This paper is part of a larger cross-sectional survey that employed a convergent parallel mixed method using both quantitative and qualitative data collection methods.^14,15 This report covers only an aspect of the quantitative method. The survey was conducted between June 15, 2020, to July 07, 2020.

The study was conducted in six contiguous districts in the Bono East region of Ghana: Kintampo North Municipal (KNM), Kintampo South District (KSD), Techiman South Municipal (TSM), Techiman North District (TND), Nkoranza South Municipal (NSM), and Nkoranza North District (NND) where the Kintampo Health and Demographic Surveillance System (Kintampo-HDSS) administered by the Kintampo Health Research Centre operates. Information about how the Kintampo-HDSS operates has been described in an earlier publication in detail.¹⁶

The study area has an estimated population of about 430,728 with males (47.8%) and females (52.2%). The various districts constitute; KNM & KSD (163,191 (38%)), NND & NSM (93,805 (22%)), and TND & TSM (173,732 (40%)) of the total population. The study area is served by 9 districts hospitals, 35 health centres, and 214 functional Community-based Health Planning and Services (CHPS). Other health care dynamics of the area are described in other publications.¹³ The study setting is largely rural and subsistence farming is the major occupation. The population is multi-ethnic and majority of them reside in rural communities. Other population indicators of the study area have been described in other publications in detail.^{16, 17–21}

Based on a survey that assessed the knowledge, attitudes and preventive practices on Ebola Virus Disease within the Kintampo districts of Ghana, 83% had knowledge on the virus disease. Using a prevalence of 0.83, 95 percent confidence level, and an error margin of 0.0324. The study calculated the sample size using the formula:

where ρ is the prevalence of the study, alpha is the significance level and epsilon is the error margin of the study. A sample of 517 is generated when the formula was employed. In order to enhance and prevent sampling biases, the sites were sampled proportionally. The proportional division of the sampling distribution according to the size of each study site is a follow; Kintampo North & South (196 (38%)), Nkoranza North & South (112 (22%)), and Techiman North & South (209 (40%)).

The 517 sample was randomly sampled from the Kintampo- HDSS database. All resident adult members of the population who are registered in the Kintampo-HDSS, aged 18 years and above were eligible to be included in the study. Non-registered resident and registered resident members who were below 18 years were not considered in the sampling.

A close-ended questionnaire was administered to participants by trained interviewers. The questionnaire was pretested before the start of the actual data collection. Each questionnaire consisted of 9 modules including questions to elicit avoidance, discrimination, anxiety and fear.

The questionnaire for data collection was designed using Survey Solutions Version 19.05 and was deployed on android tablets. Data validation checks like branching logics, range and consistencies were incorporated in the design of the data collection tool (This prevented missing data by ensuring that all fields were answered before the data was uploaded). Geographic Position System (GPS) coordinates features were captured automatically to provide evidence that interviews were actually conducted in locations that could be verified. The completed data were synchronized daily unto the database after cross-checking the response for consistency, completeness and accuracy from the tablets. Data was updated with the resolved queries then formatted and sent for statistical analysis using STATA version 14.0. Descriptive statistics such as frequencies and percentages were used to describe the data. Chi-square test was conducted between the main avoidance variable “I will not have anything to do with someone suspected of COVID-19” and selected socio-demographic characteristics of respondents, and also between the main variable for avoidance and other avoidance related variables at a statistically significance level of 0.05. Tables were generated with Microsoft Excel and presented in Word format.

The demographic characteristics of the sample compares very well with the HDSS population that it was taken from (Table 1). The majority of the participants were females (55%), and this is similar to the HDSS population (53%) from which the sample was taken. Participants ranged from 18 to 60 years of age with an average age of thirty-nine years, and this is also similar to the adult HDSS population from which the sample was taken. Majority of the respondents were above the age of forty years followed by those within the age group of thirty to forty years. About 39% of the respondents did not have any formal education which compares well with the HDSS population (36%). However, three percent of the respondents have had education up to the tertiary level. Farming or domestic worker was the main occupation of the respondents with about 43% of them falling into this category. This was followed by trading with clerical or secretarial being the least occupational status. Close to seventy percent of the respondents (68%) were married and less than 1 in 5 respondents (18%) were single. More than half of the respondents interviewed were Muslims.

From the survey about 60% of the respondents agreed that they will not have anything to do with a person suspected of COVID- 19. However, about 37% of the respondents disagreed with this assertion as shown in Table 2. On the other hand, a lower proportion of respondents (29%) agreed they will not have anything to do with someone who has recovered from COVID-19 compared to about 60% of respondents agreeing that they will not have anything to do with someone suspected of COVID-19 as shown in Table 2. Again, from Table 2, almost all of the respondents interviewed (98%) affirmed they feel compassionate towards people who have contracted COVID-19. About 67% however agreed they fear to be isolated if suspected to be infected with COVID-19. Half of the respondents interviewed agreed that there is little they can do to help COVID-19 patients. About 91% of the respondents agreed that people within the community need to adopt a far more tolerant attitude towards people that recovered from the COVID- 19 infection. Table 2 provides further detailed description of avoidance and discrimination against persons suspected of COVID-19 within the survey area. Other descriptive analysis of avoidance and discrimination by socio-demographic characteristics of respondents are provided in the Tables 3 and 4.

Respondents with some level of education showed willingness to have something to do with a suspected COVID-19 patient than those with no education. About 67% of respondents with no formal education compared to 25% with university education agreed to have nothing to do with suspected COVID-19 patient as shown in Table 3, whilst more males (65%) agreed not to have anything to do with suspected COVID-19 patient compared to females as shown in Table 3.

Respondent’s educational level appeared to be associated to one having something to do with recovered COVID-19 patient as shown in Table 3. Older people, forty years and above (61%) agreed not to have anything to do with suspected COVID-19 patients compared to those who are younger Table 4. From Tables 3 and 4, respondents’ attitude towards recovered patients appears positive irrespective of gender, age group or educational level.

From Table 5, a chi-square test results showed that there is no significant relationship between age group and occupation and willingness to associate with persons suspected to have contracted COVID-19, but the educational level of an individual significantly determines whether the person is willing to associate him/herself with persons suspected to have contracted COVID-19. A Chisquare test also showed a strong significant association between “I will not have anything to do with someone suspected of COVID- 19” and other avoidance/discrimination variables. “I will not have anything to do with someone suspected of COVID-19” is also significantly associated with “I will not have anything to do with someone who has recovered from COVID-19” and “there is little I can do to help people who have COVID-19. Other strongly significant association between “I will not have anything to do with someone suspected of COVID-19” and other avoidance/discrimination variables are provided in Table 5.

This study did not measure stigma specifically but provides the opportunity to investigate stigma against COVID-19 suspected persons. Following the findings from this study, we recommend a more in-depth study of COVID-19 stigma. Self-reported method used in this study is also a limitation as respondent may provide socially desirable responses that may bias the results. The study reports on data collected from Kintampo-HDSS catchment area which may not necessarily apply to other areas in Ghana.