Disparities and Factors Associated with Coronavirus Disease-2019-Related Public Stigma: A Cross-Sectional Study in Thailand

Coronavirus disease 2019 (COVID-19)-related public stigma is a major challenge, with scarce available evidence. This study aimed to determine the disparities and factors associated with COVID-19-related public stigma in the Thai population. We conducted a cross-sectional study involving a voluntary online survey in Thailand from 21 April 2020 to 4 May 2020. We invited 4004 participants to complete a series of questionnaires, including the validated COVID-19 public stigma scale and questions on relevant COVID-19-related psychosocial issues. Multinomial logistic regression was performed to investigate the factors associated with COVID-19-related public stigma. The prevalence of COVID-19-related public stigma was 24.2% (95% confidence interval [CI], 22.2–26.2) for no/minimal, 35.5% (95% CI, 33.4–37.6) for moderate, and 40.3% (95% CI, 38.2–42.4) for high. We observed disparities in the prevalence of COVID-19-related public stigma according to participant characteristics and psychosocial factors. Using the no/minimal group as a reference group, the six predominant risk factors significantly associated with a moderate and high degree of COVID-19-related public stigma were middle-aged or older adults, male, divorced/widowed/separated, current quarantine status, moderate/severe fear of COVID-19, and medium/high perceived risk of COVID-19. Additional risk factors significantly related to a high degree of COVID-19-related public stigma were religion (Buddhist), region of residence (non-capital city), and exposure to COVID-19-related information. Disparities in COVID-19-related public stigma due to sociodemographic and psychosocial issues are frequent in the Thai population. To reduce public stigmatization, early identification of vulnerable groups and the development of tailored mitigation strategies should be implemented during the pandemic.


Introduction
The incidence of psychosomatic illness has increased since the emergence of coronavirus disease 2019 . These conditions are expected natural psychological This was a cross-sectional analytical study based on the Health Outcomes and Mental Health Care Evaluation Survey, under the Pandemic Situation of COVID-19 (HOME-COVID- 19). The details of the protocol have been published elsewhere [8]. In brief, an open, online, voluntary survey encompassing a set of questionnaires was sent via the SurveyMonkey ® platform, which limits to one-time participation per unique Internet Protocol address. The samples were invited by a convenience and snowball sampling strategy from all the regions in Thailand through various social media networks including public websites, Facebook, LINE, Twitter, and Instagram. Eligible participants included (i) Thai citizens aged ≥18 years at the date of the survey, (ii) permanent residents or non-residents with work permits, (iii) those who could read and communicate in the Thai language, and (iv) those who could access the Internet. We excluded incomplete surveys and surveys that took <2 min or >60 min to complete. The current analysis was restricted to only wave I of information from 21 April 2020 to 4 May 2020 (during the national government's protocols under lockdown in Thailand).
Under the HOME-COVID-19, this current study was approved by the Committee of Research Ethics of the Faculty of Public Health (ET010/2020) and the Faculty of Pharmacy (23/2563), Chiang Mai University. All participants provided written informed consent for the first page of the questionnaire. This study was in line with the Strengthening the Reporting of Observational Studies in Epidemiology Statement [9] and Improving the Quality of Web Surveys: The Checklist for Reporting Results of Internet E-Surveys [10].

Sample Size
The sample sizes for both prevalence and factors related to public stigma in the Thai population were estimated using (i) the overall mean ± standard deviation (SD) of COVID-19-related public stigma (based on the validated COVID-19 Public Stigma Scale (COVIDSS)) of 24.2 ± 7.6 [11], specified type I error at 0.05, and d equal to 0.5, a total sample size of 891; and (ii) linear multiple regression, R 2 deviation from zero with a small effect size of 0.02, type I error of 0.05, 90% of power, and anticipated total predictors equal to 15, with a sample size needed of 1192 [12]. The sample size of our survey met both requirements and was sufficient to address the research questions.

Assessment Tools and Potential Risk Factors
Participants were asked to complete a set of questionnaires regarding COVID-19related public stigma and relevant psychosocial issues as follows: • Public stigma: COVID-PSS comprises ten items with three factors (stereotypes, prejudice, fear), and a possible score range of 10-50 points. The COVID-PSS revealed acceptable psychometric properties in the Thai population, with Cronbach's α of 0.85. The degree of public stigma was established and classified as no/minimal (≤18 points), moderate (19-25 points), or high (≥26 points) [11]. • Perceived social support: The Multidimensional Scale of Perceived Social Support (MSPSS-12) consists of 12 items that measure individual perceptions of external social support. This scale has excellent internal consistency, with a Cronbach's α of 0.92 [13]. Perceived social support was categorized as low (12-35 points), moderate (36-60 points), or high (61-84 points). • Resilient coping: The Brief Resilient Coping Scale (BRCS) consists of four items to capture tendencies to cope with stress in a highly adaptive manner. The BRCS revealed satisfactory reliability (Cronbach's α = 0.80) [8]. For scale interpretation, BRCS scores was classified as low-(≤13 points), medium-(14-16 points), or high (≥17 points) resilience copers [14].

•
Fear of COVID-19 and perceived risk of COVID-19 infection: A numerical rating scale (NRS) of 0-10 points was used to measure the degree of fear or perceived risk of COVID-19 infection. The degree of fear or perceived risk was classified as no/minimal fear or low perceived risk (0-3 points), moderate fear or medium perceived risk (4-6 points), and severe fear or high perceived risk (7-10 points). • A set of potential risk factors for public stigma, including sociodemographic characteristics (age, sexual identity, marital status, educational level, occupation, religion, region of residence, living status, personal income, reimbursement scheme, history of mental illness, chronic non-communicable diseases (NCDs)), and issues-related to the COVID-19 pandemic (economic burden (income loss, financial problems), duration of exposure to COVID-19-related information, confirmed cases in the community, quarantine status, and working from home status).

Statistical Analyses
All analyses were performed using Stata 14.0 (StataCorp, LP, College Station, TX, USA). Two-tailed tests were conducted with a type I error rate of 0.05. Respondents with missing data or incomplete data were excluded from the analysis. Descriptive statistics are expressed as frequency and percentage or mean ± SD, with a range (min-max). We categorized COVID-19-related public stigma into three groups according to total scores (no/minimal, moderate, and high degrees of stigma). Baseline participant characteristics, according to the degree of public stigma, were assessed using analysis of covariance for continuous data and Fisher's exact test for categorical data. We applied survey weights to all analyses to ensure that our results represented the national population and rate of Internet use based on the National Statistic Office of the Thai Ministry of Information and Communication Technology.
We estimated the prevalence rate of COVID-19-related public stigma with 95% confidence intervals (CIs) and assessed the variation in these rates by participant characteristics. Using a trend test analysis, non-overlapping 95% CIs (p for trend <0.05) indicated a statistical difference in the prevalence rates across participant characteristic strata. We applied a two-stage multinomial logistic regression approach to determine factors associated with the degree of COVID-19-related public stigma (using no/minimal as a reference group). In the first stage, the crude association between participant characteristics and the degree of public stigma (moderate or high) was analyzed using univariable multinomial logistic regression models to identify candidate risk factors. Next, candidate risk factors with a p-value < 0.200 were included in the multivariable multinomial logistic regression models using the stepwise backward method. Variance inflation factors were used to identify multicollinearity in the final model. Moreover, an ancillary analysis was performed to confirm the robustness of the main analysis using multivariable linear regression to explore the linear relationship between the potential risk factors and COVID-PSS-public stigma score. The effect estimates of the risk factor models are expressed as odds ratios (ORs) or beta coefficients with 95% CIs.

Overview of Participant Characteristics
A total of 4997 people were identified through an online survey invitation. Of those, 4381 (87.7%) were willing to participate in the survey, and 4004 participants met the eligibility criteria and completed a set of mental health and psychosocial questions (completeness rate of 92.6%, Figure 1). Participant characteristics are described in Supplementary Table S1. Most participants were female (65.4%), with a mean age of 29.1 ± 10.8 years (range 18-79). Most participants resided in a non-capital city or its environs (64.4%). Most participants had moderate/high perceived social support and were medium/high resilient copers, whereas most participants reported having moderate/severe fear of COVID-19 and medium/high perceived risk of COVID-19 infection.

Risk Factors Associated with COVID-19-Related Public Stigma
With respect to participant characteristics (using no/minimal public stigma as a reference group), the univariable multinomial regression identified 16 candidate risk factors with p-value < 0.200 (Table 2). Subsequently, the final model based on multivariable multinomial regression models revealed six independent significant risk factors of moderate degree of COVID-19-related public stigma: (i) ages 31   <0.001 † The effect estimates are presented weighted. ‡ To includes diabetes mellitus, hypertension, dyslipidemia, stroke and heart disease, chronic kidney disease, chronic lung disease, and cancer. § The currency exchange on the survey period was 1 USD = 32.5 Baht. Abbreviations: CI, confidence interval; COVID-19, coronavirus disease-2019; COVID-PSS, coronavirus disease-2019 Public Stigma Scale; NCDs, non-communicable diseases; OR, odds ratio.
With respect to the ancillary analysis, the findings showed consistent results for the set of factors associated with COVID-19-related public stigma (R 2 = 0.58), except for exposure to COVID-19-related information, which was not significant. On the other hand, resilient coping perception has become a significant protective factor for COVID-19-related public stigma, with a small effect size (beta coefficient of −0.76; 95% CI, −1.46 to −0.05; p = 0.036 for medium resilient copers; −0.82, 95% CI, −1.51 to −0.14; p = 0.019 for high resilient copers; Supplementary Table S2).

Discussion
Our findings highlight the prevalence and disparities of COVID-19-related public stigma in the general Thai population. We found that COVID-19-related public stigma was common during the pandemic in Thailand. Critically, the estimated prevalence of a high degree of COVID-19-related public stigma was frequent at 40.3% (95% CI, 38.2-42.4); this rate was highly variable by participant characteristics and psychosocial issues regarding the pandemic. Moreover, our risks set findings for the development of a medium/high degree of COVID-19-related public stigma can provide information on the target-specific population and minimize public stigmatization in public health settings.
Few studies have addressed COVID-19-related public stigma in the general population owing to the lack of a validated tool. These studies have shown that COVID-19-related public stigma is a common phenomenon in several countries [15][16][17][18][19][20]. Similar to our study, the prevalence of moderate and high degrees of COVID-19-related public stigma accounts for more than half of the surveyed population in many countries. Furthermore, our findings underscore that participant characteristics as well as psychosocial issues during the pandemic are significantly associated with the degree of COVID-19-related public stigma.
Collectively, based on common risk factors, our findings revealed that middle-aged adults or older (31-50, ≥51 years) had a higher risk of being at a moderate/high degree of COVID-19-related public stigma than young adults. Not surprisingly, the older population, particularly in advanced age and with multi-morbidities, is at risk of a severe and critical stage if infected with COVID-19, leading to greater awareness of infection. Males have also been reported to have more severe COVID-19 [21], which may be recognized as at risk of experiencing moderate/high COVID-19-related public stigma than others. Interestingly, the explanation may be related to ideas of masculinity, in which norms are social rules that expect men to be strong, and may invoke behaviors showing responsibility towards protecting their family and community [22]. According to previous reports [23,24], we found that cohabitants are more aware of COVID-19 infection due to fear of infecting their partners. Additionally, married people, particularly healthcare workers, illustrated more personalized stigma and had concerns about public attitudes [24]. People who experienced the current quarantine status certainly perceived a higher risk of COVID-19 infection and had a higher stigma score in our observation. Some studies have shown that quarantine cases are prone to self-stigma and stigmatization by society [23]. In addition, a study involving quarantined healthcare workers also reported that guilt towards family members and friends leads to avoiding contact with neighbors and the community [25]. Finally, both perceived fear and risk of COVID-19 infection were recognized as strong factors contributing to COVID-19-related public stigma in our study. Indeed, the feeling of fear and subsequently the perceived risk of a newly emerging infectious disease usually arises from the uncertain and unpredictable course of the disease. Technology, including the internet and social media, creates "infodemics" spreading the news of COVID-19 cases, mortality, and its communicability. This can accelerate more fear and perceived dangerousness to people. Moreover, stigma from perceived risk can also be mediated by fear of COVID-19.
Apart from the common factors, some unique variables, including religion-Buddhism, living in a non-capital city, and exposure to COVID-19-related information, are associated only with a high degree of COVID-19-related public stigma. In Thailand, Buddhism is the religion of most of the population (93.5%), followed by Muslims (5.4%), and Christians and others (1.1%) [26]. Buddhists are associated with the doctrine of cultivating compassion to attenuate prejudiced attitudes towards other social groups [27]. In the case of the COVID-19 pandemic, comparison with the irreligious population may reflect aspects including liberality and acceptance of the behaviors of others, including mistakes or errors. However, some hidden residual factors, such as borderline personality disorder, narcissism, or carelessness towards religion, predominantly in the younger population, which are associated with lower stigmatization [28,29], were not investigated in our study. Therefore, this finding requires further confirmation. For residential areas, we postulated that people living in the non-capital area had a higher degree of COVID-19-related public stigma because of their fear and perceived risk of COVID-19 infection spreading from the capital city, because most cases in Thailand at the time of data collection were based in the capital city and its environs. Theoretically, media exposure has been suggested as a potential factor for perceived stigmatization among people at a high risk of contagion [30]. In this case, we can determine that media images and influences may lead to prejudice and discrimination related to COVID-19, resulting in violence against some ethnic groups, such as Asian people in the US and globally.

Strengths and Limitations
To our knowledge, this is the first study to report on the prevalence of COVID-19related public stigma in Thailand. This study was based on a nationwide survey with a large sample size. Unlike previous studies, we used the validated COVID-PSS to measure COVID-19-related public stigma among the Thai population [11]. However, our results should be used with consideration of some limitations. First, we used Wave I of the HOME-COVID-19, representing only the early phase of the pandemic when circumstances could be different from other periods, for example, due to the availability of the COVIDvaccine. We also lacked information regarding knowledge of COVID-19 infection, in which misinformation or lack of knowledge may generate more fear and anxiety about the disease and increase stigma. Although stigma changes over time and context, the results of this study are believed to be beneficial for future emerging infectious diseases. Second, our findings were based on an open online survey; therefore, information bias should be considered. In addition, it may be generalized only to those with access to the Internet. Third, fear of COVID-19 and perceived risk of COVID-19 infection was assessed via a nonvalidated NRS of 0-10 points questionnaire which may limit comparison to other settings, or international comparisons with respect to these issues. However, NRS-unidimensional assessment is practical, reasonable, and applicable for capturing participants' feelings or opinions via a public survey. Fourth, despite an ancillary analysis confirming risk factors in line with the main analysis, uncertainty with respect to exposure to COVID-19-related information, resilient coping perception, and the risk of COVID-19-related public stigma need to be confirmed in further studies. Fifth, further associations between public stigma and adverse mental health (i.e., anxiety, stress, and depression) are warranted to address public health concerns. Lastly, the longer-term effect of stigma research needs to be studied, because our findings reflected only the short-term effects, and the impact of stigma may change as the pandemic evolves.

Implications for Public and Future Research
Given the high burden of mental health and psychosocial issues during the pandemic, it is crucial to minimize COVID-19-related public stigma due to the negative consequences of stigma, including an unwillingness to disclose COVID-19 infection or to test and seek treatment. Recently, a randomized trial in the general United States population by [31] suggested that video-based interventions involving reliable information on COVID-19 prevention strategies, video encouraging digital social activity, and video sensitizing to COVID-19-related stigma are effective in reducing COVID-19-related public stigma. However, its utility could be limited by its generalizability to other populations and cross-cultural adaptation to larger public health effects. To help the target population and supplement the previous intervention trial, our findings support proactive mental health surveillance by identifying the person who may be vulnerable or at risk of public stigmatization middle-aged adults or older, male sex, married/domestic partnership, divorced/widowed/separated, Buddhist, living in the non-capital city, current quarantine status, high perceived fear or risk of COVID-19 infection. To promote mental health wellbeing, multimodal mitigation strategies involving public health education and knowledge, programs for empowering and supporting vulnerable populations, and anti-stigma policies enforced in legal legislation should be promptly implemented during the pandemic.

Conclusions
COVID-19-related public stigma is highly prevalent and varies among the Thai populations. The results of this study highlight the disparities in the prevalence of COVID-19related public stigma according to sociodemographic and psychosocial issues. Our study also shows the possibility of identifying vulnerable groups and participants who are at risk of stigma during the pandemic, which should be targeted by strategies aimed at mitigating the impact of public stigma on health.