The COVID-19 pandemic has had a substantial impact on the lives and well-being of people globally and has created new stressors and disruptions to daily living for people around the world.Reference Chew, Lee and Tan 1 This includes concern for one’s own health and that of one’s close relations, as well as constant exposure to information about the pandemic and its adverse effects.Reference Forte, Favieri and Tambelli 2 Policy measures implemented by authorities to limit the spread of the disease, including stay-at-home orders, have also resulted in limits on individuals’ movements, decreased social contact, and adverse economic effects on individuals and countries at large.Reference Marčinko, Jakovljević and Jakšić 3 Many have suffered feelings of isolation and helplessness due to the enormity of the pandemic as well as control measures adopted.Reference McIntyre and Lee 4
Evidence suggests that outbreaks of communicable diseases can cause individuals to experience increased mental health problems, including symptoms of psychosis, trauma, suicidal ideation, and panic.Reference Tucci, Moukaddam and Meadows 5 , Reference Wu, Chan and Ma 6 Key among these problems is anxiety, which is a distressing feeling of nervousness or tension linked to physical changes such as increased blood pressure, sweating, trembling, dizziness, or a rapid heartbeat and defensive behaviors such as avoidance.Reference Major, Cozzarelli and Horowitz 7 Studies during the COVID-19 pandemic have showed increased levels of anxiety in the general population and in specific groups such as students and health workers.Reference Mrklas, Shalaby and Hrabok 8 -Reference Cao, Fang and Hou 15 In studies of anxiety in Canada and India conducted during the pandemic, 47% and 28% of respondents, respectively, were moderately to severely anxious.Reference Mrklas, Shalaby and Hrabok 8 , Reference Verma and Mishra 10
Differences in rates of anxiety between populations may be due to the timing of the inquiry around a particular outbreak—whether early in the outbreak, at the peak, or in the post-outbreak period. Other differences could be attributable to contextual factors, demography, and health system factors. The psychological impact of COVID-19 is higher in places with a higher prevalence of COVID-19 and can worsen with increased media exposure.Reference Moghanibashi-Mansourieh 9 , Reference Özdin and Bayrak Özdin 13 , Reference Nwachukwu, Nkire and Shalaby 16 Moreover, individuals who reside in settings with recently imposed quarantine and prolonged restrictive measures, including lockdowns, are more likely to experience anxiety.Reference Tee, Tee and Anlacan 17 , Reference Tran, Nguyen and Auquier 18 Consistent with the patterns of the prevalence of anxiety and other mental health issues more generally, women have been more affected by anxiety during the pandemic than men.Reference Mrklas, Shalaby and Hrabok 8 , Reference Moghanibashi-Mansourieh 9 , Reference Özdin and Bayrak Özdin 13 , Reference Lai, Ma and Wang 14 , Reference Tee, Tee and Anlacan 17 , Reference Zhang, Wang and Rauch 19 Studies also find that levels of anxiety during this period have been higher in younger age groups (20-40 years)Reference Mazza, Ricci and Biondi 20, potentially due to uncertainty and concerns about employment status, since that age bracket is greatly represented in the active labor force.Reference Moghanibashi-Mansourieh 9 , Reference Goodwin, Wiwattanapantuwong and Tuicomepee 12 , Reference Nwachukwu, Nkire and Shalaby 16 In some contexts, anxiety is positively associated with employment.Reference Goodwin, Wiwattanapantuwong and Tuicomepee 12 , Reference Verma and Mishra 10 For instance, studies indicate that severe anxiety and other mental health effects are more likely in occupations with intense working conditions and risk of exposure to COVID-19, such as nursing and other frontline work.Reference Lai, Ma and Wang 14 , Reference Zhang, Wang and Rauch 19 , Reference Lee, Ng and Khong 21 , Reference Bruine de Bruin 28 In other contexts, anxiety is also associated with job loss.Reference Yang, Lin and Frost 22
Findings from current and past disease outbreaks also revealed that marital status is correlated with the likelihood of experiencing anxiety. During the pandemic, individuals who were single, separated, or widowed were more likely to experience anxietyReference Tee, Tee and Anlacan 17 , Reference Yang, Lin and Frost 22; relatedly, people with strong social support were less likely to experience anxiety.Reference Hua, Huang and Bugeja 23 Findings from current and past disease outbreaks also implicate concerns about preexisting chronic disease, health of self and family, and current or preexisting psychiatric illness as correlates of anxiety. Individuals with poor health status, including a history of existing or prior medical conditions (both physical and mental), are more likely to experience anxiety. Reference Chew, Lee and Tan 1 Reference Marčinko, Jakovljević and Jakšić 3 , Reference Tee, Tee and Anlacan 17 , Reference Boateng, Phipps and Smith 24 , Reference Hao, Tan and Jiang 25 Having a high level of confidence in doctors,Reference Wang, Pan and Wan 26 observing precautionary measures,Reference Wang, Tee and Roy 27 being satisfied with the health information received from authorities,Reference Wang, Pan and Wan 26 low perceived risk of contracting COVID-19,Reference Bruine de Bruin 28 and high perceived likelihood of survivalReference Wang, Pan and Wan 26 decreased the risk of individuals experiencing anxiety during the COVID-19 pandemic.Reference Özdin and Bayrak Özdin 13
Individuals who suffer health anxiety during a pandemic may carry out disruptive behaviors such as rushing to and from overcrowded stores or may be reluctant to seek medical assistance. Conversely, anxiety may also cause individuals to visit doctors and pursue tests repeatedly for reassurance while neglecting to seek mental health assistance, believing their condition to be somatic. Higher-than-normal levels of anxiety can decrease immune system functioning, increasing the risk of infection and of severe disease. 29 Increased anxiety is also associated with suicidal behavior, and several studies have indicated a concern for possible increases in suicide attempts during the pandemic.Reference Gunnell, Appleby and Arensman 30 , Reference Sinyor, Spittal and Niederkrotenthaler 31 To the best of our knowledge, no previous study has examined the effect of the COVID-19 pandemic on anxiety levels in the general adult population in Ghana. This study aims to determine the prevalence of likely Generalized Anxiety Disorder (GAD) as well as the sociodemographic, clinical, and other COVID-19-related factors influencing likely GAD symptoms among the general population during the COVID-19 pandemic.
Methods
The methods and information about the study site have been published in a related study.Reference Adu, Wallace and Lartey 32 We employed a cross-sectional approach, utilizing a web-based survey that included demographic, social, clinical, and COVID-19-related variables adapted from survey questions used to gather baseline data from Text4hope subscribers in Canada during the COVID pandemic.Reference Agyapong, Hrabok and Vuong 33 , Reference Hrabok, Gusnowski and Vuong 34 The study assessed symptoms of anxiety using the Generalized Anxiety Disorder 7-item (GAD-7) scale.Reference Spitzer, Kroenke and Williams 35 The survey was hosted on the Qualtrics XM platform as a web-based survey. A link to the survey was forwarded to respondents primarily through WhatsApp-based platforms, including specific groups for nurses, doctors, and students, as well as for general public groups for residents in Ghana. Data were collected in November 2020 and January 2021, in between the first and second COVID-19 waves.
With an estimated population in Ghana of 30 million people, using the sample size calculator (available online: https://www.surveymonkey.com/mp/sample-size-calculator/ accessed on May 1, 2021), the sample size needed to estimate the prevalence for likely GAD with a 95% confidence interval and a 3% margin of error was 1068.
Ethics approval was obtained from the Ghana Health Service Ethics Review Committee [GHS-ERC 027/08/20]. Informed consent was sought from respondents by providing them comprehensive information about the benefits and risks of the study and asking about their agreement to participate in the study immediately after the survey was started. If they responded “no,” the survey was immediately terminated. We also included a footnote in all sections of the survey informing respondents they could terminate the survey at any time.
Data Analysis
Data were analyzed using the Statistical Package for Social Sciences (SPSS) version 26. Descriptive statistics for demographic characteristics were reported in numbers and percentages. Cross-tabular bivariate analyses with chi-square or Fisher’s exact tests were used to explore the relationship between the categorical variables and moderate/high anxiety (‘likely GAD’). Variables with a statistically significant relationship (p < 0.05, two-tailed) and variables that trended toward significance (0.05 ≤ p ≤ 0.10, two-tailed) with likely GAD on bivariate analysis were entered into a logistic regression model. Prior to performing the logistic regression analysis, correlational diagnostics were performed to identify any strong intercorrelations (Spearman’s correlation coefficient of ±0.7–1.0) among predictor variables. If high intercorrelations between two predictor variables were identified, we planned to reach a consensus among members of the study group on which variables to include in the regression model based on their relative relevance to the outcome of interest. Odds ratios from the binary logistic regression analysis were examined to determine the association between each of the variables in the model and likely GAD, controlling for the other variables in the model.
Results
Respondents were fairly balanced between males and females, with majority residing in the Ashanti Region, having a university or college degree, being married, and being employed in a government agency. The prevalence for likely GAD in our sample was higher in female respondents (10.5%) compared to male respondents (4.7%). The prevalence of likely GAD in our overall sample was 7.6% (Table 1).
Table 1. Gender distribution of demographic and COVID-19-related characteristics of respondents
Bivariate Analysis
The association between all variables examining sociodemographic factors, COVID-19 news exposure, and anxiety-related variables and likely GAD is illustrated in Table 2. Gender, employment, and relationship status, fear of contracting COVID-19, having a family member or relative sick with COVID-19, loss of job during the pandemic, level of support received from employer, and seeking mental health support were significantly associated with likely GAD (p ≤ 0.05). For example, respondents who were female, unemployed, or single had a higher prevalence of likely GAD compared to respondents with other characteristics. Similarly, respondents who expressed that they had been fearful of contracting COVID-19 as well as those who reported they had lost their jobs during the pandemic had higher prevalence of likely GAD compared to respondents with other characteristics. Other variables such as frequency of exposure to COVID-19-related news were not significantly associated with likely GAD.
Table 2. Chi-square/Fisher’s exact* tests of association between the demographic antecedents and likely GAD
Logistic Regression
The full model containing all twelve predictors was significant, Χ Reference Forte, Favieri and Tambelli 2 (21, N = 475) = 93.2, p < 0.001 indicating the model was able to distinguish between individuals who reported moderate/high anxiety symptoms versus those who reported low anxiety symptoms. The model explained between 17.8% (Cox and Snell RReference Forte, Favieri and Tambelli 2) and 44.2% (Nagelkerke RReference Forte, Favieri and Tambelli 2) of the variance. Additionally, 94.5% of all cases were correctly classified. As shown in Table 3, respondents who were female were 2.8 times more likely to have likely GAD compared to male respondents. Similarly, respondents who did not lose their job due to COVID-19 were 0.6 times less likely to present with likely GAD compared to those who lost their jobs due to COVID-19. This implies that respondents who lost their job during the COVID pandemic were 1.67 (1/0.6) times more likely to present with likely GAD symptoms compared to respondents who did not lose their job. Similarly, respondents who did not seek counseling during the pandemic were 0.13 times less likely to present with likely GAD compared to respondents who sought counseling. This also implies that respondents who sought counseling were 7.7 (1/0.13) times more likely to present with likely GAD symptoms compared to respondents who did not seek counseling.
Table 3. Logistic regression predicting likelihood for respondents to present with likely GAD
Employment status, relationship, religion, working in health care, having a family member or friend who is sick from COVID, fear of contracting COVID-19 infection, and level of employer support were not independently significantly associated with likely GAD, controlling for all other factors in the model.
Discussion
The prevalence of likely GAD in this study (7.6%) was lower than that reported in Canada (47%), where a similar survey that utilized the same scale and cutoff scores was carried out, as well as in India (28%).Reference Mrklas, Shalaby and Hrabok 8 , Reference Verma and Mishra 10 The disparity between our findings and those of the Indian study, which used the DASS tool rather than the GAD-7, may be attributed to differences in the anxiety assessment instrument used. The disparity between our findings and those reported in Canada, which used the same scale and cutoff scores to assess anxiety, could be due to differences in cultural factors and contextual factors, such as the higher rate of COVID-19 mortality and morbidity in Canada compared to Ghana, as well as differences in health systems’ responses to the pandemic.Reference Analyses 36
The prevalence for likely GAD in our sample was 10.5% in women compared with 4.7% in men. Women were 2.8 times more likely to have moderate to high anxiety symptoms as compared with men. Both biological and social factors are expected to play a role in females being more anxious about COVID-19 than males. Women are reported to be more sensitive to stress hormones and threats and less likely to use adaptive coping strategies than men.Reference Tang, Liang and Zhang 37 It has also been reported that women tend to assume more caregiving responsibilities and tend to be employed in lower paying jobs with less job security. Our findings are in line with those of other studies conducted during the COVID-19 pandemic in Iran,Reference Moghanibashi-Mansourieh 9 China,Reference Wang, Pan and Wan 38 , Reference Liu, Yang and Zhang 39 and Turkey.Reference Özdin and Bayrak Özdin 13
Losing one’s job due to COVID-19 pandemic measures was associated with likely GAD.
This finding is consistent with the work of others, who suggest that those who lose their jobs are likely to be more anxious due to their inability to cater for themselves and their families.Reference Morris, Moment and Thomas 40 , Reference Islam, Ferdous and Potenza 41 A nationally representative survey of partial lockdown districts in Ghana in June 2020 reported that some job losses were a direct result of COVID-19-related work suspensions. 42 In contrast, this study found no significant difference in likely GAD prevalence between those who were unemployed and those who were employed in a government agency; this contrasts with a Canadian study, which reported higher anxiety levels in respondents who were unemployed compared to those employed in a government agency.Reference Hrabok, Gusnowski and Vuong 34 Our study did not show that health care workers were more likely to experience GAD. Working in the health sector did not show any relationship with GAD. Our findings are in contrast with other studies that reported higher stress, anxiety, and other mental health symptoms among health workers during the pandemic.Reference Mrklas, Shalaby and Hrabok 8 , Reference Moghanibashi-Mansourieh 9 The higher level of anxiety among health-care workers elsewhere may be linked to the higher rate of COVID-19 morbidity and mortality. Another reason for relatively lower levels of anxiety among health-care workers in Ghana may be the institution of motivation packages for health-care workers during the first wave of the pandemic, which may have boosted their mental health.
Respondents who received mental health counseling were more than 7 times likely to be anxious compared with those who did not receive mental health counseling. This is consistent with longitudinal studies conducted in China by Tang et al.Reference Tang, Liang and Zhang 37 and Wang et al.Reference Wang, Pan and Wan 38 Individuals who seek counseling tend to be already mentally unwell.Reference Tang, Liang and Zhang 37 , Reference Wang, Pan and Wan 38
Relationship status did not independently predict likelihood for respondents to experience moderate to high anxiety in this study. This is consistent with findings from a Canadian study during the pandemicReference Hrabok, Gusnowski and Vuong 34 but contrasts with findings from studies conducted during the COVID-19 pandemic among nonworking women in PakistanReference Spitzer, Kroenke and Williams 35 , Reference Abbas, Aqeel and Abbas 43 and among women and men in Bangladesh,Reference Analyses 36 which indicated that respondents who were single had higher levels of anxiety, possibly due to the fear of not having companionship or support from partners if they got infected with the virus.
Having a family member or friend with COVID-19 and being fearful of contracting the COVID-19 infection did not independently predict moderate to high anxiety in respondents after controlling for other factors such as employment and region in the regression model. These findings are in contrast with those reported by Moghanibashi-Mansourieh et al.Reference Moghanibashi-Mansourieh 9 and Mrklas et al.,Reference Mrklas, Shalaby and Hrabok 8 who studied anxiety in Iran and Canada, respectively, during the pandemic. Our study did not specifically ask about contact with the infected person, which might account for the differences in anxiety levels reported with the other studies.
Strengths and Limitations of the Study
A major strength of the study is the use of GAD-7, a validated self-reported scale with high reliability for the assessment of likely GAD in the general population.Reference Garabiles, Lao and Yip 44 However, this study has some limitations. First, the study achieved a sample size of 756 instead of the anticipated sample size of 1068. As a result, the margin of error for our prevalence estimates for likely GAD increased from 3% to 4%. Second, since data collection occurred primarily with an online questionnaire, there is the possibility of selection bias since individuals without internet access, smart devices, and computers may not have been able to access the survey. Distribution of survey links on WhatsApp groups means a large section of Ghanaians who were not members of these select social media groups or affiliated with members of the group were excluded from the survey.
Third, the cross-sectional nature of the study does not allow for a direct causal relationship to be established between the variables in the regression model and likely GAD and also makes it difficult to draw conclusions regarding long-term effects of COVID-19 on likely GAD. Fourth, the demographics of our respondents are not representative of the demographics of the entire Ghanaian population; therefore, our findings on the prevalence of anxiety may not be nationally representative. For example, Ghana’s most recent census reports that about 50.7% of Ghanaians are women, whereas 51.6% of our sample are women. Also, about 42.1% of Ghanaians are married, whereas 45.6% of our sample are married.Reference Ghana 49
Also, all the variables were evaluated using self-reports and hence may suffer recall biases. Due to the large number of variables in our regression model, our study outcome may be prone to type 1 error, necessitating statistical correction. However, routine use of statistical correction methods such as the Bonferroni correction has been criticized as deleterious to sound statistical judgment and reducing the chance of a type I error at the expense of a type II error.Reference Ghaemi 48 Finally, the absence of nationally representative pre-pandemic data means that we are not able to confidently attribute the high prevalence of anxiety in our study to the pandemic.
Policy Implications
This study provides useful information on the prevalence and predictors of likely GAD among a section of the Ghanaian population during the COVID-19 pandemic. Likely GAD has a high potential negative health burden. Without prompt intervention, anxiety symptoms may evolve into long-term depression. We recommend collecting information about GAD in the general population through existing periodic national surveys (such as the Ghana Demographic and Health Survey) to inform local understanding of the determinants of GAD in Ghana generally and for relevant mitigation approaches. The government of Ghana could adopt internet-based cognitive behavior therapy (CBT), which is very cost effective, easily scalable, and geographic-location independent, in treating psychiatric symptoms to reduce mental health issues, including anxiety, among the general population in Ghana.Reference Agyapong, Hrabok and Vuong 33 , Reference Armstrong 45 , Reference Zhang and Ho 46 , Reference Ho, Chee and Ho 47 Internet-based CBT has been used in other settings to manage stress, anxiety, and depression and could serve as a useful tool in Ghana and other low- and middle-income settings during public health emergencies.Reference Agyapong, Hrabok and Vuong 33 , Reference Armstrong 45 , Reference Zhang and Ho 46
Conclusions
Our findings suggest that the pandemic may have greater effect on women, those who lost their jobs due to the COVID-19 pandemic, and those who sought mental health counseling. Priority must therefore be attached to psychological support measures for members of these groups.
