Association between preventive behaviour and anxiety at the start of the COVID-19 pandemic in Slovenia
Article Category: Research Article
Published Online: Dec 31, 2020
Page range: 17 - 24
Received: Jun 30, 2020
Accepted: Nov 04, 2020
DOI: https://doi.org/10.2478/sjph-2021-0004
Keywords
© 2021 National Institute of Public Health, Slovenia
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.
After the coronavirus outbreak at the end of December 2019 in China, the first cases of infection in Europe appeared after 24 January 2020, with the first large outbreak occurring in Northern Italy (1). As a neighbouring country of Italy with a high rate of population mobility, the relatively fast spread of the infection to Slovenia was expected. The Slovenian National Institute of Public Health published recommendations to prevent the spread of infection, including hand disinfection (2) and social distancing. However, the isolation of people who had been to Italy, but who had not been in direct contact with infected persons and were asymptomatic, was not recommended. As awareness of the infectiousness of the virus grew, citizens and institutions took a range of preventive activities to limit social contact even before these were officially recommended. The first case of infection in Slovenia was reported on 4 March. On 13 March, when this survey began, the Slovenian government declared an epidemic. With a prevalence of 67.8 cases/million population, Slovenia was ranked 11th on that day (3). On 14 March, the country reported its first coronavirus death. The government adopted confinement measures, including the temporary closure of nursery schools and schools, and the temporary prohibition of public gatherings.
Preventive measures such as mask wearing, hand hygiene practices, social distancing, case detection, contact tracing, and quarantine were proposed to reduce infection transmission (4, 5) according to the available data on virus infectivity (6), its persistence, and the efficacy of chemical (7) or heat disinfection (8, 9). The elderly and those with certain underlying medical conditions, which required more attention and care (6, 10), especially in clinical environments (11), were the most vulnerable population. Potential drugs are still being tested in clinical trials. Infected persons therefore mainly receive symptomatic treatment and supportive care (4, 6).
Preventive behaviour can be explained by the evolutionary disease-avoidance model that proposes a behavioural immune system as a proactive mechanism that protects us against the infection, working in tandem with the physiological immune system (12, 13, 14, 15). The behavioural immune system is a motivational system that helps minimise infection risk by changing cognition (e.g. vulnerability beliefs, group stereotypes), affect (e.g. disgust, fear, worry, anger) and behaviour (e.g. illness avoidance) in ways that promote pathogen avoidance. It is marked by both contextual sensitivity and biases that aid adaptive response (16).
In the context of disease-avoidance mechanisms, fear is associated significantly with specific attitudes and reactions (17). Anxiety appears to have a highly degree of impact on preventive behaviour. People with high levels of anxiety are likely to develop maladaptive behaviours, such as excessive avoidance and the persistent, repetitive and unnecessary seeking of medical reassurance, which is an added burden on an already overtaxed healthcare system during a pandemic. On the other hand, it is important to note that health anxiety, defined as an enduring tendency or trait, varies along a continuum. People with very low health anxiety may fail to engage in basic hygiene or other recommended health measures because they do not perceive their health to be at risk. In times of a pandemic, such people are liable to spread infections. In the context of planning preventive measures to reduce the scope of an epidemic, such psychological elements are often overlooked when considered within the context of medical diseases (18).
The purpose of this study, performed in the early stage of the spread of Covid-19 in Slovenia, was to evaluate respondents’ opinions on the efficacy and use of preventive measures, and to assess how these opinions were associated with their anxiety level, perceived vulnerability to disease, germ aversion and a number of demographic characteristics.
The research was based on a survey in which data was obtained through the voluntary participation of anonymous participants. A questionnaire exploring knowledge and perception of different aspects of the epidemic in Slovenia was used to identify the factors influencing the taking of preventive measures in the early phase of coronavirus spread. The study protocol was reviewed and approved by the Ethics Committee of the Faculty of Health Sciences, University of Novo mesto (approval no FZV-98/2020128).
The online survey was distributed using snowball sampling, which is in line with similar previous studies (19, 20, 21). The initial group of respondents was contacted via project members’ professional (healthcare professionals) and personal contacts (general population), and the link to the survey was disseminated further via the Facebook social network. The participants were asked to complete a structured, self-administered electronic questionnaire. The online survey began on 13 March at 2.20 pm and was active for 24 hours.
The survey was accessed by 18,760 individuals: 12,305 responded with 8,023 responses appropriate for further analysis, and 7,764 completed the questionnaire in full. There were 79% female and 21% male respondents. Their ages ranged from 13 to 83 years (mean 40.5 years, SD 13.2; male participants: mean age 41.2 years, SD 13.63; female participants: mean age 40.36 years, SD 13.16), and 4.8% were aged 65 years or over. Forty-five per cent of respondents had secondary school education, 41% had an undergraduate degree, and 15% a postgraduate degree. Fourteen per cent of respondents worked in the healthcare sector. One quarter of respondents reported having household member(s) over 65 years old, 52% reported having children in the household and 14% reported having both. Among the respondents, 17 had had a coronavirus infection, 1% had been in contact with an infected person and 10% knew somebody that had been infected (but had not had direct contact with that person).
Ranking of opinions on the efficacy of different preventive measures against coronavirus infection.
| Preventive measure | Estimated efficacy – respondents’ opinion |
Share of respondents considering preventive measure to be efficient or very efficient (N=7875) |
|---|---|---|
| Mean rank a higher rank means a more efficient measure |
% | |
| 8.59 | 92.4 | |
| 7.68 | 84.2 | |
| 7.62 | 83.7 | |
| 7.44 | 78.5 | |
| 5.27 | 33.0 | |
| 4.92 | 27.3 | |
| 3.84 | 14.3 | |
| 3.41 | 14.6 | |
| 3.40 | 10.2 | |
| 2.83 | 7.1 |
The opinions on preventive measures were obtained using a cluster of ten preventive measures (listed in Table 2) that were assessed from 1 (absolutely inefficient) to 5 (absolutely efficient) (Cronbach’s alpha=0.7).
Ranking of preventive measures being taken on the day the epidemic was declared.
| Preventive measure | Estimated efficacy (Friedman’s test) |
Share of respondents that agreed or absolutely agreed with the preventive measure statement (N=7934) |
|---|---|---|
| Mean rank a higher rank means a more frequently used preventive measure |
% | |
| 7.37 | 97.0 | |
| 7.32 | 97.3 | |
| 6.92 | 93.5 | |
| 6.74 | 89.6 | |
| 6.55 | 89.9 | |
| 5.41 | 67.8 | |
| 5.00 | 70.8 | |
| 4.59 | 59.3 | |
| 2.82 | 24.9 | |
| 2.26 | 19.9 |
For the assessment of the preventive behaviour, respondents evaluated their level of agreement with ten statements (listed in Table 2) from 1 (absolutely not) to 5 (absolutely yes). For further analysis, the sum was calculated to obtain the preventive behaviour score (Cronbach’s alpha=0.75).
The Thermometer for Mental Health was used to assess the psychological burden, i.e. one’ own experience of physical, emotional, psychosocial burden, and the burden of everyday life during the preceding seven days (22). Respondents had to assess these on a continuous visual scale from 0 (no burden) to 10 (extremely strong burden). To assess anxiety, the seven-item, GAD-7 Generalized Anxiety Disorder (23) was used. GAD-7 consists of seven questions based in part on the DSM-IV (Diagnostic and Statistical Manual of Mental Disorders, 4th edition) criteria for GAD (generalised anxiety disorder), and reflects the frequency of symptoms during the preceding two-week period. For each symptom queried, it provides the following response options: “not at all,” “several days,” “over half the days” and “nearly every day”, and these are scored 0, 1, 2 or 3 respectively (Cronbach’s alpha=0.92).
The Perceived Vulnerability to Disease Questionnaire (PVDQ), developed by Duncan et al. (24), a 15-item self-report using a seven-point scale response (with endpoints labelled as “strongly disagree” and “strongly agree”) was used. It measures two factors: perceived infectability (assesses beliefs in one’s own susceptibility to infectious diseases, e.g. “If an illness is going around, I will get it”; seven items, Cronbach’s alpha=0.68) and Germ Aversion (assesses emotional discomfort in contexts where disease-causing germs might be transmitted, e.g. ”It really bothers me when people sneeze without covering their mouth”; eight items, Cronbach’s alpha=0.56). The first translation from English to Slovenian was done by a group of Slovenian psychologists, with a backward translation provided by an English reader, as recommended by the guidelines on the translation of tests. One item from Germ Aversion was changed from ”I avoid using public telephones because of the risk that I may catch something from a previous user” into ”I avoid using public transport or other objects for public use because of the risk...”, to be more reflective of current reality, as in recent research in China (25).
To assess an association between numerical variables, a Pearson correlation was performed, and ANOVA was used to assess the differences between the groups in terms of their demographic characteristics. For ranking analysis, a Friedman non-parametric ANOVA was computed. IBM SPSS Statistics for Windows, Version 25.0 was used for all analyses. A p-value of <0.05 was considered statistically significant.
Non-parametric ANOVA (Friedman’s test) showed statistically significant different opinions on the efficacy of preventive measures (p<0.001). Quarantine ranked the highest, followed by hand disinfection (and hand washing). The least efficient were seasonal flu vaccination, herbal infusions and essential oils.
The non-parametric ANOVA (Friedman’s test) shows a statistically significant difference in the preventive measures taken (p<0.001). The respondents adhered strictly to hand and cough hygiene (>97%), avoided crowded places (93.5%) and people at risk of infection (89.6%), and two thirds preferred to stay at home. They regularly disinfected their hands (89.9%) and objects (70.8%), almost 60% prepared food supplies for at least two weeks, and one quarter used face masks.
We computed the scores of the variables in the questionnaire concerning preventive behaviour and the scales used: PVDQ (for both factors: Perceived Infectability and Germ Aversion), Thermometer for Mental Health to assess psychological burden, and GAD-7. The scores were also analysed with regard to the respondents’ different demographic characteristics (Table 3).
Descriptive statistics for analysed scores (preventive behaviour, perceived infectability, germ aversion, psychological burden and anxiety) and ANOVA between the analysed scores and demographic characteristics (results presented as mean score (SD); shading represents a statistically significant difference between the demographic categories with ANOVA – level of significance p<0.05)
| Variables | N | Preventive behaviour | Germ aversion | Perceived infectability | Anxiety | Psychological burden | |
|---|---|---|---|---|---|---|---|
| 1,642 | Male | 36.0 (5.95) | 4.68 (1.02) | 3.33 (0.92) | 3.12 (4.12) | 3.65 (2.50) | |
| 6,089 | Female | 38.9 (5.19) | 5.06 (1.01) | 3.41 (0.98) | 4.94 (5.16) | 4.56 (2.47) | |
| 3,460 | Undergraduate | 38.3 (5.70) | 4.92 (1.04) | 3.43 (0.95) | 4.52 (5.12) | 4.25 (2.52) | |
| 1,383 | Bachelor’s degree or equivalent | 37.9 (5.21) | 5.04 (1.00) | 3.39 (1.00) | 4.56 (4.90) | 4.38 (2.50) | |
| 2,418 | Master’s degree or equivalent | 38.4 (5.35) | 5.01 (1.03) | 3.35 (0.99) | 4.60 (4.96) | 4.53 (2.48) | |
| 400 | Doctoral degree or equivalent | 38.1 (5.31) | 5.04 (0.97) | 3.31 (0.94) | 4.64 (4.93) | 4.41 (2.56) | |
| 857 | Yes | 38.7 (5.43) | 4.97 (1.01) | 3.45 (1.03) | 4.59 (5.16) | 4.66 (2.51) | |
| 6,316 | No | 38.2 (5.52) | 4.97 (1.02) | 3.38 (0.96) | 4.53 (4.99) | 4.34 (2.50) | |
| 517 | Unemployed | 38.8 (5.20) | 4.99 (1.09) | 3.47 (0.97) | 4.80 (5.12) | 4.16 (2.58) | |
| 714 861 | alone with older>65 | 37.7 (5.98) 38.7 (5.84) | 4.84 (1.05) 4.95 (1.00) | 3.38 (1.03) 3.47 (0.93) | 4.04 (4.74) 4.72 (5.47) | 4.15 (2.56) 4.50 (2.53) | |
| 2,905 | with children | 38.5 (5.26) | 4.99 (1.00) | 3.31 (0.94) | 4.47 (4.89) | 4.42 (2.52) | |
| 1,108 | with older>65 and children | 38.8 (5.65) | 5.01 (1.03) | 3.48 (1.00) | 5.00 (5.34) | 4.46 (2.47) | |
| 2,106 | with partner or other | 37.8 (5.34) | 5.00 (1.06) | 3.44 (0.98) | 4.56 (4.90) | 4.26 (2.48) | |
| p | <0.001 | 0.003 | <0.001 | 0.001 | 0.006 | ||
| 885 | Yes | 39.0 (5.54) | 5.07 (1.00) | 3.51 (1.02) | 5.66 (5.60) | 4.77 (2.57) | |
| 6,846 | No | 38.2 (5.48) | 4.96 (1.03) | 3.38 (0.96) | 4.41 (4.91) | 4.31 (2.49) | |
Preventive measures were taken more strictly by women and by respondents living in a household with elderly people aged over 65 (either with or without children), whereas the level of education or whether a respondent worked in the healthcare sector did not have any effect on the measure-taking score. We also found a strong (p<0.001) positive correlation between age and the taking of preventive measures.
People who knew somebody who had been infected had higher perceived infectability, greater germ aversion, were more anxious, took more preventive measures, and felt greater psychological burden compared to those who had not undergone this experience. The same was also characteristic of those living in a shared household with people aged over 65 and those living with both elderly persons and children.
We found strong positive correlations between preventive behaviour and all tested scores. Greater preventive behaviour was found among individuals who experienced greater psychological distress, were more anxious, and expressed greater perceived infectability and greater germ aversion (Table 4).
Correlation between preventive behaviour score and perceived infectability, germ aversion, psychological burden and anxiety scores.
| Pearson correlation of preventive behaviour score with: | Pearson correlation | Sig. (2-tailed) |
|---|---|---|
| 0.094 | p<0.001 | |
| 0.288 | p<0.001 | |
| 0.130 | p<0.001 | |
| 0.128 | p<0.001 |
Statistical analyses also revealed a significant correlation between anxiety and psychological distress (r=0.555, p<0.001), and more anxious individuals also expressed greater perceived infectability (r=0.267, p<0.001) and germ aversion (r=0.153, p<0.001).
The purpose of this study was to evaluate people’s opinions on the efficacy of preventive measures and the extent to which taking preventive measures was associated with people’s perception of the situation, their anxiety level, including perceived vulnerability to disease, and a number of demographic characteristics in the initial phase of the Covid-19 epidemic. Infection prevention and control was a widely highlighted topic in research publications in the initial stage of the spread of Covid-19 as reviewed by Adhikari et al. (4). However, there has been no study on how preventive measures are applied. Compliance with preventive measures is one of the most important factors that should be considered in epidemic modelling in order to cover all the aspects that could influence it (26).
Preventive measures that were ranked by the respondents as more efficient were those recommended in the current pandemic situation (4, 5, 10, 27): quarantine, disinfection and washing of hands, disinfection of surfaces and objects, and mask wearing. The measures traditionally used in the treatment of respiratory infections, such as herbs (28) and vitamin C (29) were ranked lower; herbal teas and essential oils were ranked even lower than antiviral drugs, since there are none available against Covid-19, along with the influenza vaccine, which targets another virus.
The results of this study show a very strong association between taking preventive measures and levels of anxiety, the self-assessed level of psychological burden, perceived infectability and greater germ aversion. Greater preventive behaviour is found in individuals who experience greater psychological distress, are more anxious, and express greater perceived infectability and greater germ aversion. A higher anxiety level was associated with increased use of preventive measures against Covid-19, which has also been described by Wong et al (30).
Germ aversion and pathogen disgust sensitivity were also the two variables most consistently associated with Covid-19 concern and preventive health behaviour in a similar study by Schook et al. (31).
However, a recently published study by Wang et al. (32), which also investigates the association between the psychological impact of the outbreak and the taking of precautionary measures in China, showed opposite findings, i.e. greater adherence to preventive measures was linked to lower anxiety, stress and depression.
In our study, perceived infectability (3.40) is comparable to the European average (3.48) (17, 24), whereas germ aversion is much higher (4.98) than the expected European average (3.55). This is likely influenced by the current situation and the real threat posed by Covid-19, which has been confirmed by Diaz Martinez et al., who state that the germ aversion score is more situationally conditioned (33). This can explain the higher scores on the Germ Aversion scale in those who know someone already infected with Covid-19 (5.07).
Women expressed significantly higher levels of preventive behaviour, but also higher levels of anxiety, perceived vulnerability and germ aversion compared to men, which accords with other studies that suggest the existence of gender-based differences in immune behaviour (24, 34, 35).
Those residing in a household together with family members aged over 65 and those who live with both children and elderly persons expressed the highest level of anxiety, together with higher perceived psychological burden and infectability, and the lowest opinion on the efficacy of preventive measures. It is certainly the case that the fact that the highest mortality rate is among the elderly and that children are often asymptomatic (but are nevertheless possible carriers of the virus) (6) can worsen distress in such households.
With SARS-CoV-2, the greatest risk is transmission to healthcare workers (5). The observation that medical workers are facing enormous pressure and severe challenges, which include a high risk of infection, has already been reported for Hubei, China (6). In our study, healthcare workers did not declare higher levels of preventive behaviour and germ aversion, nor did they express a higher level of anxiety compared to other respondents. However, they did admit to higher levels of psychological burden and perceived infectability.
More anxious individuals express greater mental distress, as well as greater belief in their own susceptibility to infectious diseases and much more emotional discomfort in contexts that connote an especially high potential for pathogen transmission. Higher anxiety and perceived vulnerability apparently lead to a higher level of preventive behaviour, which is desirable during an epidemic. However, in the case of very high anxiety, it can lead to excessive health anxiety that gives rise to various maladaptive behaviours (18).
Our results show that at the time the epidemic was declared, as many as 20.5% of respondents were anxious, far exceeding the expected prevalence of anxiety disorders, which is 10.4% (7–15.5%) in Euro/Anglo cultures (36). Increased anxiety during the Covid-19 epidemic has also been described by other authors (32, 34).
Mental distress is expected to continue to deepen as the epidemic continues, also as a result of social isolation and associated distress and, in many cases, increased loneliness.
Quarantine was evaluated as the most efficient preventive measure, and people took it relatively seriously even before it became an official measure. It is to be expected that while social distancing is required to slow the spread of the coronavirus, it may have detrimental effects on an individual’s physical and psychological health in the long run, as social isolation increases the risk of coronary heart disease, stroke and mortality, and can have a negative impact on psychological health, leading to depressive symptoms (37).
We should acknowledge some of the limitations of the survey we used. These include the oversampling of a particular network due to the random snowball sampling method, e.g. 79% of participants were women. However, Ekman et al. (38) claim that the bias associated with collecting information using online questionnaires is no greater than that caused by paper questionnaires. Nevertheless, it should be pointed out that although this survey is not representative of the entire adult population in Slovenia, we did manage to cover all age groups and educational levels.
While prevention and control was a widely highlighted topic in research publications in the early stage of the spread of Covid-19, there has been no study on the extent to which preventive measures are applied. This study reveals that quarantine has been evaluated as the most efficient preventive measure and that people took it relatively seriously even before it became an official measure. Women, those living with elderly people over the age of 65 and those who knew an infected person took preventive measures more seriously.
This study also reveals a significant association between adherence to preventive measures and anxiety, perceived psychological burden, perceived vulnerability and germ aversion. These findings may have implications for the development of interventions intended to increase preventive health behaviours. A follow-up study of these phenomena in the next stages of the epidemic will show how attitude towards and compliance with preventive measures will change, and whether the changes occur interdependently with anxiety and other scores that reveal one’s perception of the situation.