key: cord-0730402-9309aig5
authors: van der Velden, Peter G.; Marchand, Miquelle; Cuelenaere, Boukje; Das, Marcel
title: Pre-outbreak determinants of perceived risks of corona infection and preventive measures taken. A prospective population-based study
date: 2020-04-06
journal: nan
DOI: 10.1101/2020.04.01.20049957
sha: 686135083f06748330ae6bcd283556b00ca6925c
doc_id: 730402
cord_uid: 9309aig5

Objectives. Assess how people perceive the risks of coronavirus infection, whether people take preventive measures, and which (pre-outbreak) factors contribute to the perceived risks and measures taken, such as (pre-outbreak) respiratory problems, heart problems, diabetes, anxiety and depression symptoms, loneliness, age, gender, marital status and education level. Methods. Data were collected in the longitudinal LISS panel, based on a random sample of the Dutch population. The coronavirus survey started on March 2, and the data collection ended on March 17 2020. Data were linked with surveys on health and social integration conducted at the end of 2019 (Nstudy sample=3,540). Results. About 15% perceived the risk of infection as high, and 11% the risk becoming ill when infected. Multivariate logistic regression analyses showed the following. Older age-groups perceived the risk for coronavirus infection as lower (all adjusted Odd Ratios [aOR] smaller than .058). In total, 43.8% had taken preventive measures, especially females (aOR=1.47, 95% CI=1.27-1.69). Those with lower education levels less often used preventive measures (aOR=0.56, 95% CI=0.46-0.69). Those with pre-outbreak respiratory problems (aOR=2.89, 95% CI=2.24-3.73), heart problems (aOR=2.17, 95% CI=1.49-3.16) and diabetes (aOR=3.10, 95% CI=2.02-4.74) perceived the risk becoming ill when infected as higher than others. However, respondents with pre-outbreak respiratory problems and diabetes did not more often take preventive measures. Conclusions. Vulnerable patients more often recognize that they are at risk becoming ill when infected by the coronavirus, but many do not take preventive measures. Interventions to stimulate the use of preventive measures should pay additional attention to physically vulnerable patients, men and those with lower education levels.

On December 31 2019, the WHO China Country Office was informed of cases of pneumonia with a then unknown etiology. The Chinese authorities identified the etiology: a new type of corona virus (COVID-19 or SARS-CoV-2) which was isolated on January 7 [1] . In the first two months after the first report, 79,968 persons in China were infected by the virus (confirmed cases [2] . The number of confirmed cases across the globe on March 1 2020 was raised to 87,137. With respect to the spectrum of the disease, Wu and McGoogan [3] reported that, based on the 44,415 confirmed cases in China, 81% was mild, 14% severe and 5% critical. The overall case-fatality rate (CFR) in China was 2.3% (among 44,472 confirmed cases). Meanwhile, the corona virus outbreak also severely affects the production facilities, transport, the global economy and financial markets.

To prevent and reduce infection by the new coronavirus (COVID-19 or SARS-CoV-2) health organizations such as the WHO, governmental health agencies and journals offer information about possible preventive measures [1] [2] [3] [4] [5] . To target and implement interventions to stimulate preventive behaviour against infection, more insight is needed in how people perceive the risks of being infected by this new coronavirus, if they use of preventive measures, and especially which (pre-outbreak) factors determine the perceived risks and measures taken [6] . However, to the best of our knowledge to date peer-reviewed population-based studies assessing the perceived risks of corona infection, measures and their determinants are absent. Aim of the present population-based study conducted is to shed light on this gap of scientific knowledge.

With respect to perceived risks, we made a distinction between risk for infection and risk of becoming ill when infected [9] . With respect to potential determinants, we first focused on pre-outbreak respiratory, heart problems and diabetes because they increase the risk for severe health problems when infected [8] .

We furthermore assessed pre-outbreak anxiety and depression symptoms, and loneliness because they may impact the perceived threat of infection and perceived likelihood to become ill when infected [9] [10] [11] [12] . We finally assessed demographics such as age because older people are more at risk to become ill [6] , and sex and education level. This study is conducted in the Netherlands and during the data collection period 

The study was conducted using the Dutch Longitudinal Internet studies for the Social Sciences (LISS) panel [13] . Netherlands. The 16 profiles were constructed using the following demographic characteristics: sex (2 categories), age categories (4 categories) and marital status (2 categories) totalling 2*4*2=16 exclusive demographic profiles. In case a profile in our study sample differed from the general population, a weighting factor was computed and applied. All results are based on the weighted sample.

According to the Dutch Medical Research Involving Human Subjects Act (WMO) the present study did not require ethical approval. In accordance with the General Data Protection Regulation, participants gave explicit consent for the use of the collected data for scientific and policy relevant research.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.01.20049957 doi: medRxiv preprint

The Corona survey (T2) started with the following brief introduction "The next question are about the new corona virus. There is currently an outbreak of this virus in China. Now, also people in the Netherland and in other countries have become ill".

We administered two questions, developed for this study, to gain insight in how adults perceived the risks of the coronavirus. Respondents were asked: What do you think is the chance that you … in the next two months?: 1.) become infected with this coronavirus, and 2.) get severely ill, if you become infected with this coronavirus.

Both questions had a 7-points answer scales (see Table 2 ).

After completing these questions, respondents were asked "In the past two months did you do things to prevent infection by this coronavirus as much as possible? (1= yes, 2=no)". In case respondents answered "yes", they were asked to indicate what they exactly did. The answer categories were (partly) based on WHO recommendations ((1=the purchase of mouth masks, 2=wash hand more often and longer, 3= not going to certain (busy) places, 4= cancelled a journey, 5=otherwise, namely, (open answer category)). When respondents answered "no", they were asked why not (1=because I

do not know what I should do, 2=but maybe I will do this still, 3=because I have not thought about it yet, 4=because I find it nonsense, 5= because, namely; open answer category). For both questions respondents could choose for more than one answer.

The Health survey (T1 b ) assessed several Physician-diagnosed Diseases (PD) in the past year (1= yes, 2=no) and Health Problems (HP) respondents regularly suffer from (1=yes, 0=no). For the present study we focused on reported: 1.) respiratory problems ((PD=chronic lung disease such as chronic bronchitis or emphysema or asthma) or (HP=short of breath, problems with breathing, or coughing, a stuffy nose or flurelated complaints)); 2.) heart problems ((PD=angina, pain in the chest a heart attack including infarction or coronary thrombosis or another heart problem including heart failure) or (HP =heart complaints or angina, pain in the chest due to exertion); and 3.) diabetes (PD=diabetes or a too high blood sugar level).

Loneliness at T1 a was assessed using the six-item De Jong Gierveld Loneliness Scale (Cronbach's Alpha=.85) [14] . Respondents are asked to rate items such as 'I often feel deserted' and 'there are enough people I can count on in case of a misfortune' on three-point Likert scales (1 = yes, 2 = more or less, 3 = no). We calculated the total score after recoding the three negative formulated items and lower scores reflect more loneliness. For the present study we dichotomized scores into low (≥ 15) and high loneliness (≤ 14). About 20% of the respondents have scores of 14 or lower (two lowest percentiles).

Anxiety and depressive symptoms in the past months were examined at T1 b using the 5-item Mental Health Index or Inventory (MHI-5) [15, 16] . The MHI-5 ask respondents to rate the presence of symptoms during the past month on 6-point Likert . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint . https://doi.org/10.1101/2020.04.01.20049957 doi: medRxiv preprint scales (1 = never to 6 = continuously). A cut-off of ≤ 59 was used to identify respondents with moderate to high anxiety and depression-symptom levels (Cronbach's Alpha=.86) [17] .

We monitored when respondents completed the corona questions.

Chi-square tests and multivariable logistic regression analyses were conducted with pre-outbreak medical health problems, symptoms, loneliness, demographics and time elapsed as predictors, and perceived risks and measures taken as dependent variables.

The elapsed time was recoded into 1=0-4 days, 2=5-9 days, and 3=10-15 days. Due to low cell counts in the extremes of perceived risks (see Table 1 ), we recoded the perceived risks into the following three categories. To optimize readability, hereafter we label these three categories of perceived risks as low (no to small chance), medium (between small and big chance) and high (big chance to absolute certain). After this recoding we assessed to what extent the predictors were associated with the perceived medium and high risk.

A similar strategy was used to assess which factors were associated with whether respondents took preventive measures People may perceive the risks as high and therefore take measures, but the opposite may also be true. People may perceive the risk as lower because they take measures. Since the perceived risks and preventive measures taken were assessed at . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint . https://doi.org/10.1101/2020.04.01.20049957 doi: medRxiv preprint the same time, we therefore did not add the perceived risk to the list of predictors in the multivariable logistic regression analyses predicting preventive measures taken.

All analyses were conducted with IBM SPSS version 26. Table 1 provides an overview of the characteristics of the weighted study sample, e.g. the prevalence of pre-outbreak health problems, symptoms, loneliness and demographics. The increase in respondents after day 9 can be attributed to the reminder mail. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint . https://doi.org/10.1101/2020.04.01.20049957 doi: medRxiv preprint

In Table 2 shows that a minority (15.0%) perceived the risk of being infected as high.

A somewhat lower proportion perceived the risk for becoming ill when infected as high (10.6%). On the other hand, very few respondents perceived the risk of infection and becoming ill as zero (4.4% and 5.5% respectively). . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint . https://doi.org/10.1101/2020.04.01.20049957 doi: medRxiv preprint

The results of the chi-square test and the stepwise multivariable regression analyses are presented in Table 3 . We focus on the results on the stepwise regression analyses (adjusted Odds Ratios). They show that respondents with pre-outbreak respiratory and heart problems more often perceive the risk of infection as medium and high than respondents without these health problems. Anxiety and depression symptoms and loneliness were not independently associated with the perceived risk. Older and low educated respondents less often perceived the risk of infection as high than younger respondents and higher educated respondents respectively. Respondents who participated later, more often perceived the risk of infection as medium and high than those who participated in the first 4 days. Females more often than males perceived the risk of infection as medium and high. Respondents who participated later and females, more often perceived the risk of infection as high.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint . https://doi.org/10.1101/2020.04.01.20049957 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint . https://doi.org/10.1101/2020.04.01.20049957 doi: medRxiv preprint Table 4 contains the results of the same analyses but with the perceived risk for becoming ill when infected in the next two months as dependent variable (right side). On a bi-variate level, almost all predictors were significantly associated. The multivariable analyses showed that respondents with pre-outbreak physical health problems, anxiety and mental health problems and loneliness, more often perceived the risk for becoming ill when infected as high than others. Older respondents more often, in contrast to the perceived risk of infection, perceived the risk for becoming ill as medium and high than younger respondents.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint . https://doi.org/10.1101/2020.04.01.20049957 doi: medRxiv preprint 

Of the total study sample, 43.8% took preventive measures (see Table 2 ) such as washing hands more often and longer (92.2%), not going to work of avoid certain (busy) places (53.6%), purchase of mouth masks (5.9%) and cancelled a journey (8.2%). Of the respondents who did not take preventive measures, 42.5% reported that they find it nonsense or useless, 24.9% that maybe will do this still, 20.4% have not thought about it yet, and 15.4% that they do not know what they should do. Table 5 shows which factors predicted the use of preventive measures against infection by the coronavirus. With respect to pre-outbreak physical health problems:

only respondents with heart problems took preventive measures more often. Females more often took preventive measures, and medium and high educated respondents more often than low educated respondents. Finally, respondents who filled in the survey more recently, more often took preventive measures.

We repeated the regression analyses among those who participated 10-15 days after the start of the corona survey, showing almost similar results. Having heart problems was no longer significantly associated with preventive measures.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint . https://doi.org/10.1101/2020.04.01.20049957 doi: medRxiv preprint 

Main results of this prospective population based-study are that during the 2-week study period (March 2 to March 17 2020) the number of respondents who perceived the risk of being infected by the coronavirus as high, increased sharply (10% to 44%).

Multivariable logistic regression analyses showed that respondents with pre-outbreak respiratory and heart problems, diabetes, anxiety and depression symptoms and loneliness, and older respondents more often perceived the risk becoming ill when infected as high. In line with the increased perceived risk to be infected, the number of respondents who took preventive measures increase too. However, respondents with pre-outbreak respiratory and heart problems did not more often take preventive measures than others. In addition, analyses of respondents who participated 10-15

after the the start of the study showed that respondents with respiratory and heart problems and diabetes did not differ in the proportion of people who took preventive measures.

Our findings are somewhat similar to the results of a study reported by the WHO Regional Office for Europe [6] . This serial cross-sectional study conducted in Germany in almost the same period as our study (week 10 and 11 2020) showed that the prevalence of respondents who perceived the risk to be infected by the coronavirus as high, increased from 16.8% to 21.4%. They furthermore reported, like us, that older respondents (60+) felt less likely be infected. Importantly, in our study the effects of other factors that are associated with the perceived risk of corona infection were controlled for such as pre-outbreak respiratory and heart problems, and education level. Asmundson and Taylor [18] reported that, according to polls, in the US 56%

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. To date many studies on our research topic are initiated and conducted.

However, when finalizing this study we were not aware of any peer-reviewed population-based studies on perceived risks, the use of preventive measures and their pre-outbreak determinants, to compare our findings with.

Strength of the present study are the use of a large traditional probability based sample drawn from the Dutch population, the prospective study-design, data on preoutbreak physician-diagnosed diseases, and use of well validated instruments on anxiety and depression symptoms, and loneliness.

We deliberately choose to use the data that was collected in the first two weeks of the survey (response was 70.1%), to be able to share our results rapidly given the threatening global developments. However, although we distinguished three subsequent periods during these two weeks suggesting an increase in preventive measures taken, we do not know from this study if and when all respondents have taken preventive measures. In addition, we do not know from this study to what extent respondents who have taken preventive measures, will continue to comply with protection guidelines from governmental health agencies. Another limitation is that we not were able to include children. It is unknown to what extent children's perceptions of the risks and the measures they taken resembles those of adults and especially parents and other family members. We did not systematically examine whether respondents were in quarantine, e.g. were separated and restricted in movement because they had been potentially infected by the coronavirus and their . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint . https://doi.org/10.1101/2020.04.01.20049957 doi: medRxiv preprint effects on perceived risks [19] . The present study does not provide information on this topic, nor how quarantine affects post-quarantine preventive behaviour.

Nevertheless, we believe that our results are also of relevance for future outbreaks. 

The results of this study, based on a random sample of the general adult population, are partly reassuring and positive, and partly negative. Positive is the finding that the number of respondents who have taken preventive measures during the brief 2-weeks study period increased, while taking other significant predictors of the use of preventive measures into account. It is very likely that the daily stream of information . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.01.20049957 doi: medRxiv preprint about the pandemic and advice on this matter provided by Dutch governmental health agencies, physicians and media, contributed to this finding. A negative finding is that respondents with respiratory problems and diabetes, who are considered groups at severe risk for complicated health problems when infected, did not take preventive measures more often than others. In addition, we found no indications that people took preventive measures irrespective of their education level and sex. The last findings suggest that specific education level and sex-related interventions should be developed and offered to increase preventive behavior among men and those with a lower education level.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.01.20049957 doi: medRxiv preprint

Novel Coronavirus (2019-nCoV) SITUATION REPORT-1, 21

Novel Coronavirus (2019-nCoV) SITUATION REPORT-41

Characteristics of and Important Lessons From the Coronavirus Disease 2019 (COVID-19) Outbreak in China: Summary of a Report of 72 314 Cases From the Chinese Center for Disease Control and Prevention

World Health Organisation (WHO)

Pre-emptive low cost social distancing and enhanced hygiene implemented before local COVID-19 transmission could decrease the number and severity of cases

WHO Regional Office for Europe. Guidance and protocol. Rapid, simple, flexible, cost effective behavioural insights on COVID-19

Meta-analysis of the relationship between risk perception and health behavior: the example of vaccination

If You Are at Higher Risk

The relationship between psychological states and health perception in individuals at risk for cardiovascular disease

Social disconnectedness, perceived isolation, and symptoms of depression and anxiety among older Americans (NSHAP): a longitudinal mediation analysis

Risk Perception and Anxiety Regarding Radiation after the 2011 Fukushima Nuclear Power Plant Accident: A Systematic Qualitative Review

Immediate Psychological Responses and Associated Factors during the Initial Stage of the 2019 Coronavirus Disease (COVID-19) Epidemic among the General Population in China

True longitudinal and probability based internet panels: evidence from The Netherlands

A 6-item scale for overall, emotional, and social loneliness confirmatory tests on survey data

An efficient method of identifying major depression and panic disorder in primary care

The MOS 36-item short-form health survey, SF-36: conceptual framework and item selection

Een beschrijving van de MHI-5 in de gezondheidsmodule van het Permanent Onderzoek Leefsituatie [A Description of the MHI-5 in the Health Module of Permanent Research of Living Conditions

Fear and the 2019-nCoV outbreak

The psychological impact of quarantine and how to reduce it: rapid review of the evidence