refers to subjective assessment of one's own vulnerability or probability of contracting a disease. Societal-level risk perception, however, is the global estimation of a serious risk to other people, i.e., probability of outbreak. Generally, precautionary behaviors are likely to depend directly upon personal-level risk perception; however that of societal-level risk perception may not affect behaviors directly. [6] [7] [8] Some researchers have suggested that individuals tend to be optimistic about personal-level risk and pessimistic about societal-level risk. 9
Newly emerging contagious diseases have created a novel chance to examine how people perceive risk during an epidemic. Since the occurrence of the index case of Middle East
Respiratory Syndrome (MERS) on 20 May 2015, a total of 186 persons were diagnosed with the disease, 38 of whom had died, and 16,693 patients were quarantined in South Korea. 10 The epidemic of MERS coronavirus (MERS-CoV) has had its largest outbreak outside of the Middle East in South Korea. 11 The occurrence of multiple transmissions after the first secondary infection and the failure of the government's on risk communication resulted in the increased concern of the general public. 12 However, many studies have not distinguished between personal and societal level of risk perception regarding contagious diseases during outbreaks. [15] [16] [17] [18] The first objective of the present study was to assess the personal-level and societal-level risk perception of MERS in the general population, during the MERS outbreak period. The second objective was to identify trends and factors associated with risk perception across the epidemic period.
Between June 3 and July 2, 2015, a total of 5,015 participants who were older than 19 years were monitored using a serial cross-sectional study design in five consecutive surveys, covering the MERS epidemic. All surveys were conducted by Gallup Korea, an affiliation of Gallup International. The first survey was conducted between June 2 and 4, 2015 after the June 1, 2015 occurrence of the first tertiary infected case. The last was conducted just two days before the last confirmed patient on July 4, 2015. Additional details (period, number of respondents successfully interviewed, and response rate) for each of the five surveys are provided in Table 1 . All surveys were conducted using mobile (85%) or landline (15%) random digit dialing numbers in eight regions which was representative of nationwide. Samples were
The interviews were conducted based on two levels of the risk perception, which are personallevel and societal-level risk perceptions. 17 Personal-level risk perception was assessed using the question "How worried are you that you may contract MERS-COV infection?" Responses were assessed on a four-point scale, with four points indicating "very worried" and one point indicating "not worried at all" (reclassified as 1-2 points = "not worried"; 3-4 points = "worried"). Personal-level risk perception proportion was defined as the number of participants who were "worried" by the number of eligible respondents. Societal-level risk perception was evaluated using the question "Do you think the MERS epidemic will subside or spread within a few days?" and required the following responses: "controlled," "uncontrolled," or "no opinion". Questions about societal-level perception were included since the survey 2. Societal- level risk perception proportion was defined as the number of participants whose response was "uncontrolled" by the number of eligible respondents.
Demographic factors evaluated as respondents' characteristics included gender, age, educational attainment, occupation, perceived household economic status, residential area, and political orientation/party (party identification). Educational attainment was classified into five levels (less than middle school, high school, university, graduate school or higher). Educational attainment was investigated in all surveys except survey 1. Occupation was classified as either unemployed, farming and fishery, self-employed, blue-collar worker, white-collar worker, fulltime housewife, or student. Perceived household economic status was classified into five levels (lower, lower middle, middle, upper middle, upper). Respondents were classified as either metropolitan or non-metropolitan residents; and distinguished by whether they resided in an area where MERS had occurred or not. Party identification was classified based on the support for the president or the political parties. Support or lack of support for the president was assessed using the options of "approval", "disapproval", or "no opinion", while support for the party identification was assessed based on alignment either with the ruling party, with the opposition party, or no opinion.
Response rates according to personal-level or societal-level risk perception were calculated over time. Univariate analyses using chi-square test were performed in the five consecutive surveys, entirely and respectively, to identify the relationships between risk perception and each demographic variable. We used multivariable logistic regression analyses to explore ; hence, these were dropped from both the descriptive and logistic regression analyses. Using logistic regression analysis for each personal-level and societal level risk perception, "y = 1" was used respectively when "worried" in personal-level and when "uncontrolled" in societal-level, otherwise "y = 0" was used.
The general characteristics of the participants are shown in Table 2 . There were no statistically significant differences between surveys. Nearly half of the participants were female, aged <50 years, were educated up to high school or below, were from the affected area, showed disapproval of the president or the ruling party, and had personal-level risk perception. Majority of the participants were employed, were in the middle economic status, metropolitan, without societal-level risk perception (controlled). Figure 1 reports how the outbreak proceeded, with three overlapping transmission periods, the timing of the five independent surveys, and the risk perception rates. Noticeable differences were investigated between personal-level and societal-level risk proportions throughout the epidemic periods. Overall risk perception at personal-level proportion (56.5%) was nearly two times higher than at societal-level (30.3%). Personal-level risk perception proportions were always higher than societal-level during the present study periods. Of the personal-level risk perception, proportion was initially high during survey 1 (67.3%), declined during survey 2 (55.1%), temporally rose during survey 3 (62.8%), and declined again during surveys 4 and 5 (52.2% and 44.9%, respectively). A similar trend was observed in the societal-level risk perception proportions. The percentages of respondents who reported as being "worried" or "uncontrolled" decreased gradually after survey 3. Societal-level risk perception proportions decreased more rapidly than personal-level, over time, from 52.6% and 62.8% in survey 3 to 9.0% and 44.9% in survey 5, respectively. At the beginning of the occurrences of tertiary and quaternary cases, we identified high perceived risk in both the personal-level and societal-level proportions. 29 ; 95% CI 1.27-6.66). Higher level of education was also associated with lower level of risk perception at the personal-level, but was not statistically significant except university degree in the overall survey (aOR 0.73; 95% CI 0.55-0.96). Lower economic status and those living in metropolitan cities paid more attention to the personal-level risk of MERS-CoV in the overall model. Those who disapproved of the president and the ruling party had higher risk perception at the personal-level; the peak of disapproval was found in survey 2. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 u.a.: unavailable data; * P < 0.05. a There was small sample size of those who perceived societal-level risk those in the upper economic level in survey 5, the perceived household economic status was excluded from the survey 5 model.
The present study found that personal-level risk perception was more than twice the level at the societal-level. Risk perception increased with new generations of transmission, such as with the tertiary and quaternary infection. Both risk perceptions tended to decrease over time and the societal-level risk perception declined more rapidly. Given that external stimuli are extreme events, two different reactions can occur: the affective reaction (risk-as-feelings) and cognitive reaction (risk-as-analysis). 19 20 Previous studies suggest that affective reaction is quick, intuitive, automatic, while cognitive reaction is slow, deliberate, and probability calculative. In the early phase of the outbreak, people may be experiencing challenges when attempting to 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 F o r p e e r r e v i e w o n l y 15 quantify the risk, which might lead to an affective reaction. 3 21 In contrast, cognitive reaction may occur during the late stage of the epidemic. The traditional impersonal-impact hypothesis proposes that cognitive reaction is more likely to correlate with societal-level risk perception than that of the personal-level. 5 8 Because the societal-level risk perception decreased rapidly, this study does not seem to support that cognitive reaction is more closely related to societallevel risk perception. While the affective or cognitive reaction do not individually make an impact on the different stages of epidemic; however, they can affect it together, simultaneously, indicating both personal and societal level risk perceptions. 17 Additional research is needed to understand why the personal-level risk perception was doubled and lasted longer than that of the societal-level.
According to multiple logistic regression analyses, being female was predisposed to greater risk perception at the personal-level, but not at the societal level. Previous studies that investigated risk perception by gender also showed that a higher risk perception was associated with the female gender. 15 16 22-24 However, previous studies did not distinguish between the level of risk perception. Possible explanation for the higher perception of risk by the female may be explained by worldviews and culture-identity protection. 25 Further research is needed to determine why the same female group showed differences in perceived risk for personal and societal levels.
The older the respondents, the higher the perceived personal-level risk, but the opposite occurred at the societal-level. In the early stage of the epidemic, the older the respondent, the lower the risk perception, but this increased with time at the personal-level. Some researchers suggested that numeracy skills may be correlated with risk perception. 26 27 It is necessary to further investigate the impact of age on risk perception. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 F o r p e e r r e v i e w o n l y 16 There were no significant differences in the proportions of those with risk perception according to the major socioeconomic characteristics (education, income level, occupation, or area).
However, given that some hierarchy-specific trends in income level were observed in the overall model of personal-level risk perception, these results might be due to the limited number of study participants Those who did not support the president, or the ruling party were reported to have had higher risk perception in both the personal and societal levels. In the group that did not approve of the president, the probabilities of risk perception were higher at the individual-level than at the societal-level, but not for the party identification. In the early days of the MERS-COV outbreak, the government did not specify details regarding scientifically uncertain information in order to reduce public anxiety over the crisis, nor did the government disclose which hospitals the confirmed patients had visited. This resulted in increased public distrust in the government. 12 13 Similar pattern of distrust in the government was associated with the spread of infection, during the outbreak of Ebola. 28 29 This study, which used a serial cross-sectional study design had some limitations. First, the study used a cross-sectional study design. Thus, causal relations between personal characteristics and risk perceptions could not be determined-rather, it could only suggest their relevance. Particularly, it was difficult to consider that presidential approval rating and party identification would actually lead to greater risk perception. Nevertheless, consecutive crosssectional surveys may be a better option than a single cross-sectional survey. Second, this study could not evaluate the intensity of risk perception, because it only included questions focusing on whether or not participants recognized the risk at the different levels. It would be useful to
This study is the first to evaluate the differences in risk perception between the personal-and societal-level during the MERS-CoV outbreak in South Korea; and also reported various factors influencing risk perception. Quality of risk communication can create conditions for modulating the easy spread of emerging contagious diseases. To prevent the failure of epidemic management, further efforts are needed to understand the mechanism behind the general public's risk perception by the governmental public health sector as well as by the society of academy. Planning and implementation of strategies that consider the risk awareness mechanism will be a significant step in the right direction during national infectious disease crises.
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
None declared.
WMJ participated and analyzed the data, interpreted the data, drafted and amended the manuscript. UNK contributed to the analysis of the data, interpreted the data, drafted and amended the manuscript. DHJ contributed to questionnaire design, coordinated data collection, and data interpretation. SJE contributed to study design, supervised the research, data interpretation, and amended the manuscript. JYL contributed to study design, supervised the research, data interpretation, and amended the manuscript.
We would like to thank Gallup Korea, an affiliation of Gallup International, for supporting surveys and data collection for this manuscript.
No additional data are available. Discuss the generalisability (external validity) of the study results 17
Give the source of funding and the role of the funders for the present study and, if applicable, for the original study on which the present article is based *Give information separately for cases and controls in case-control studies and, if applicable, for exposed and unexposed groups in cohort and cross-sectional studies. 2.00; 95% confidence interval [CI] 1.14-2.86) and lower perceived economic status respondents (aOR 1.84-2.28; 95% CI 1.03-4.14) were more likely to perceive the affective risk.
The older adults, the higher the affective risk perception, but the lower the cognitive risk perception compared to the younger adults. The respondents who had low trust in the president or the ruling party had a higher risk perception in both affective and cognitive..
This study suggests that even if cognitive risk perception is dissolved, affective risk perception can continue during MERS-CoV epidemic. Risk perceptions associating factors (ie gender, age, perceived economic status) appear to be noticeably different between the affective and cognitive dimensions. It also adds findings that trust in the president not only affective risk perception but also cognitive risk perception. There is need of further efforts to understand the mechanism regarding the general public's risk perception for effective risk communication.
Keywords: Middle East respiratory syndrome coronavirus, risk and perception, epidemics, surveys
This is the first study to evaluate the difference in risk perception between the affective and cognitive dimensions during MERS-CoV outbreak in South Korea.
We used five consecutive cross-sectional surveys using nationwide representative samples.
The validity of the questionnaire used in the survey was not evaluated because of the urgency of the outbreak.
This study could not confirm causal relationship between personal characteristics and risk perception due to the limitation of the cross-sectional study design. During contagious disease epidemics, perceived risk can have a significant impact on precautionary behaviors that might affect disease transmission. [10] [11] [12] Relevant empirical study was emphasized that informing public about the disease outbreak, such as the Ebola virus, could reduce worry about contracting the virus and take more preventive measures. 13 Evaluating the public risk perception of disease helps us know what knowledge the public needs. Therefore, understanding characteristics of risk perception and factors relating to how people perceive the risk is important in terms of minimizing the impact of spread of infectious disease.
Given that external stimuli are extreme events, two different reactions can occur: the affective reaction (risk-as-feelings) and cognitive reaction (risk-as-analysis). [14] [15] [16] Previous studies suggest that affective reaction is quick, intuitive, automatic, while cognitive reaction is slow, deliberate, and probability calculative. In the early phase of the outbreak, people may be experiencing challenges when attempting to quantify the risk, which might lead to an affective reaction. 12 17 In contrast, cognitive reaction may occur during the late stage of the epidemic.
Most people may not conduct deliberate risk analysis when they cope with lack of knowledge about risk, such as new disease outbreak, but rely on simple heuristics. 18 19 Heuristic processing can be understood as simple decision rule of thumb or mental shortcut that can reduce the complexity of decision making. When risk management decisions are needed, trust in the institutions can be used as one of the heuristics. 20 People having trust in the responsible risk manager, such as the government, may perceive less risk in a particular issue than people not having trust. 21 22 Regarding the MERS epidemic in South Korea, less trust in the government affected increasing of individuals` risk perception. [23] [24] [25] Trust is known to be related not only to cognitive risk perception but also to affective risk perception. 26 27 However, when assessing the influence of trust in risk perception, many studies have not distinguished between affective and cognitive reaction regarding contagious diseases during outbreaks. 3 12 23 24 28-30 We hypothesized that (1) affective risk perception would increase and decrease faster than cognitive risk perception over time and that (2) low trust in government would be related with high risk perception (both affective and cognitive).
Between June 3 and July 2, 2015, a total of 5,015 participants who were older than 19 years were monitored using a serial cross-sectional study design in five consecutive surveys, covering the MERS epidemic. All surveys were conducted using mobile (85%) or landline (15%) random digit dialing numbers in eight regions which was representative of nationwide.
Samples were selected post-stratification by gender, age, and province. The total number of weighted cases in this survey equals the total number of unweighted cases at the national level.
The weights were normalized in order to calculate proportions and ratios; however; not for estimating the number of the subtotal populations. Trained interviewers conducted all interviews using computer assisted telephone interviewing (CATI). The first survey was Details including period, number of respondents successfully interviewed, and response rate for each of the five surveys are provided in Table 1 .
The interviews were conducted based on two aspects of the risk perception, which are affective and cognitive risk perceptions (Supplementary file). Affective risk perception was assessed using the question "How much worried are you that you could get MERS?" Responses were assessed on a four-point scale, with four points indicating "very worried" and one point indicating "not worried at all" (reclassified as 1-2 points = "not worried"; 3-4 points = "worried"). Affective risk perception proportion was defined as the number of participants who were "worried" by the number of eligible respondents. Cognitive risk perception was evaluated using the question "Do you think MERS epidemic will settled down in the next few days or spread further?" and required the following responses: "will settle down," "will spread further".
Questions about cognitive perception were included since the survey 2. Cognitive risk perception proportion was defined as the number of participants whose response was "will spread further" by the number of eligible respondents. Trust in government was assessed using presidential job approval rating. Trust in government includes expectations of government`s competence to prevent people from risk and develop and implement follow-up measures. 31 It can be termed this trust concept as competence-based trust. 32 33 We tried to assess the competence-based trust using presidential job approval rating. Presidential job approval was evaluated using the question "Do you approve or disapprove of the way President Park Geunhye is handling her job as president?" and required the following responses: "approval", "disapproval". The development of questionnaires on risk perception and trust in the government had not gone through a valid procedure due to the urgency of the outbreak. We also imposed survey items on existing questionnaire developed by Gallup Korea, an affiliation of Gallup International.
Response rates according to affective or cognitive risk perceptions were calculated over time.
Univariate analyses using chi-square test were performed in the five consecutive surveys, entirely and respectively, to identify the relationships between risk perception and each demographic variable. We used multivariable logistic regression analyses to explore factors influencing risk perceptions (affective and cognitive) in the five surveys, entirely and respectively. Basic multivariable logistic regression model was adjusted for gender, age, educational attainment, occupation, perceived household economic status, affected area, residential area, presidential job approval, and party identification. Basic model was used in analysis for each affective and cognitive risk perception, "y = 1" was used respectively when "worried" in affective and when "spread" in cognitive, otherwise "y = 0" was used.
This study was reviewed and approved by the Institutional Review Board (IRB) of Seoul
Metropolitan Government-Seoul National University Boramae Medical Center (IRB No.
20190515/07 -2019 -11/062). The need for informed consent was waived by the board.
No patient or public was involved in the design or planning of this study.
The general characteristics of the participants are shown in Table 2 . There were no statistically significant differences between surveys except perceived economic status, affective risk perception, cognitive risk perception. Nearly half of the participants were female, aged <50 years, were educated up to high school or below, were from the affected area, showed disapproval of the president or the ruling party. Majority of the participants were employed, were in the middle economic status, metropolitan. More than half of participants were worried, but had views that epidemic will subside. Figure 1 reports how the outbreak proceeded, with three overlapping transmission periods, the timing of the five independent surveys, and the risk perception rates. Differences were investigated between affective and cognitive risk proportions throughout the epidemic periods.
Overall risk perception of the five surveys at affective proportion (56.5%) was nearly two times higher than at cognitive dimension (30.3%). Affective risk perception proportions were always higher than cognitive dimension during the present study periods. Of the affective risk perception, proportion was initially high during survey 1 (67.3%), declined during survey 2 (55.1%), temporally rose during survey 3 (62.8%), and declined again during surveys 4 and 5 (52.2% and 44.9%, respectively). A similar trend was observed in the cognitive risk perception proportions. The percentages of respondents who reported as being "worried" or "spread further" decreased gradually after survey 3. Cognitive risk perception proportions decreased more rapidly than affective aspect, over time, from 52.6% and 62.8% in survey 3 to 9.0% and 44.9% in survey 5, respectively. At the beginning of the occurrences of tertiary and quaternary cases, we identified high perceived risk in both the affective and cognitive aspects proportions. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59
Unlike the cognitive risk perception, no difference was found by gender in the cognitive risk perception (Table 4) . Furthermore, respondents aged >30 years were consistently less aware of the cognitive risk during MERS-CoV epidemic. Generally, no not statistically significant association was found with educational attainment, occupation, perceived economic status, MERS-CoV affected area, and metropolitan area. Similar to the affective dimension, those who 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 u.a.: unavailable data; * P < 0.05. a There was small sample size of those who perceived societal-level risk those in the upper economic level in survey 5, the perceived household economic status was excluded from the survey 5 model.
The aims of the present study were to explore the differences in risk perception at affective and cognitive dimension and examine the relationship between trust in government and the both risk perceptions. To do this end, we investigated the pattern of affective and cognitive risk perception proportions during MERS-CoV epidemic, respectively; analyzed the correlations of presidential job approval rating, party identification and risk perceptions (affective and cognitive).
First, we found that affective risk perception responded faster and lasts longer. The affective risk perception proportions were always higher than at the cognitive dimension. Risk However, our results are inconsistent with previous studies that affective reaction tends to appear in early epidemic periods. 12 17 Relevant research in risk perception have proposed that affective reaction is fast, efficient, automatic, experiential compared to cognitive reaction. [14] [15] [16] We can consider the possibility that damaged trust in government as a responsible risk manager may had further evoked the emotional risk perception. 8 9 23-25 While the affective or cognitive reaction do not individually make an impact on the different stages of epidemic; however, they can affect it together, simultaneously, indicating both affective and cognitive risk perceptions. 12 17 26 29 Additional research is needed to understand why the affective risk perception was higher and lasted longer than that of the cognitive risk perception during MERS-CoV epidemic in South Korea.
Second, our study shows that low trust in government had influenced both affective and cognitive risk perceptions. We tried to assess the competence-based trust using presidential job approval rating. After party identification was adjusted, we examine correlation with trust and risk perception. It is consistent with previous studies that trust in government could shape the public`s risk perception (both affective and cognitive). 21 22 26 27 However, the previous studies have not distinguished between affective and cognitive reaction when evaluating the impact of trust regarding contagious diseases during outbreaks. 3 12 23 24 28-30 Our findings suggest that trust in government is correlated with both affective and cognitive risk perception and it is important to understand the relationship between trust in government and two different aspects of risk perceptions. Those who did not support the president were reported to have had higher risk perception in both the affective and cognitive levels. In the group that did not approve of the president, the probabilities of risk perception were higher at the cognitive dimension than at affective dimension. In the early days of the MERS-COV outbreak, the government did not specify details regarding scientifically uncertain information in order to reduce public anxiety over the crisis, nor did the government disclose which hospitals the confirmed patients had visited. This resulted in increased public distrust in the government. 4 5 8 9 Similar pattern of distrust in the government was associated with the spread of infection, during the outbreak of Ebola. 34 35 Those who disapproved of the ruling party had also higher risk perceptions.
Identification of party is can be classified in the political aspect of trust. 36 It need to investigate further comprehensive understanding of trust`s effect on risk perception.
Third, we found that gender, age, perceived economic status, residential area, party identification correlated significantly with risk perception. According to multiple logistic regression analyses, being female was predisposed to greater risk perception at the affective risk perception, but not at the cognitive dimension. Previous studies that investigated risk perception by gender also showed that a lower risk perception was associated with the male gender. 3 28 37-39 Possible explanation for lower perception of risk by male are that male have more to gain from risky behaviors. 40 However, previous studies did not distinguish between the level of risk perception. Further research is needed to determine why the same female group showed differences in perceived risk for affective and cognitive levels. The older the respondents, the lower the perceived cognitive dimension, but the opposite occurred weakly at the affective risk perception. The correlation with age and affective risk perception was not significant in the most model (survey 1, survey 2, survey 3, survey 4 models). We found that the higher the age, the higher president's job approval rating. The effect of trust may lead to a reduction in the cognitive risk perception among older respondents. Further research is needed as to why the effect of trust in government had not been shown in the affective risk perception.
Given that some hierarchy-specific trends in income level were observed only in the overall model of personal-level risk perception, these results consistent with previous studies. [41] [42] [43] The location effect on risk perception also was evaluated in this study, however it was not clear the correlate with risk proximity and risk perception. 44 There were no significant differences in the proportions of those with risk perception according to the major socioeconomic characteristics (education, income level, occupation). It is necessary to further investigate the correlation with demographic factors and risk perception.
This study, which used a serial cross-sectional study design had some limitations. First, the study used a cross-sectional study design. Thus, causal relations between personal characteristics and risk perceptions could not be determined-rather, it could only suggest their relevance. Second, this study could not evaluate the intensity of risk perception, because it only included questions focusing on whether or not participants recognized the risk at the different levels. It would be useful to evaluate risk perceptions of respondents qualitatively if questions about the circumstances and characteristics of risk perception were surveyed in future studies.
Third, because of the rapidly evolving epidemic, this study could not evaluate the validity of the questionnaire using a test-retest design. Fourth, small sample size of some variables once stratified, i.e. perceived household economic status, led to the exclusion of major socioeconomic characteristics from further analyses. This study is the first to evaluate the differences in risk perception at affective and cognitive dimension and the relationship between trust in government and the both risk perceptions.
during the MERS-CoV outbreak in South Korea; and also reported various factors influencing risk perception. We found that affective risk perception responded faster and lasts longer; and low trust in government had influenced both affective and cognitive risk perceptions. Quality of risk communication can create conditions for modulating the easy spread of emerging contagious diseases. To prevent the failure of epidemic management, further efforts are needed to understand the mechanism behind the general public's risk perception by the governmental public health sector as well as by the society of academy. Planning and implementation of strategies that consider the risk awareness mechanism will be a significant step in the right direction during national infectious disease crises.
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
None declared.
WMJ participated and analyzed the data, interpreted the data, drafted and amended the manuscript. UNK, SC and HJ contributed to the analysis of the data, interpreted the data, drafted and amended the manuscript. DHJ contributed to questionnaire design, coordinated data collection, and data interpretation. SJE and JYL contributed to study design, supervised the research, data interpretation, and amended the manuscript.
We would like to thank Gallup Korea, an affiliation of Gallup International, for supporting surveys and data collection for this manuscript.
No additional data are available. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 Discuss the generalisability (external validity) of the study results 17
Give the source of funding and the role of the funders for the present study and, if applicable, for the original study on which the present article is based *Give information separately for cases and controls in case-control studies and, if applicable, for exposed and unexposed groups in cohort and cross-sectional studies.
This study suggests that even if cognitive risk perception is dissolved, affective risk perception can continue during MERS-CoV epidemic. Risk perception associating factors (i.e. gender, age, self-reported economic status) appear to be noticeably different between affective and cognitive dimensions. It also indicates that trust in the government influence not only affective risk perception but also cognitive risk perception. There is a need for further 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 F o r p e e r r e v i e w o n l y 3 efforts to understand the mechanism regarding the general public's risk perception for effective risk communication.
Keywords: Middle East respiratory syndrome coronavirus, risk and perception, epidemics, surveys
This is the first study to evaluate the difference in risk perception between the affective and cognitive dimensions during MERS-CoV outbreak in South Korea.
We used four consecutive cross-sectional surveys using nationwide representative samples.
The validity of the questionnaire used in the survey was not evaluated because of the urgency of the outbreak.
This study could not confirm causal relationship between personal characteristics and risk perception due to the limitation of the cross-sectional study design. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 During contagious disease epidemics, perceived risk can have a significant impact on precautionary behaviors that might affect disease transmission. [10] [11] [12] A relevant empirical study emphasized that informing public about the disease outbreak, such as the Ebola virus, could reduce worry about contracting the virus and take more preventive measures. 13 The evaluation of public risk perception of disease helps us to know what knowledge the public needs.
Therefore, understanding characteristics of risk perception and factors relating to how people perceive the risk is important in terms of minimizing the impact of spread of infectious disease.
Given that external stimuli are extreme events, two different reactions can occur: the affective reaction (risk-as-feelings) and cognitive reaction (risk-as-analysis). [14] [15] [16] Previous studies suggest that affective reaction is quick, intuitive, automatic, while cognitive reaction is slow, deliberate, and probably calculative. In the early phase of the outbreak, people may experience 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 F o r p e e r r e v i e w o n l y 5 challenges when attempting to quantify the risk, which may lead to an affective reaction. 12 17 In contrast, cognitive reaction may occur during the late stage of the epidemic.
Most people may not conduct deliberate risk analysis when they cope with lack of knowledge about risk, such as new disease outbreak, but rely on simple heuristics. 18 19 Heuristic processing can be understood as simple decision rule of thumb or mental shortcut that can reduce the complexity of decision making. When risk management decisions are needed, trust in the institutions can be used as one of the heuristics. 20 People having trust in the responsible risk manager, such as the government, may perceive less risk in a particular situation than people not having trust. 21 22 Regarding the MERS epidemic in South Korea, less trust in the government affected increasing number of individuals` risk perception. [23] [24] [25] Trust is known to be related not only to cognitive risk perception but also to affective risk perception. 26 27 However, when assessing the influence of trust in risk perception, many studies have not distinguished between affective and cognitive reaction regarding contagious diseases during outbreaks. 3 12 23 24 28-30 We hypothesized that (1) affective risk perception would increase and decrease faster than cognitive risk perception over time and that (2) low trust in government would be related with high risk perception (both affective and cognitive).
Between June 9 and July 2, 2015, a total of 4,010 participants who were older than 19 years were monitored using a serial cross-sectional study design in four consecutive surveys, covering the MERS epidemic. All surveys were conducted using mobile (85%) or landline (15%) random digit dialing numbers in eight regions which was representative of nationwide. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 Details including period, number of respondents successfully interviewed, and response rate for each of the four surveys are provided in Table 1 . classified based on the support for the political parties. Support for the party identification was assessed based on alignment either with the ruling party, with the opposition party, or no opinion.
The interviews were conducted based on two aspects of the risk perception, which are affective and cognitive risk perceptions (Supplementary file). Affective risk perception was assessed using the question "How much worried are you that you could get MERS?" Responses were assessed on a four-point scale, with four points indicating "very worried" and one point indicating "not worried at all" (reclassified as 1-2 points = "not worried"; 3-4 points = "worried"). Affective risk perception proportion was defined as the number of participants who were "worried" by the number of eligible respondents. Cognitive risk perception was evaluated using the question "Do you think MERS epidemic will settled down in the next few days or spread further?" and required the following responses: "will settle down," "will spread further".
Cognitive risk perception proportion was defined as the number of participants whose response was "will spread further" by the number of eligible respondents. Trust in government was assessed using presidential job approval rating. Trust in government includes expectations of government`s competence to prevent people from risk and develop and implement follow-up measures. 31 This trust concept can be termed competence-based trust. 32 33 We tried to assess the competence-based trust in the government using presidential job approval rating.
Presidential job approval was evaluated using the question "Do you approve or disapprove of the way President Park Geun-hye is handling her job as president?" and required the following responses: "approval", "disapproval". The development of questionnaires on risk perception and trust in the government had not gone through a validity procedure due to the urgency of the outbreak. We also imposed survey items on existing questionnaire developed by Gallup
Korea, an affiliation of Gallup International.
Response rates according to affective or cognitive risk perceptions were calculated over time.
Univariate analyses using chi-square test were performed in the four consecutive surveys, entirely and respectively, to identify the relationships between risk perception and each demographic variable. We used multivariable logistic regression analyses to explore factors influencing risk perceptions (affective and cognitive) in the four surveys, entirely and respectively. Multivariable logistic regression model was adjusted for gender, age, educational attainment, occupation, self-reported household economic status, affected area, residential area, presidential job approval, and party identification. The self-reported household economic status was excluded in survey 4 model with cognitive risk perception. These exclusions were because there was small sample size of those who perceived cognitive risk in the upper economic level in survey 4. Missing values of any variable were ≤2.7%. Using logistic regression analysis for each affective and cognitive risk perception, "y = 1" was used respectively when "worried" in affective and when "spread" in cognitive, otherwise "y = 0" was used.
This study was reviewed and approved by the Institutional Review Board (IRB) of Seoul
Metropolitan Government-Seoul National University Boramae Medical Center (IRB No.
20190515/07 -2019 -11/062). The need for informed consent was waived by the board.
No patient or public was involved in the design or planning of this study.
The general characteristics of the participants are shown in Table 2 . There were no statistically significant differences between surveys except self-reported economic status, affective risk perception, cognitive risk perception. Nearly half of the participants were female, aged <50 years, educated up to high school or below, from the affected area, and showed disapproval of the president or the ruling party. Majority of the participants were employed, of middle economic status, and metropolitan. More than half of participants were worried but had views that the epidemic would subside. Figure 1 reports how the outbreak proceeded, with three overlapping transmission periods, the timing of the four independent surveys, and the risk perception rates. Differences were investigated between affective and cognitive risk proportions throughout the epidemic periods.
Overall risk perception of the four surveys at affective proportion (53.8%) was nearly two times higher than at cognitive dimension (30.3%). Affective risk perception proportions were always higher than cognitive dimension during the present study periods. Of the affective risk perception, proportion was initially high during survey 1 (55.0%), rose during survey 2 (62.8%), and declined again during surveys 3 and 4 (52.2% and 44.9%, respectively). A similar trend was observed in the cognitive risk perception proportions. The percentages of respondents who reported as being "worried" or "spread further" decreased gradually after survey 2. Cognitive risk perception proportions decreased more rapidly than affective aspect, over time, from 52.6% and 62.8% in survey 2 to 9.0% and 44.9% in survey 4, respectively. At the beginning of the occurrences of tertiary and quaternary cases, we identified high perceived risk in both the affective and cognitive aspects proportions. Table 3 shows the association between variables and risk perception of MERS-CoV at the affective dimension. The result showed that gender, age, educational attainment, self-reported economic status, area, presidential job approval rating, and party identification were significantly associated with affective risk perception. Women (adjusted OR [aOR] 1.72-2.00; 95% confidence interval [CI] 1.14-2.86) were more likely to perceive MERS-CoV risk at affective dimension, which decreased with time, and subsequently increased again. Groups of older than 40 years were less aware of the risk (aOR 0.58-0.76; 95% CI 0.28-1.56) in survey 1; however, they perceived the risk more over time (aOR 2.84-3.29; 95% CI 1.27-6.66) in survey 4. The association of education with affective risk perception was non-significant except university degree in the overall survey (aOR 0.73; 95% CI 0.55-0.96). Lower economic status and those living in metropolitan cities paid more attention to the affective risk of MERS-CoV in the overall model. Those who disapproved of the president and the ruling party had higher risk perception at the affective dimension; the peak of disapproval was found in survey 1. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59
Unlike the cognitive risk perception, no difference was found by gender in the cognitive risk perception (Table 4) . Furthermore, respondents aged >30 years were consistently less aware of the cognitive risk during MERS-CoV epidemic. Generally, no not statistically significant association was found with educational attainment, occupation, self-reported economic status, MERS-CoV affected area, and metropolitan area. Similar to the affective dimension, those who disapproved of the president and the ruling party had higher risk perceptions at the cognitive dimension. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 First, we found that affective risk perception responded faster and lasts longer. The affective risk perception proportions were always higher than at the cognitive dimension. Risk perception increased with new generations of transmission, such as with the tertiary and quaternary infection. Both risk perceptions tended to decrease over time and the cognitive risk perception declined more rapidly.
However, our results that affective reaction tends to decrease before cognitive reaction are inconsistent with those of previous studies. 12 17 Relevant research in risk perception have proposed that affective reaction is fast, efficient, automatic, experiential compared to cognitive reaction. [14] [15] [16] We can consider the possibility that damaged trust in government as a responsible risk manager might have further evoked the emotional risk perception. 8 9 23-25 While the affective or cognitive reaction do not individually have an impact on the different stages of the epidemic; they can, however, affect it together, simultaneously, indicating both affective and 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 F o r p e e r r e v i e w o n l y 16 cognitive risk perceptions. 12 17 26 29 Additional research is needed to understand why the affective risk perception was higher and lasted longer than that of the cognitive risk perception during MERS-CoV epidemic in South Korea.
Second, our study shows that low trust in government had influenced both affective and cognitive risk perceptions. After party identification was adjusted for, we examined correlation with trust and risk perception. It is consistent with previous studies that trust in government could shape the public`s risk perception (both affective and cognitive). 21 22 26 27 However, the previous studies have not distinguished between affective and cognitive reaction when evaluating the impact of trust regarding contagious diseases during outbreaks. 3 12 23 24 28-30 Our findings suggest that trust in government is correlated with both affective and cognitive risk perception and it is important to understand the relationship between trust in government and two different aspects of risk perceptions. Those who did not support the president were reported to have had higher risk perception in both the affective and cognitive levels. In the group that did not approve of the president, the probabilities of risk perception were higher at the cognitive dimension than at affective dimension. In the early days of the MERS-COV outbreak, the government did not specify details regarding scientifically uncertain information in order to reduce public anxiety over the crisis, nor did the government disclose which hospitals the confirmed patients had visited. This resulted in increased public distrust in the government. 4 5 8 9 Similar patterns of distrust in the government was associated with the spread of infection, during the outbreak of Ebola. 34 35 Those who disapproved of the ruling party had also higher risk perceptions. Identification of party can be classified in the political aspect of trust. 36 There is need to investigate further comprehensive understanding of trust`s effect on risk perception. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 F o r p e e r r e v i e w o n l y 17 Third, we found that gender, age, self-reported economic status, residential area, party identification correlated significantly with risk perception. According to multiple logistic regression analyses, being female predisposed to greater risk perception at the affective risk perception, but not at the cognitive dimension. Previous studies that investigated risk perception by gender also showed that a lower risk perception was associated with the male gender. 3 28 37-39 Possible explanation for lower perception of risk by male are that male have more to gain from risky behaviors. 40 However, previous studies did not distinguish between the level of risk perception. Further research is needed to determine why the same female group showed differences in perceived risk for affective and cognitive levels. The older the respondents, the lower the perceived cognitive dimension, but the opposite occurred weakly in the affective risk perception. The correlation with age and affective risk perception was not significant in most model (survey 1, survey 2, survey 3 models). After trust in government was adjusted for, we found correlation between older age and lower cognitive risk perception.
Further research is needed as to why the effect of trust in the government had not been shown in the affective risk perception.
Given that some hierarchy-specific trends in income level were observed only in the overall model of affective risk perception, these results were consistent with previous studies. [41] [42] [43] The location effect on risk perception also was evaluated in this study, but it was not clear the correlation with risk proximity and risk perception. 44 There were no significant differences in the proportions of those with risk perception according to the major socioeconomic characteristics (education, income level, occupation). It is necessary to further investigate the correlation with demographic factors and risk perception. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 F o r p e e r r e v i e w o n l y 18 This study, which used a serial cross-sectional study design had some limitations. First, the study used a cross-sectional study design. Thus, causal relations between personal characteristics and risk perceptions could not be determined-rather, it could only suggest their relevance. Second, this study could not evaluate the intensity of risk perception, because it only included questions focusing on whether or not participants recognized the risk at the different levels. It would be useful to evaluate risk perceptions of respondents qualitatively if questions about the circumstances and characteristics of risk perception were surveyed in future studies.
Third, because of the rapidly evolving epidemic, this study could not evaluate the validity of the questionnaire using a test-retest design. Fourth, small sample size of some variables once stratified (e.g., self-reported household economic status)_led to the exclusion of major socioeconomic characteristics from further analyses.
This study is the first to evaluate the differences in risk perception at affective and cognitive dimension and the relationship between trust in the government and both risk perceptions during the MERS-CoV outbreak in South Korea. The study also reported various factors influencing risk perception. We found that affective risk perception responded faster and lasts longer; and low trust in the government influenced both affective and cognitive risk perceptions.
Quality of risk communication can create conditions for modulating the easy spread of emerging contagious diseases. To prevent the failure of epidemic management, further efforts are needed to understand the mechanism behind the general public's risk perception, the governmental public health sector, as well as the society of academy. Planning and implementation of strategies that consider the risk awareness mechanism will be a significant step in the right direction during national infectious disease crises. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
None declared.
WMJ participated and analyzed the data, interpreted the data, drafted and amended the manuscript. UNK, SC and HJ contributed to the analysis of the data, interpreted the data, drafted and amended the manuscript. DHJ contributed to questionnaire design, coordinated data collection, and data interpretation. SJE and JYL contributed to study design, supervised the research, data interpretation, and amended the manuscript.
We would like to thank Gallup Korea, an affiliation of Gallup International, for supporting surveys and data collection for this manuscript.
No additional data are available. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 F o r p e e r r e v i e w o n l y 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 F o r p e e r r e v i e w o n l y 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Introduction Background/rationale 2 Explain the scientific background and rationale for the investigation being reported 4
Objectives 3 State specific objectives, including any prespecified hypotheses 5
Present key elements of study design early in the paper 5 Discuss the generalisability (external validity) of the study results 17
Give the source of funding and the role of the funders for the present study and, if applicable, for the original study on which the present article is based *Give information separately for cases and controls in case-control studies and, if applicable, for exposed and unexposed groups in cohort and cross-sectional studies. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
|
