Perceived stress associated with COVID-19 epidemic in Colombia: an online survey Estrés percibido relacionado con la epidemia de COVID-19 en Colombia: una encuesta en línea Percepção de estresse relacionado à epidemia da COVID-19 na Colômbia: uma pesquisa eletrônica

This study aimed to assess the prevalence and variables related to perceived stress associated with the COVID-19 pandemic in a sample of Colombian adults using a designed online cross-sectional survey. Adults answered a version of the Perceived Stress Scale (PSS-10) modified for COVID-19 (PSS-10-C), with Cronbach alpha equal to 0.86. In total, 406 individuals aged between 19 and 88 years (M = 43.9; SD = 12.4) agreed to participate in the survey: 61.8% were females, 90.6% had a university degree, 44.1% were health professionals, and 45.7% considered public health policies for preventing the spread of the disease inconsistent with scientific recommendations. PSS-10-C scores ranged from 0 to 36 (M = 16.5; SD = 7.3); 58 individuals (14.3%) scored for high perceived stress (cut-off point = 25). The inconsistency between policies and scientific evidence was significantly related to high perception of stress associated with COVID-19 (OR = 2.36; 95%CI: 1.32-4.20), after adjusting for gender. We concluded that the study group presented the prevalence of perceived stress associated with COVID-19 at high levels, arising from the inconsistent strategies developed by health authorities in view of scientific recommendations. Further researches must address the psychosocial aspects of epidemics. COVID-19; Quarantine; Psychological Distress; Cross-Sectional Studies Correspondence A. Campo-Arias Universidad del Magdalena. Carrera 32 No 22-08, Santa Marta, Magdalena 470004, Colombia. acampoa@unimagdalena.edu.co 1 Universidad del Magdalena, Santa Marta, Colombia. 2 Pontificia Universidad Javeriana, Bogotá, Colombia. doi: 10.1590/0102-311X00090520 Cad. Saúde Pública 2020; 36(5):e00090520 This article is published in Open Access under the Creative Commons Attribution license, which allows use, distribution, and reproduction in any medium, without restrictions, as long as the original work is correctly cited. COMUNICAÇÃO BREVE BRIEF COMMUNICATION Pedrozo-Pupo JC et al. 2 Cad. Saúde Pública 2020; 36(5):e00090520 Introduction Since the first case, in early December 1, the perceived vulnerability to coronavirus infection (COVID-19) has radically changed in all countries. In Colombia, it switched from a distant threat to a real problem when the country reported the first infected, in March 2020 2. Based on other countries’ experiences, both scientific organizations and the media demanded public health decisions that were consistent with the best available evidence to control the spread of the infection 1,2. However, for financial interests, decisions were taken late 3. Data suggest that public health policies often underestimated scientific evidence and responded to political and capitalist interests 4. On March 20th, the Colombian National Government decided to prevent COVID-19 transmission: the President of the Republic decreed a quarantine 5. Quarantines have negative psychological effects, often deemed as marginal, including symptoms related to anxiety, depression, acute stress, and even manifestations of post-traumatic stress disorder 6. The restriction of mobility can act as a psychosocial stressor 7 due to prolonged confinement, difficulties in obtaining daily-life necessary supplies, potential financial losses, and access to conflicting and inadequate information on the Internet 6,8. Yet, few studies have quantified perceived stress during quarantines. In Australia, Taylor et al. 6 evaluated the frequency of and factors associated with psychosocial stress; by applying the Kessler Psychosocial Distress Scale during the equine influenza epidemic they observed that 34% of participants presented high levels of perceived distress, compared to 12% of the general population. They also reported that the psychosocial distress was significantly higher for those residing closer to infection outbreaks, workers in the equine sector, and younger and less educated people 9. Similarly, for the current COVID-19 pandemic, 52,730 people in China responded to the COVID-19 Peritraumatic Distress Index, which includes symptoms of anxiety, depression, related to stress and specific phobias, as well as physical symptoms reported during the last week. Researchers found that 35% of participants scored high for psychological distress, more frequent among women, people aged between 18 and 30 years, participants with higher education level and residents near affected areas 9. They observed that psychological distress was associated with local availability of medical services, regional health system efficiency, and measures to control the epidemic 10. Political decisions may affect emotional health; thus, measures should consider the perceived stress of citizens in emergency situations 10. These decisions seem relevant in countries like Colombia, where mistrust in public institutions explains most of the limited social capital 11. This study aimed to assess the prevalence and variables related to perceived stress associated with the COVID-19 pandemic in a sample of Colombian adults.


Introduction
Since the first case, in early December 1 , the perceived vulnerability to coronavirus infection  has radically changed in all countries. In Colombia, it switched from a distant threat to a real problem when the country reported the first infected, in March 2020 2 .
Based on other countries' experiences, both scientific organizations and the media demanded public health decisions that were consistent with the best available evidence to control the spread of the infection 1,2 . However, for financial interests, decisions were taken late 3 . Data suggest that public health policies often underestimated scientific evidence and responded to political and capitalist interests 4 .
On March 20th, the Colombian National Government decided to prevent COVID-19 transmission: the President of the Republic decreed a quarantine 5 . Quarantines have negative psychological effects, often deemed as marginal, including symptoms related to anxiety, depression, acute stress, and even manifestations of post-traumatic stress disorder 6 .
The restriction of mobility can act as a psychosocial stressor 7 due to prolonged confinement, difficulties in obtaining daily-life necessary supplies, potential financial losses, and access to conflicting and inadequate information on the Internet 6,8 . Yet, few studies have quantified perceived stress during quarantines. In Australia, Taylor et al. 6 evaluated the frequency of and factors associated with psychosocial stress; by applying the Kessler Psychosocial Distress Scale during the equine influenza epidemic they observed that 34% of participants presented high levels of perceived distress, compared to 12% of the general population. They also reported that the psychosocial distress was significantly higher for those residing closer to infection outbreaks, workers in the equine sector, and younger and less educated people 9 .
Similarly, for the current COVID-19 pandemic, 52,730 people in China responded to the COVID-19 Peritraumatic Distress Index, which includes symptoms of anxiety, depression, related to stress and specific phobias, as well as physical symptoms reported during the last week. Researchers found that 35% of participants scored high for psychological distress, more frequent among women, people aged between 18 and 30 years, participants with higher education level and residents near affected areas 9 . They observed that psychological distress was associated with local availability of medical services, regional health system efficiency, and measures to control the epidemic 10 .
Political decisions may affect emotional health; thus, measures should consider the perceived stress of citizens in emergency situations 10 . These decisions seem relevant in countries like Colombia, where mistrust in public institutions explains most of the limited social capital 11 . This study aimed to assess the prevalence and variables related to perceived stress associated with the COVID-19 pandemic in a sample of Colombian adults.

Method
The authors performed a cross-sectional study approved by the Ethics Research Committee of the University of Magdalena (Act 002-2020). The research was based on the Declaration of Helsinki and Colombian legislations.
The study was formed by a nonprobability sample. Professors and students from a public university in Santa Marta, Colombia, and health professionals were invited to participate via email. The recipients could forward the link to friends and acquaintances, in a snowball strategy. We expected at least 386 people to engage in the survey and a 50% prevalence of perceived stress associated with the COVID-19 pandemic at high levels. Inclusion criteria were: Colombian individuals, over 18 years old, and able to respond online questionnaires.
Participants filled out demographic information on age, gender, education level, marital status, employment, and socioeconomic status. They were further asked to answer yes or no to: whether they were health professionals; suffered from a chronic condition (comorbidities such as diabetes); followed the precautionary measures; and considered the public health policies for preventing the spread of the disease consistent with scientific recommendations.
Several Colombian studies have used the previous Spanish version of the PSS-10, which has shown acceptable internal consistency, with Cronbach alpha between 0.65 and 0.86 15,16 . In our study, the COVID-PSS-10 presented a high internal consistency, with Cronbach alpha equal to 0.86.
Between 8:00p.m. on March 21 and 8:00p.m. on March 23, 2020 (the first weekend of the quarantine in Colombia), the leading researcher sent 60 emails among his contacts. The questionnaire was disclosed by email, Facebook, and WhatsApp.
In the descriptive component, frequencies and percentages were computed for categorical variables, and mean (M) and standard deviation (SD) were calculated for quantitative variables. Table 1 shows all dichotomized variables. To establish the crude associations between the quantified variables and high perceived stress associated with COVID-19, odds ratios (OR) were estimated with a 95% confidence interval (95%CI). Associations were adjusted using logistic regression. The IBM-SPSS Statistics 23.0 (https://www.ibm.com/) helped in the analysis.

Table 1
Demographic characteristics of the participants.

Variable
Frequency % Regarding high perceived stress associated with COVID-19, scores ranged between 0 and 36 (M = 16.5; SD = 7.3); 58 participants (14.3%) scored for high perceived stress, which was significantly related to the inconsistency between adopted policies and scientific evidence. After adjusting for gender, the relationship remained significant (OR = 2.36; 95%CI: 1.32-4.20). Table 2 presents other associations.

Discussion
In total, 15% of the participants scored for high perceived stress associated with COVID-19, which was significantly related to participants' perception on the inconsistency between scientificallyverified recommendations and adopted public health measures by the government authorities. In this study, the prevalence of high perceived stress was relatively lower than in previous studies. In Australia, Taylor et al. 6 found that 34% of participants reported high perceived distress during an equine influenza quarantine. Qiu et al. 9 reported that 35% suffered from high psychological distress during the last week of the current COVID-19 epidemic. The discrepancy in prevalence may be explained by the spectrum of symptoms explored by each research instrument and the epidemic context 17 .
In our research, high perceived stress was solely associated with the perceived inconsistency between government arrangements and scientific recommendations, compatible with other papers. In China, Qiu et al. 9 reported that measures adopted by health authorities to control the epidemic were associated with perceived distress. The data suggest that reliance in national health authorities outlines emotional responses during epidemic situations 11 .
Other studies noticed that high psychosocial stress was more common among young people 9,10 . Although our study expected a higher prevalence of perceived stress among younger people, due to their information overload by social networks 8 , it was similar among the three age groups.
Findings regarding education are inconsistent: in Australia, distress was higher among less educated people 6 ; whereas in China, stress was higher among more educated participants 9 . Studies currently under developments aim to clarify the role of education in the emotional response to crises such as COVID-19 epidemic.
Other variables may mediate emotional responses to epidemics. In this investigation, the perceived stress did not depend on gender, contrarily to a previous study performed in China, which found that women reported greater psychosocial distress than men 9 , suggesting that biological and sociocultural factors associated with gender may mediate the association 18 .
Our study omitted proximity to most affected areas by the epidemic and access to medical services as modifiers to emotional responses 9,17 . We found health professionals to equally report perceived stress when compared to other jobs 9,10 , contrarily to the equine epidemic in Australia, in which healthcare professionals showed a significantly higher psychological distress 9 . This disparity may be explained by biosecurity measures followed by health personnel in hospitals 19 .
The emotional response to epidemics and their control measures relate to several variables, many of which have been little investigated. Our results suggest that a distal variable, as conceived in the model of social determinants of health, such as mistrusting government institutions, may portray a stress factor for citizens 11,19 .
This research provides new information on acute emotional responses to quarantine, which must be considered by health authorities alongside the epidemiological aspects of the infection. Emotional aspects are essential for infected people, their family members, and healthy citizens who must remain quarantined 5 . Yet, this study present some limitations and potential biases: snowball sampling is nonrandom, so results should be interpreted with caution; the chosen method is undershoot, so that it did not reach the general population; and the authors adapted an existing scale to measure perceived stress associated with COVID-19 20 .
In conclusion, 15% of the participants reported high perceived stress associated with COVID-19. High levels of stress were related to perceived inconsistent strategies adopted by health authorities in view of scientifically-verifired recommendations. Further researches must address the psychosocial aspects of epidemics.