Significant outcomes
-
Findings indicate a significant decrease in well-being and an increase in psychopathology from before to during the COVID-19 pandemic.
-
Financial loss, endorsing substance use, and direct/personal contact as very important coping strategies, were significantly associated with worsening of both well-being and psychopathology.
-
Being of Black African ethnicity and endorsing studying/learning as a very important coping strategy were protective against worsening of both well-being and psychopathology.
Limitations
-
The cross-sectional design of the study does not allow for conclusions about causality to be made.
-
The convenience sample and lack of representative participation of all racial/ethnic groups limits the generalisability of findings.
-
Retrospective recall of pre-pandemic status may be vulnerable to recall bias.
Introduction
Since its outbreak, the COVID-19 virus has infected over 771 million individuals and resulted in 6.9 million deaths globally up until 22 October 2023 (World Health Organization, 2023). In South Africa, over 4 million individuals have been infected with the virus (National Department of Health, 2023). In response to the outbreak, the South African government implemented a nationwide lockdown in March 2020 with stringent measures put in place to contain the spread of the virus. These measures included restrictions on movement, closure of schools and non-essential businesses and services, a travel ban, social distancing, and wearing of face masks in public (Salim & Karim, Reference Salim and Karim2020; Obasa et al., Reference Obasa, Singh, Chivunze, Burgess, Masiye, Mtande, Ochieng, Chalwe, Mokgatla, Rennie and Moodley2021; Rwafa-Ponela et al., Reference Rwafa-Ponela, Price, Nyatela, Nqakala, Mosam, Erzse, Lalla-Edward, Hove, Kahn, Tollman and Hofman2022). Five levels of lockdown were introduced with corresponding restrictions. Lockdown restrictions eased gradually between March (level 5) and September (level 1) in 2020 (South African Government, 2023). Following the implementation of the lockdown, many South Africans endured social and financial hardships. For example, the number of employed persons declined by 2.2 million from quarter 1 to quarter 2 of 2020 (Statistics South Africa, 2020). There was a rise in gender-based violence (GBV), with 2,320 cases reported in the first week of the lockdown (African Union Commission - Women, Gender and Development Directorate (AUC-WGDD), United Nations Entity for Gender Equality and the Empowerment of Women (UN Women), Office of the United Nations High Commissioner for Human Rights (OHCHR) and United Nations Population Fund (UNFPA), 2020). Moreover, health services prioritised COVID-19 cases over other illnesses (Obasa et al., Reference Obasa, Singh, Chivunze, Burgess, Masiye, Mtande, Ochieng, Chalwe, Mokgatla, Rennie and Moodley2021) and access to chronic medications was limited, particularly in informal settlements and rural areas, disrupting care for people with chronic conditions (Human Sciences Research Council (HSRC), 2020).
Although the COVID-19 pandemic was recognised as an infectious disease crisis, associated restrictive measures had a significant impact on well-being and mental health globally (United Nations, 2020). Psychological responses to the pandemic included increased anxiety, stress, depression, and sleep disturbances (Luo et al., Reference Luo, Guo, Yu, Jiang and Wang2020; Wang et al., Reference Wang, Pan, Wan, Tan, Xu, Ho and Ho2020; Kinser et al., Reference Kinser, Jallo, Amstadter, Thacker, Jones, Moyer, Rider, Karjane and Salisbury2021; Nguse & Wassenaar, Reference Nguse and Wassenaar2021; Rwafa-Ponela et al., Reference Rwafa-Ponela, Price, Nyatela, Nqakala, Mosam, Erzse, Lalla-Edward, Hove, Kahn, Tollman and Hofman2022). According to a systematic review, high rates of symptoms of anxiety (6.33% to 50.9%), depression (14.6% to 48.3%), PTSD (7% to 53.8%), psychological distress (34.43% to 38%), and stress (8.1% to 81.9%) were reported in the general population across eight countries during the COVID-19 pandemic (Xiong et al., Reference Xiong, Lipsitz, Nasri, Lui, Gill, Phan, Chen-Li, Iacobucci, Ho, Majeed and McIntyre2020). Another study found that adults in South Africa reported significantly higher levels of anxiety, depression, and stress symptoms compared to individuals in the United Kingdom, Australia, and Ireland (Meaklim et al., Reference Meaklim, Junge, Varma, Finck and Jackson2021). In South Africa, psychosocial support was disrupted in both urban and rural settings, and the use of mental healthcare services decreased considerably due to fear and anxiety among patients of contracting the virus, loss of income and transportation difficulties (Salim and Karim, Reference Salim and Karim2020; Nguse & Wassenaar, Reference Nguse and Wassenaar2021; Rwafa-Ponela et al., Reference Rwafa-Ponela, Price, Nyatela, Nqakala, Mosam, Erzse, Lalla-Edward, Hove, Kahn, Tollman and Hofman2022).
Although the pandemic impacted most individuals, vulnerable groups were disproportionally impacted. Studies in high-income countries identified several risk factors for poor mental health and well-being during COVID-19, including being female, under 30, single or separated, lower education level, pre-existing health conditions, low SES, financial loss, and exposure to COVID-19-infected individuals (Rahman et al., Reference Rahman, Hoque, Alif, Salehin, Islam, Banik, Sharif, Nazim, Sultana and Cross2020; Wang et al., Reference Wang, Pan, Wan, Tan, Xu, Ho and Ho2020; Nagasu et al., Reference Nagasu, Muto, Yamamoto and Hashimoto2021; Banna et al., Reference Banna, Al Sayeed, Kundu, Christopher, Hasan, Begum, Kormoker, Dola, Hassan, Chowdhury and Khan2022; Wang et al., Reference Wang, Pan, Wan, Tan, Xu, Ho and Ho2020). Conversely, protective factors included resilience (Gilleen et al., Reference Gilleen, Santaolalla, Valdearenas, Salice and Fusté2021), taking precautionary measures (e.g., hand hygiene, wearing a mask; Wang et al., Reference Wang, Pan, Wan, Tan, Xu, Ho and Ho2020) and positive coping strategies such as faith-based support (Budimir et al., Reference Budimir, Probst and Pieh2021), keeping busy (Fuller & Huseth-Zosel, Reference Fuller and Huseth-Zosel2021), and engaging in exercise (Ogueji et al., Reference Ogueji, Okoloba and Demoko Ceccaldi2022). However, limited research exists in low- and middle- income countries (LMICs), particularly South Africa. In a resource-constrained setting, it is integral to identify the risk and protective factors of well-being and mental health during a pandemic to guide future intervention and prevention strategies in times of crises.
Moreover, studies of COVID-19 have primarily assessed mental health and well-being during the pandemic (Cao et al., Reference Cao, Fang, Hou, Han, Xu, Dong and Zheng2020; Gloster et al., Reference Gloster, Lamnisos, Lubenko, Presti, Squatrito, Constantinou, Nicolaou, Papacostas, Aydın, Chong, Chien, Cheng, Ruiz, Garcia-Martin, Obando-Posada, Segura-Vargas, Vasiliou, McHugh, Höfer, Baban, Dias Neto, Nunes da Silva, Monestès, Alvarez-Galvez, Paez-Blarrina, Montesinos, Valdivia-Salas, Ori, Kleszcz, Lappalainen, Ivanović, Gosar, Dionne, Merwin, Kassianos, Karekla and Francis2020; Alonso et al., Reference Alonso, Vilagut, Mortier, Ferrer, Alayo, Aragón-Peña, Aragonès, Campos, Cura-González, Emparanza, Espuga, Forjaz, González-Pinto, Haro, López-Fresneña, Salázar, Molina, Ortí-Lucas, Parellada, Pelayo-Terán, Pérez-Zapata, Pijoan, Plana, Puig, Rius, Rodríguez-Blázquez, Sanz, Serra, Kessler, Bruffaerts, Vieta and Pérez-Solà2021; O’Connor et al., Reference O’Connor, Wetherall, Cleare, McClelland, Melson, Niedzwiedz, O’Carroll, O’Connor, Platt, Scowcroft, Watson, Zortea, Ferguson and Robb2021; Mudiriza & De Lannoy, Reference Mudiriza and De Lannoy2020), while very few studies have examined the change in these outcomes from the pre-pandemic period. Studies that have assessed this change have been conducted in HICs and found a significant decline in mental health and well-being (Bäuerle et al., Reference Bäuerle, Steinbach, Schweda, Beckord, Hetkamp, Weismüller, Kohler, Musche, Dörrie, Teufel and Skoda2020; Wanberg et al., Reference Wanberg, Csillag, Douglass, Zhou and Pollard2020; Gilleen et al., Reference Gilleen, Santaolalla, Valdearenas, Salice and Fusté2021; Nogueira et al., Reference Nogueira, Gerardo, Silva, Pinto, Barbosa, Soares, Baptista, Paquete, Cabral-Pinto, Vilar, Simões and Freitas2022). There is a lack of published empirical research on risk and protective factors influencing the change in mental health and well-being over the course of the pandemic in LMICs. A change in these outcomes may have been influenced by various demographic, socio-economic, clinical, behavioural and environmental factors, either non-modifiable or modifiable (Moreno et al., Reference Moreno, Wykes, Galderisi, Nordentoft, Crossley, Jones, Cannon, Correll, Byrne, Carr, Chen, Gorwood, Johnson, Kärkkäinen, Krystal, Lee, Lieberman, López-Jaramillo, Männikkö, Phillips, Uchida, Vieta, Vita and Arango2020; Solmi et al., Reference Solmi, Estradé, Thompson, Agorastos, Radua, Cortese, Dragioti, Leisch, Vancampfort, Thygesen, Aschauer, Schloegelhofer, Akimova, Schneeberger, Huber, Hasler, Conus, Cuénod, von Känel, Arrondo, Fusar-Poli, Gorwood, Llorca, Krebs, Scanferla, Kishimoto, Rabbani, Skonieczna-Żydecka, Brambilla, Favaro, Takamiya, Zoccante, Colizzi, Bourgin, Kamiński, Moghadasin, Seedat, Matthews, Wells, Vassilopoulou, Gadelha, Su, Kwon, Kim, Lee, Papsuev, Manková, Boscutti, Gerunda, Saccon, Righi, Monaco, Croatto, Cereda, Demurtas, Brondino, Veronese, Enrico, Politi, Ciappolino, Pfennig, Bechdolf, Meyer-Lindenberg, Kahl, Domschke, Bauer, Koutsouleris, Winter, Borgwardt, Bitter, Balazs, Czobor, Unoka, Mavridis, Tsamakis, Bozikas, Tunvirachaisakul, Maes, Rungnirundorn, Supasitthumrong, Haque, Brunoni, Costardi, Schuch, Polanczyk, Luiz, Fonseca, Aparicio, Valvassori, Nordentoft, Vendsborg, Hoffmann, Sehli, Sartorius, Heuss, Guinart, Hamilton, Kane, Rubio, Sand, Koyanagi, Solanes, Andreu-Bernabeu, Cáceres, Arango, Díaz-Caneja, Hidalgo-Mazzei, Vieta, Gonzalez-Peñas, Fortea, Parellada, Fullana, Verdolini, Fárková, Janků, Millan, Honciuc, Moniuszko-Malinowska, Łoniewski, Samochowiec, Kiszkiel, Marlicz, Sowa, Marlicz, Spies, Stubbs, Firth, Sullivan, Darcin, Aksu, Dilbaz, Noyan, Kitazawa, Kurokawa, Tazawa, Anselmi, Cracco, Machado, Estrade, De Leo, Curtis, Berk, Ward, Teasdale, Rosenbaum, Marx, Horodnic, Oprea, Alexinschi, Ifteni, Turliuc, Ciuhodaru, Bolos, Matei, Nieman, Sommer, van Os, van Amelsvoort, Sun, Guu, Jiao, Zhang, Fan, Zou, Yu, Chi, de Timary, van Winke, Ng, Pena, Arellano, Roman, Sanchez, Movina, Morgado, Brissos, Aizberg, Mosina, Krinitski, Mugisha, Sadeghi-Bahmani, Sadeghi, Hadi, Brand, Errazuriz, Crossley, Ristic, López-Jaramillo, Efthymiou, Kuttichira, Kallivayalil, Javed, Afridi, James, Seb-Akahomen, Fiedorowicz, Carvalho, Daskalakis, Yatham, Yang, Okasha, Dahdouh, Gerdle, Tiihonen, Shin, Lee, Mhalla, Gaha, Brahim, Altynbekov, Negay, Nurmagambetova, Jamei, Weiser and Correll2022). Non-modifiable factors can serve to identify individuals at increased risk of mental health-related problems during a pandemic who might require specific interventions, while modifiable factors are actionable targets for prevention/intervention strategies (Solmi et al., Reference Solmi, Estradé, Thompson, Agorastos, Radua, Cortese, Dragioti, Leisch, Vancampfort, Thygesen, Aschauer, Schloegelhofer, Akimova, Schneeberger, Huber, Hasler, Conus, Cuénod, von Känel, Arrondo, Fusar-Poli, Gorwood, Llorca, Krebs, Scanferla, Kishimoto, Rabbani, Skonieczna-Żydecka, Brambilla, Favaro, Takamiya, Zoccante, Colizzi, Bourgin, Kamiński, Moghadasin, Seedat, Matthews, Wells, Vassilopoulou, Gadelha, Su, Kwon, Kim, Lee, Papsuev, Manková, Boscutti, Gerunda, Saccon, Righi, Monaco, Croatto, Cereda, Demurtas, Brondino, Veronese, Enrico, Politi, Ciappolino, Pfennig, Bechdolf, Meyer-Lindenberg, Kahl, Domschke, Bauer, Koutsouleris, Winter, Borgwardt, Bitter, Balazs, Czobor, Unoka, Mavridis, Tsamakis, Bozikas, Tunvirachaisakul, Maes, Rungnirundorn, Supasitthumrong, Haque, Brunoni, Costardi, Schuch, Polanczyk, Luiz, Fonseca, Aparicio, Valvassori, Nordentoft, Vendsborg, Hoffmann, Sehli, Sartorius, Heuss, Guinart, Hamilton, Kane, Rubio, Sand, Koyanagi, Solanes, Andreu-Bernabeu, Cáceres, Arango, Díaz-Caneja, Hidalgo-Mazzei, Vieta, Gonzalez-Peñas, Fortea, Parellada, Fullana, Verdolini, Fárková, Janků, Millan, Honciuc, Moniuszko-Malinowska, Łoniewski, Samochowiec, Kiszkiel, Marlicz, Sowa, Marlicz, Spies, Stubbs, Firth, Sullivan, Darcin, Aksu, Dilbaz, Noyan, Kitazawa, Kurokawa, Tazawa, Anselmi, Cracco, Machado, Estrade, De Leo, Curtis, Berk, Ward, Teasdale, Rosenbaum, Marx, Horodnic, Oprea, Alexinschi, Ifteni, Turliuc, Ciuhodaru, Bolos, Matei, Nieman, Sommer, van Os, van Amelsvoort, Sun, Guu, Jiao, Zhang, Fan, Zou, Yu, Chi, de Timary, van Winke, Ng, Pena, Arellano, Roman, Sanchez, Movina, Morgado, Brissos, Aizberg, Mosina, Krinitski, Mugisha, Sadeghi-Bahmani, Sadeghi, Hadi, Brand, Errazuriz, Crossley, Ristic, López-Jaramillo, Efthymiou, Kuttichira, Kallivayalil, Javed, Afridi, James, Seb-Akahomen, Fiedorowicz, Carvalho, Daskalakis, Yatham, Yang, Okasha, Dahdouh, Gerdle, Tiihonen, Shin, Lee, Mhalla, Gaha, Brahim, Altynbekov, Negay, Nurmagambetova, Jamei, Weiser and Correll2022). Due to socio-economic inequalities and a fragile health care system in South Africa, government strategies during a pandemic need to be evidence-based and tailored to the needs of at-risk population groups.
Current study
The Collaborative Outcomes Study on Health and Functioning During Infection Times (COH-FIT) is a global multi-language online survey measuring the impact of the COVID-19 pandemic on physical and mental health and well-being (Solmi et al., Reference Solmi, Estradé, Thompson, Agorastos, Radua, Cortese, Dragioti, Leisch, Vancampfort, Thygesen, Aschauer, Schloegelhofer, Akimova, Schneeberger, Huber, Hasler, Conus, Cuénod, von Känel, Arrondo, Fusar-Poli, Gorwood, Llorca, Krebs, Scanferla, Kishimoto, Rabbani, Skonieczna-Żydecka, Brambilla, Favaro, Takamiya, Zoccante, Colizzi, Bourgin, Kamiński, Moghadasin, Seedat, Matthews, Wells, Vassilopoulou, Gadelha, Su, Kwon, Kim, Lee, Papsuev, Manková, Boscutti, Gerunda, Saccon, Righi, Monaco, Croatto, Cereda, Demurtas, Brondino, Veronese, Enrico, Politi, Ciappolino, Pfennig, Bechdolf, Meyer-Lindenberg, Kahl, Domschke, Bauer, Koutsouleris, Winter, Borgwardt, Bitter, Balazs, Czobor, Unoka, Mavridis, Tsamakis, Bozikas, Tunvirachaisakul, Maes, Rungnirundorn, Supasitthumrong, Haque, Brunoni, Costardi, Schuch, Polanczyk, Luiz, Fonseca, Aparicio, Valvassori, Nordentoft, Vendsborg, Hoffmann, Sehli, Sartorius, Heuss, Guinart, Hamilton, Kane, Rubio, Sand, Koyanagi, Solanes, Andreu-Bernabeu, Cáceres, Arango, Díaz-Caneja, Hidalgo-Mazzei, Vieta, Gonzalez-Peñas, Fortea, Parellada, Fullana, Verdolini, Fárková, Janků, Millan, Honciuc, Moniuszko-Malinowska, Łoniewski, Samochowiec, Kiszkiel, Marlicz, Sowa, Marlicz, Spies, Stubbs, Firth, Sullivan, Darcin, Aksu, Dilbaz, Noyan, Kitazawa, Kurokawa, Tazawa, Anselmi, Cracco, Machado, Estrade, De Leo, Curtis, Berk, Ward, Teasdale, Rosenbaum, Marx, Horodnic, Oprea, Alexinschi, Ifteni, Turliuc, Ciuhodaru, Bolos, Matei, Nieman, Sommer, van Os, van Amelsvoort, Sun, Guu, Jiao, Zhang, Fan, Zou, Yu, Chi, de Timary, van Winke, Ng, Pena, Arellano, Roman, Sanchez, Movina, Morgado, Brissos, Aizberg, Mosina, Krinitski, Mugisha, Sadeghi-Bahmani, Sadeghi, Hadi, Brand, Errazuriz, Crossley, Ristic, López-Jaramillo, Efthymiou, Kuttichira, Kallivayalil, Javed, Afridi, James, Seb-Akahomen, Fiedorowicz, Carvalho, Daskalakis, Yatham, Yang, Okasha, Dahdouh, Gerdle, Tiihonen, Shin, Lee, Mhalla, Gaha, Brahim, Altynbekov, Negay, Nurmagambetova, Jamei, Weiser and Correll2022; Solmi et al., Reference Solmi, Thompson, Estradé, Agorastos, Radua, Cortese, Dragioti, Leisch, Vancampfort, Thygesen, Aschauer, Schlögelhofer, Aschauer, Schneeberger, Huber, Hasler, Conus, Do Cuénod, von Känel, Arrondo, Fusar-Poli, Gorwood, Llorca, Krebs, Scanferla, Kishimoto, Rabbani, Skonieczna-Żydecka, Brambilla, Favaro, Takamiya, Zoccante, Colizzi, Bourgin, Kamiński, Moghadasin, Seedat, Matthews, Wells, Vassilopoulou, Gadelha, Su, Kwon, Kim, Lee, Papsuev, Manková, Boscutti, Gerunda, Saccon, Righi, Monaco, Croatto, Cereda, Demurtas, Brondino, Veronese, Enrico, Politi, Ciappolino, Pfennig, Bechdolf, Meyer-Lindenberg, Kahl, Domschke, Bauer, Koutsouleris, Winter, Borgwardt, Bitter, Balazs, Czobor, Unoka, Mavridis, Tsamakis, Bozikas, Tunvirachaisakul, Maes, Rungnirundorn, Supasitthumrong, Haque, Brunoni, Costardi, Schuch, Polanczyk, Luiz, Fonseca, Aparicio, Valvassori, Nordentoft, Vendsborg, Hoffmann, Sehli, Sartorius, Heuss, Guinart, Hamilton, Kane, Rubio, Sand, Koyanagi, Solanes, Andreu-Bernabeu, Cáceres, Arango, Díaz-Caneja, Hidalgo-Mazzei, Vieta, Gonzalez-Peñas, Fortea, Parellada, Fullana, Verdolini, Andrlíková, Janků, Millan, Honciuc, Moniuszko-Malinowska, Łoniewski, Samochowiec, Kiszkiel, Marlicz, Sowa, Marlicz, Spies, Stubbs, Firth, Sullivan, Darcin, Aksu, Dilbaz, Noyan, Kitazawa, Kurokawa, Tazawa, Anselmi, Cracco, Machado, Estrade, De Leo, Curtis, Berk, Ward, Teasdale, Rosenbaum, Marx, Horodnic, Oprea, Alexinschi, Ifteni, Turliuc, Ciuhodaru, Bolos, Matei, Nieman, Sommer, van Os, van Amelsvoort, Sun, Guu, Jiao, Zhang, Fan, Zou, Yu, Chi, de Timary, van Winkel, Ng, Pena, Arellano, Roman, Sanchez, Movina, Morgado, Brissos, Aizberg, Mosina, Krinitski, Mugisha, Sadeghi-Bahmani, Sheybani, Sadeghi, Hadi, Brand, Errazuriz, Crossley, Ristic, López-Jaramillo, Efthymiou, Kuttichira, Kallivayalil, Javed, Afridi, James, Seb-Akahomen, Fiedorowicz, Carvalho, Daskalakis, Yatham, Yang, Okasha, Dahdouh, Gerdle, Tiihonen, Shin, Lee, Mhalla, Gaha, Brahim, Altynbekov, Negay, Nurmagambetova, Jamei, Weiser and Correll2023). More than 185,000 surveys were initiated across the globe. In the current study, we report on survey findings from the South African adult population. We aimed to assess the change in well-being and mental health from pre- to intra-COVID-19 lockdown periods and to identify the non-modifiable and modifiable risk and protective factors associated with healthier and poorer outcomes.
Methods
Design
The COH-FIT study was a cross-sectional survey, accessible online on the COH-FIT website (www.coh-fit.com) and was available in English, isiXhosa, and 28 additional languages. The current study reports on data collected during the COVID-19 pandemic from 26 April 2020 to 22 April 2021.
Measures
Based on previous literature, several non-modifiable and modifiable risk factors that were found to influence well-being and mental health were selected (Solmi et al., Reference Solmi, Estradé, Thompson, Agorastos, Radua, Cortese, Dragioti, Leisch, Vancampfort, Thygesen, Aschauer, Schloegelhofer, Akimova, Schneeberger, Huber, Hasler, Conus, Cuénod, von Känel, Arrondo, Fusar-Poli, Gorwood, Llorca, Krebs, Scanferla, Kishimoto, Rabbani, Skonieczna-Żydecka, Brambilla, Favaro, Takamiya, Zoccante, Colizzi, Bourgin, Kamiński, Moghadasin, Seedat, Matthews, Wells, Vassilopoulou, Gadelha, Su, Kwon, Kim, Lee, Papsuev, Manková, Boscutti, Gerunda, Saccon, Righi, Monaco, Croatto, Cereda, Demurtas, Brondino, Veronese, Enrico, Politi, Ciappolino, Pfennig, Bechdolf, Meyer-Lindenberg, Kahl, Domschke, Bauer, Koutsouleris, Winter, Borgwardt, Bitter, Balazs, Czobor, Unoka, Mavridis, Tsamakis, Bozikas, Tunvirachaisakul, Maes, Rungnirundorn, Supasitthumrong, Haque, Brunoni, Costardi, Schuch, Polanczyk, Luiz, Fonseca, Aparicio, Valvassori, Nordentoft, Vendsborg, Hoffmann, Sehli, Sartorius, Heuss, Guinart, Hamilton, Kane, Rubio, Sand, Koyanagi, Solanes, Andreu-Bernabeu, Cáceres, Arango, Díaz-Caneja, Hidalgo-Mazzei, Vieta, Gonzalez-Peñas, Fortea, Parellada, Fullana, Verdolini, Fárková, Janků, Millan, Honciuc, Moniuszko-Malinowska, Łoniewski, Samochowiec, Kiszkiel, Marlicz, Sowa, Marlicz, Spies, Stubbs, Firth, Sullivan, Darcin, Aksu, Dilbaz, Noyan, Kitazawa, Kurokawa, Tazawa, Anselmi, Cracco, Machado, Estrade, De Leo, Curtis, Berk, Ward, Teasdale, Rosenbaum, Marx, Horodnic, Oprea, Alexinschi, Ifteni, Turliuc, Ciuhodaru, Bolos, Matei, Nieman, Sommer, van Os, van Amelsvoort, Sun, Guu, Jiao, Zhang, Fan, Zou, Yu, Chi, de Timary, van Winke, Ng, Pena, Arellano, Roman, Sanchez, Movina, Morgado, Brissos, Aizberg, Mosina, Krinitski, Mugisha, Sadeghi-Bahmani, Sadeghi, Hadi, Brand, Errazuriz, Crossley, Ristic, López-Jaramillo, Efthymiou, Kuttichira, Kallivayalil, Javed, Afridi, James, Seb-Akahomen, Fiedorowicz, Carvalho, Daskalakis, Yatham, Yang, Okasha, Dahdouh, Gerdle, Tiihonen, Shin, Lee, Mhalla, Gaha, Brahim, Altynbekov, Negay, Nurmagambetova, Jamei, Weiser and Correll2022; Solmi et al., Reference Solmi, Thompson, Estradé, Agorastos, Radua, Cortese, Dragioti, Leisch, Vancampfort, Thygesen, Aschauer, Schlögelhofer, Aschauer, Schneeberger, Huber, Hasler, Conus, Do Cuénod, von Känel, Arrondo, Fusar-Poli, Gorwood, Llorca, Krebs, Scanferla, Kishimoto, Rabbani, Skonieczna-Żydecka, Brambilla, Favaro, Takamiya, Zoccante, Colizzi, Bourgin, Kamiński, Moghadasin, Seedat, Matthews, Wells, Vassilopoulou, Gadelha, Su, Kwon, Kim, Lee, Papsuev, Manková, Boscutti, Gerunda, Saccon, Righi, Monaco, Croatto, Cereda, Demurtas, Brondino, Veronese, Enrico, Politi, Ciappolino, Pfennig, Bechdolf, Meyer-Lindenberg, Kahl, Domschke, Bauer, Koutsouleris, Winter, Borgwardt, Bitter, Balazs, Czobor, Unoka, Mavridis, Tsamakis, Bozikas, Tunvirachaisakul, Maes, Rungnirundorn, Supasitthumrong, Haque, Brunoni, Costardi, Schuch, Polanczyk, Luiz, Fonseca, Aparicio, Valvassori, Nordentoft, Vendsborg, Hoffmann, Sehli, Sartorius, Heuss, Guinart, Hamilton, Kane, Rubio, Sand, Koyanagi, Solanes, Andreu-Bernabeu, Cáceres, Arango, Díaz-Caneja, Hidalgo-Mazzei, Vieta, Gonzalez-Peñas, Fortea, Parellada, Fullana, Verdolini, Andrlíková, Janků, Millan, Honciuc, Moniuszko-Malinowska, Łoniewski, Samochowiec, Kiszkiel, Marlicz, Sowa, Marlicz, Spies, Stubbs, Firth, Sullivan, Darcin, Aksu, Dilbaz, Noyan, Kitazawa, Kurokawa, Tazawa, Anselmi, Cracco, Machado, Estrade, De Leo, Curtis, Berk, Ward, Teasdale, Rosenbaum, Marx, Horodnic, Oprea, Alexinschi, Ifteni, Turliuc, Ciuhodaru, Bolos, Matei, Nieman, Sommer, van Os, van Amelsvoort, Sun, Guu, Jiao, Zhang, Fan, Zou, Yu, Chi, de Timary, van Winkel, Ng, Pena, Arellano, Roman, Sanchez, Movina, Morgado, Brissos, Aizberg, Mosina, Krinitski, Mugisha, Sadeghi-Bahmani, Sheybani, Sadeghi, Hadi, Brand, Errazuriz, Crossley, Ristic, López-Jaramillo, Efthymiou, Kuttichira, Kallivayalil, Javed, Afridi, James, Seb-Akahomen, Fiedorowicz, Carvalho, Daskalakis, Yatham, Yang, Okasha, Dahdouh, Gerdle, Tiihonen, Shin, Lee, Mhalla, Gaha, Brahim, Altynbekov, Negay, Nurmagambetova, Jamei, Weiser and Correll2023).
Non-modifiable factors
Non-modifiable factors were separated into sociodemographic and environmental/clinical domains. Sociodemographic data including age, gender, race/ethnicity, education level, marital status, employment status, and SES were collected. SES was measured on a visual analogue scale (VAS) where 0 = extremely below average, 50 = average, and 100 = extremely above average. Environmental/clinical factors included urbanicity, current level of restrictions (no restrictions, recommendations to stay home only, sued or fined if leave home unless necessary, arrested if leave home unless necessary, and cannot go out under any circumstances), living with a family member, prior mental health condition, prior medical condition, family history of mental illness, and COVID-19 symptoms. Participants were also asked if they were exposed to any traumatic events during childhood and adulthood, such as physical, sexual, or verbal abuse, a major accident/injury/illness, and witnessing a traumatic event. Childhood and adulthood adversity were dichotomised as yes or no.
Modifiable factors
Modifiable factors were separated into behavioural (coping strategies) and environmental/clinical domains. Participants were asked to rate the importance of various coping strategies during the pandemic on a 3-point Likert scale (1 = not important, 2 = somewhat important, and 3 = very important). Coping strategies included direct physical contact or interactions, exercise/walking, gaming, internet use, keeping informed about the COVID-19 pandemic, meaningful hobby, media (TV, movies, radio, and music) usage, social media usage, physical intimacy/sexual activity, spending time with pets, taking prescribed medications, religion/meditation/spirituality, studying or learning something new, use of substances (tobacco, alcohol, other), work (on site or from home) and other coping strategies. Environmental/clinical factors included financial loss measured on a 0 (no loss) to 100 (extreme loss) VAS, and access to protective devices (e.g., masks, gloves, soap, and so on) measured on a 0–100 VAS.
Outcomes
Outcome measures included well-being and a composite psychopathology score (p-score). Well-being was measured with the full WHO-5 questionnaire, (Topp et al., Reference Topp, Østergaard, Søndergaard and Bech2015), but with response options converted from a six-point Likert scale to a 0 (never) to 100 (every day) VAS. The well-being scale showed excellent internal consistency (a = .86). The p-score is a validated psychopathology measure comprised of symptoms across five mental health domains including anxiety, depression, post-traumatic stress disorder (PTSD), psychosis (hallucinations and delusions), and psychophysiological problems (sleep problems, stress and concentration problems). P-score items were extracted from validated measures, including the Generalised Anxiety Disorder-7 scale (Spitzer et al., Reference Spitzer, Kroenke, Williams and Lowe2006), Patient Health Questionnaire-9 (Kroenke et al., Reference Kroenke, Spitzer and Williams2001), PTSD checklist for DSM-5 (Blevins et al., Reference Blevins, Weathers, Davis, Witte and Domino2015), Prodromal Questionnaire-16 (Ising et al., Reference Ising, Veling, Loewy, Rietveld, Rietdijk, Dragt, Klaassen, Nieman, Wunderink, Linszen and van der Gaag2012), and the World Health Organization-5 well-being index (Heitor Dos Santos et al., Reference Heitor dos Santos, Moreira, Carreiras, Cooper, Smeed, Reis and Pereira Miguel2018). Globally, the COVID-19 pandemic resulted in a wide range of mental health problems (Dragioti et al., Reference Dragioti, Li, Tsitsas, Lee, Choi, Kim, Choi, Tsamakis, Estradé, Agorastos, Vancampfort, Tsiptsios, Thompson, Mosina, Vakadaris, Fusar‐Poli, Carvalho, Correll, Han, Park, Il Shin and Solmi2022), suggesting the need for a multidimensional measure of mental health (Solmi et al., Reference Solmi, Thompson, Estradé, Agorastos, Radua, Cortese, Dragioti, Leisch, Vancampfort, Thygesen, Aschauer, Schlögelhofer, Aschauer, Schneeberger, Huber, Hasler, Conus, Do Cuénod, von Känel, Arrondo, Fusar-Poli, Gorwood, Llorca, Krebs, Scanferla, Kishimoto, Rabbani, Skonieczna-Żydecka, Brambilla, Favaro, Takamiya, Zoccante, Colizzi, Bourgin, Kamiński, Moghadasin, Seedat, Matthews, Wells, Vassilopoulou, Gadelha, Su, Kwon, Kim, Lee, Papsuev, Manková, Boscutti, Gerunda, Saccon, Righi, Monaco, Croatto, Cereda, Demurtas, Brondino, Veronese, Enrico, Politi, Ciappolino, Pfennig, Bechdolf, Meyer-Lindenberg, Kahl, Domschke, Bauer, Koutsouleris, Winter, Borgwardt, Bitter, Balazs, Czobor, Unoka, Mavridis, Tsamakis, Bozikas, Tunvirachaisakul, Maes, Rungnirundorn, Supasitthumrong, Haque, Brunoni, Costardi, Schuch, Polanczyk, Luiz, Fonseca, Aparicio, Valvassori, Nordentoft, Vendsborg, Hoffmann, Sehli, Sartorius, Heuss, Guinart, Hamilton, Kane, Rubio, Sand, Koyanagi, Solanes, Andreu-Bernabeu, Cáceres, Arango, Díaz-Caneja, Hidalgo-Mazzei, Vieta, Gonzalez-Peñas, Fortea, Parellada, Fullana, Verdolini, Andrlíková, Janků, Millan, Honciuc, Moniuszko-Malinowska, Łoniewski, Samochowiec, Kiszkiel, Marlicz, Sowa, Marlicz, Spies, Stubbs, Firth, Sullivan, Darcin, Aksu, Dilbaz, Noyan, Kitazawa, Kurokawa, Tazawa, Anselmi, Cracco, Machado, Estrade, De Leo, Curtis, Berk, Ward, Teasdale, Rosenbaum, Marx, Horodnic, Oprea, Alexinschi, Ifteni, Turliuc, Ciuhodaru, Bolos, Matei, Nieman, Sommer, van Os, van Amelsvoort, Sun, Guu, Jiao, Zhang, Fan, Zou, Yu, Chi, de Timary, van Winkel, Ng, Pena, Arellano, Roman, Sanchez, Movina, Morgado, Brissos, Aizberg, Mosina, Krinitski, Mugisha, Sadeghi-Bahmani, Sheybani, Sadeghi, Hadi, Brand, Errazuriz, Crossley, Ristic, López-Jaramillo, Efthymiou, Kuttichira, Kallivayalil, Javed, Afridi, James, Seb-Akahomen, Fiedorowicz, Carvalho, Daskalakis, Yatham, Yang, Okasha, Dahdouh, Gerdle, Tiihonen, Shin, Lee, Mhalla, Gaha, Brahim, Altynbekov, Negay, Nurmagambetova, Jamei, Weiser and Correll2023). A single all-encompassing measure of mental health, as opposed to individual full-length measures, was intended to reduce the time taken to complete an online survey and increase completion rates (Solmi et al., Reference Solmi, Thompson, Estradé, Agorastos, Radua, Cortese, Dragioti, Leisch, Vancampfort, Thygesen, Aschauer, Schlögelhofer, Aschauer, Schneeberger, Huber, Hasler, Conus, Do Cuénod, von Känel, Arrondo, Fusar-Poli, Gorwood, Llorca, Krebs, Scanferla, Kishimoto, Rabbani, Skonieczna-Żydecka, Brambilla, Favaro, Takamiya, Zoccante, Colizzi, Bourgin, Kamiński, Moghadasin, Seedat, Matthews, Wells, Vassilopoulou, Gadelha, Su, Kwon, Kim, Lee, Papsuev, Manková, Boscutti, Gerunda, Saccon, Righi, Monaco, Croatto, Cereda, Demurtas, Brondino, Veronese, Enrico, Politi, Ciappolino, Pfennig, Bechdolf, Meyer-Lindenberg, Kahl, Domschke, Bauer, Koutsouleris, Winter, Borgwardt, Bitter, Balazs, Czobor, Unoka, Mavridis, Tsamakis, Bozikas, Tunvirachaisakul, Maes, Rungnirundorn, Supasitthumrong, Haque, Brunoni, Costardi, Schuch, Polanczyk, Luiz, Fonseca, Aparicio, Valvassori, Nordentoft, Vendsborg, Hoffmann, Sehli, Sartorius, Heuss, Guinart, Hamilton, Kane, Rubio, Sand, Koyanagi, Solanes, Andreu-Bernabeu, Cáceres, Arango, Díaz-Caneja, Hidalgo-Mazzei, Vieta, Gonzalez-Peñas, Fortea, Parellada, Fullana, Verdolini, Andrlíková, Janků, Millan, Honciuc, Moniuszko-Malinowska, Łoniewski, Samochowiec, Kiszkiel, Marlicz, Sowa, Marlicz, Spies, Stubbs, Firth, Sullivan, Darcin, Aksu, Dilbaz, Noyan, Kitazawa, Kurokawa, Tazawa, Anselmi, Cracco, Machado, Estrade, De Leo, Curtis, Berk, Ward, Teasdale, Rosenbaum, Marx, Horodnic, Oprea, Alexinschi, Ifteni, Turliuc, Ciuhodaru, Bolos, Matei, Nieman, Sommer, van Os, van Amelsvoort, Sun, Guu, Jiao, Zhang, Fan, Zou, Yu, Chi, de Timary, van Winkel, Ng, Pena, Arellano, Roman, Sanchez, Movina, Morgado, Brissos, Aizberg, Mosina, Krinitski, Mugisha, Sadeghi-Bahmani, Sheybani, Sadeghi, Hadi, Brand, Errazuriz, Crossley, Ristic, López-Jaramillo, Efthymiou, Kuttichira, Kallivayalil, Javed, Afridi, James, Seb-Akahomen, Fiedorowicz, Carvalho, Daskalakis, Yatham, Yang, Okasha, Dahdouh, Gerdle, Tiihonen, Shin, Lee, Mhalla, Gaha, Brahim, Altynbekov, Negay, Nurmagambetova, Jamei, Weiser and Correll2023). As part of the COH-FIT study, we were able to validate the p-score used in our analyses (Solmi et al., Reference Solmi, Thompson, Estradé, Agorastos, Radua, Cortese, Dragioti, Leisch, Vancampfort, Thygesen, Aschauer, Schlögelhofer, Aschauer, Schneeberger, Huber, Hasler, Conus, Do Cuénod, von Känel, Arrondo, Fusar-Poli, Gorwood, Llorca, Krebs, Scanferla, Kishimoto, Rabbani, Skonieczna-Żydecka, Brambilla, Favaro, Takamiya, Zoccante, Colizzi, Bourgin, Kamiński, Moghadasin, Seedat, Matthews, Wells, Vassilopoulou, Gadelha, Su, Kwon, Kim, Lee, Papsuev, Manková, Boscutti, Gerunda, Saccon, Righi, Monaco, Croatto, Cereda, Demurtas, Brondino, Veronese, Enrico, Politi, Ciappolino, Pfennig, Bechdolf, Meyer-Lindenberg, Kahl, Domschke, Bauer, Koutsouleris, Winter, Borgwardt, Bitter, Balazs, Czobor, Unoka, Mavridis, Tsamakis, Bozikas, Tunvirachaisakul, Maes, Rungnirundorn, Supasitthumrong, Haque, Brunoni, Costardi, Schuch, Polanczyk, Luiz, Fonseca, Aparicio, Valvassori, Nordentoft, Vendsborg, Hoffmann, Sehli, Sartorius, Heuss, Guinart, Hamilton, Kane, Rubio, Sand, Koyanagi, Solanes, Andreu-Bernabeu, Cáceres, Arango, Díaz-Caneja, Hidalgo-Mazzei, Vieta, Gonzalez-Peñas, Fortea, Parellada, Fullana, Verdolini, Andrlíková, Janků, Millan, Honciuc, Moniuszko-Malinowska, Łoniewski, Samochowiec, Kiszkiel, Marlicz, Sowa, Marlicz, Spies, Stubbs, Firth, Sullivan, Darcin, Aksu, Dilbaz, Noyan, Kitazawa, Kurokawa, Tazawa, Anselmi, Cracco, Machado, Estrade, De Leo, Curtis, Berk, Ward, Teasdale, Rosenbaum, Marx, Horodnic, Oprea, Alexinschi, Ifteni, Turliuc, Ciuhodaru, Bolos, Matei, Nieman, Sommer, van Os, van Amelsvoort, Sun, Guu, Jiao, Zhang, Fan, Zou, Yu, Chi, de Timary, van Winkel, Ng, Pena, Arellano, Roman, Sanchez, Movina, Morgado, Brissos, Aizberg, Mosina, Krinitski, Mugisha, Sadeghi-Bahmani, Sheybani, Sadeghi, Hadi, Brand, Errazuriz, Crossley, Ristic, López-Jaramillo, Efthymiou, Kuttichira, Kallivayalil, Javed, Afridi, James, Seb-Akahomen, Fiedorowicz, Carvalho, Daskalakis, Yatham, Yang, Okasha, Dahdouh, Gerdle, Tiihonen, Shin, Lee, Mhalla, Gaha, Brahim, Altynbekov, Negay, Nurmagambetova, Jamei, Weiser and Correll2023). In a validation study by Solmi and colleagues (2023), exploratory factor analysis and confirmatory factor analysis was used to extract a single “p-score” factor from multiple COH-FIT survey items across the five domains. Fig. 1 illustrates the factor structure of the p-score and Table 1 indicates the domain item descriptions and scoring. Domain items were measured on a 0–100 VAS. The mean item score for each of the five domains was computed and then averaged to create an overall p-score (0–100). The p-score was found to have good concurrent validity and high internal reliability (ω = 0.95). Study outcomes were assessed for the retrospectively recalled period of two weeks of normal life prior to the local onset of the COVID-19 pandemic and for the last two weeks prior to taking the survey during the COVID-19 pandemic.
Figure 1. Factor structure of the p-score.
Table 1. P-score items, descriptions and scoring
Participants and data collection
Non-probability convenience and snowball sampling procedures were used. There were three main methods of survey dissemination and participant recruitment in South Africa: (1) the survey was released to staff and students via Stellenbosch University’s SunSurvey platform, (2) the survey was released on social media platforms, and (3) the survey was administered in communities by a research nurse. The sample consisted of 1866 adults aged 18 years and older (M = 44.26 ± 17.36 years).
Statistical analyses
Only respondents who provided responses to all items of two item outcomes, or to all items but one of three or more-item outcomes, were included in the main analyses. Multiple imputation by chained equations (using MICE algorithm with mice (Van Buuren & Groothuis-Oudshoorn, Reference Van Buuren and Groothuis-Oudshoorn2011) package in R) was used to impute the remaining missing items. Boxplots were used to identify outliers. Identified outliers were winsorized to the next highest non-outlying value.
IBM SPSS Statistics (Version 28, IBM Corp, 2021) was used to analyse the data. Sum scores for well-being and the p-score before and during the pandemic were computed. Paired samples t-tests were performed to compare outcomes before and during the pandemic. A significance level of p < 0.05 (2-sided) was used. Cohen’s d was used to determine small (≤0.2), medium (≤0.5), and large (≤0.8) effect sizes (Sawilowsky, Reference Sawilowsky2009). Difference scores were computed, indicating the change from before to during the pandemic, and used as outcomes for further analyses.
Descriptive statistics for all variables were computed. One-way ANOVA and independent samples t-tests were performed to explore significant relationships between categorical variables and continuous outcomes. Welch’s test was used for groups with unequal variances. Given the large number of comparisons, Bonferroni corrections were applied to reduce the likelihood of Type 1 error. A corrected p-value of 0.002 (0.05/31 tests per outcome) was used to determine statistical significance. Pearson correlation analyses were conducted to assess the relationships between continuous variables and outcomes. Statistically significant factors were included in multivariate linear regression. Dummy variables were created for variables with multilevel categories. Two separate regression models were tested, one for each outcome. A significant p-value of < 0.05 was used.
There was no issue of multicollinearity between independent variables in each of the models, indicated by variance inflation factor (VIF) scores <10 and tolerance values >0.1. Given the large sample size of the study, we assume that violation of the normal distribution did not negatively impact regression estimates (Schmidt and Finan, Reference Schmidt and Finan2018). Homoscedasticity was established by visually inspecting residual plots.
Ethical considerations
The study received ethics approval from the Health Research Ethics Committee (HREC) at Stellenbosch University (ref: N20/04/007_COVID-19). Participants were asked to provide informed consent prior to participating in the survey.
Results
Sample characteristics
Sample characteristics are presented in Table 2. The majority of participants were in the 50-to-69-year age group (35.6%), followed by the 30-to-49-year age group (30.4%). Most participants were female (78.9%), identified as White (75.6%), had a college/university degree (66.1%), and were married or co-living with a partner (43.8%). Over half of the sample was unemployed (51.9%) and the overall SES was slightly above average (M = 60.88; SD = 19.26).
Table 2. Sample characteristics
Note. Restrictions level 1= recommendations to stay home only, level 2 = sued or fined if leaving home unless necessary, level 3 = arrested if leaving home unless necessary, and level 4 = cannot go out under any circumstances.
a Mean and standard deviation.
b Missing data.
c Symptoms include fever, cough, shortness of breath, sore throat, fatigue.
The majority of participants lived in a large city/town (53.4%). The most commonly reported restrictive measure was “recommendations to stay home only” (68.2%), and the least common was “cannot go out under any circumstances” (.4%). Nearly three quarters of the sample lived with a family member (72%). Half of the sample had a prior medical condition (50.3%) and were exposed to adversity during childhood (50.6%) and adulthood (49.6%). Nearly a third of the sample had a prior mental health condition (31.5%) and 42% had a family history of mental illness. Most participants did not report having any COVID-19 symptoms at the time the survey was taken (88.7%).
Change in outcomes from before to during the COVID-19 pandemic
Change in well-being and the p-score from before to during the COVID-19 pandemic are presented in Table 3. Results of the paired samples t-tests revealed a significant decrease in well-being (t(1866) = −34.93, p < .001) with a large effect size (Cohen’s d = .81) and a significant increase in the p-score (t(1866) = 30.48, p < .001) with a medium effect size (Cohen’s d = .71).
Table 3. Paired samples t-test of the change in well-being and p-score from before to during the COVID-19 pandemic
*p<0.05, **p < 0.01, ***p < 0.001.
Non-modifiable sociodemographic factors
Sociodemographic group differences in well-being and p-score are presented in Table 4. Participants identifying as Hispanic (n = 1), non-binary (n = 10), and transgender/intersex (n = 2) were excluded from analyses due to their small n. Although the inclusion of diverse groups provides greater generalisability, small sample sizes limit the ability to draw accurate conclusions. Independent samples t-tests indicated significant differences in the p-score between males and females (t(1835) = −3.6, p = .001). Analysis of variance indicated significant differences in the p-score between racial/ethnic groups (F(5, 137.0) = 8.5, p < .001). Additionally, significant differences in both well-being (F(2, 279.7) = 8.8, p < .001) and the p-score (F(2, 293.3) = 9.4, p < .001) were found across education level. Moreover, participants with a higher SES had significantly higher well-being scores (r = 0.099, p < .001).
Table 4. Change in well-being and p-score across sociodemographic groups
dPearson correlation coefficient.
Bonferroni corrected p = 0.002.
Non-modifiable environmental/clinical factors
Group differences in well-being and p-score across non-modifiable environmental/clinical factors are presented in Table 5. Analysis of variance revealed significant differences in well-being scores between COVID-19 restriction levels (F(3, 210.9) = 5.1, p = .002). Participants who reported restriction level 4 (n = 7) were excluded from analyses due to the small n. Independent samples t-tests revealed significant differences in the p-score for family history of mental illness (t(1861) = −3.4, p < .001), prior mental health condition (t(1857) = −4.8, p < .001), and exposure to childhood (t(1857.8) = −3.8, p < .001) and adulthood (t(1840.1) = −5.1, p < .001) adversity. Similarly, t-test results indicated significant differences in well-being for prior mental health condition (t(1023.9) = 4.0, p < .001) and adulthood adversity (t(1843.7) = 5.0, p < .001).
Table 5. Change in well-being and p-score across non-modifiable environmental/clinical factors
Bonferroni corrected p = 0.002.
Modifiable behavioural factors
Table 6 presents the level of perceived importance of various coping strategies used during the pandemic. The most important coping strategy reported was internet usage (76.6%). Over 50% of participants reported that direct personal contact, exercise, hobbies, spending time with pets, media usage, and work were very important coping strategies during the pandemic. Approximately 40% of participants reported that keeping informed about the pandemic and studying/learning something new were somewhat important, and over 30% reported that physical contact, exercise, hobbies, media usage, social media usage, physical intimacy, work, and other coping strategies were somewhat important during the pandemic. The least important coping strategy was gaming (71.5%), followed by substance use (62.2%).
Table 6. Change in well-being and p-score across importance levels of coping strategies
Bonferroni corrected p = 0.002.
Analysis of variance revealed significant differences in well-being between importance levels of direct personal contact/interactions (F(2, 631.9) = 20.1, p < .001), spending time with pets (F(2, 924.8) = 8.6, p < .001), using prescribed medication (F(2, 1863) = 8.9, p < .001), studying/learning something new (F(2, 1863) = 14.6, p < .001), and using substances (F(2, 677.1) = 42.6, p < .001). Similarly, significant differences in the p-score were found between importance levels of direct personal contact/interactions (F(2, 657.1) = 22.2, p < .001), spending time with pets (F(2, 908.7) = 10.3, p < .001), studying/learning something new (F(2, 1863) = 9.5, p < .001), and using substances (F(2, 649.2) = 37.9, p < .001).
Modifiable environmental/clinical factors
Pearson correlation analyses were performed to assess whether change in well-being and p-score were correlated with financial loss during COVID-19 and access to protective devices. Financial loss (M = 35.0, SD = 31.9) was significantly negatively correlated with well-being (r = −.179) and positively correlated with the p-score (r = .166). There was no significant correlation between access to protective devices (M = 89.3; SD = 20.5) and either outcome variable.
Multivariate regression models
The association between non-modifiable and modifiable factors and change in well-being is presented in Table 7. Factors significantly associated with greater worsening in well-being included having a prior mental health condition (p = .027), adulthood adversity (p = .023), endorsing direct personal contact/interactions as a very important coping strategy (p < .001), endorsing substance use as a very important coping strategy (p < .001), and greater financial loss since the COVID-19 pandemic (p < .001). Factors significantly associated with a smaller decrease in well-being included having a PhD (p = .006), higher SES (p = .019), and endorsing studying/learning something new as a somewhat important (p = .004) and very important coping strategy (p < .001). The model explained 11% of the variance in the change in well-being (adjusted R 2 = .11).
Table 7. Non-modifiable and modifiable factors associated with a change in well-being
*p< 0.05, **p < 0.01, ***p < 0.001
Note: Negative change values indicate worsening of well-being.
The association between non-modifiable and modifiable factors and change in the p-score is presented in Table 8. Factors significantly associated with greater worsening in the p-score included being female (p = .04), adulthood adversity (p = .02), endorsing direct personal contact as a very important coping strategy (p < .001), endorsing substance use as a very important coping strategy (p < .001), and greater financial loss since the COVID-19 pandemic (p < .001). Factors significantly associated with a smaller increase of the p-score included being of African descent (p < .001), having a PhD (p = .001), and endorsing studying/learning something new as a very important (p = .01) and somewhat important (p = .03) coping strategy. The model explained 11% of the variance in the change in p-score (adjusted R 2 = .11).
Table 8. Non-modifiable and modifiable factors associated with a change in p-score
*p< 0.05, **p < 0.01, ***p < 0.001
Note: Positive change values indicate greater psychopathology.
Discussion
The current study assessed the impact of the COVID-19 pandemic on well-being and psychopathology, measured with the p-score (consisting of anxiety, depression, PTSD, psychosis, sleep problems, stress, and concentration problems), in a South African adult population. To our knowledge, this is the first South African study to assess the risk and protective factors associated with the change in well-being and mental health from before to during the COVID-19 pandemic. The current study found that both outcomes significantly deteriorated during the COVID-19 pandemic compared to the last two weeks before the outbreak. There was a significant decrease in well-being, with a large effect size, and an increase in the p-score, with a medium effect size. It is important to note that retrospective evaluation introduces recall bias and participants may have perceived their mental health and well-being more positively before than during the pandemic. Additionally, an expectancy effect may have occurred, such that individuals’ expectations about the impact of the pandemic might have resulted in more negative responses (Cabeleira et al., Reference Cabeleira, Steinman, Burgess, Bucks, MacLeod, Melo and Teachman2014). Nonetheless, change in mental health and well-being found in the current study are supported by COVID-19 research conducted across various other populations (Bäuerle et al., Reference Bäuerle, Steinbach, Schweda, Beckord, Hetkamp, Weismüller, Kohler, Musche, Dörrie, Teufel and Skoda2020; Wanberg et al., Reference Wanberg, Csillag, Douglass, Zhou and Pollard2020; Gilleen et al., Reference Gilleen, Santaolalla, Valdearenas, Salice and Fusté2021).
Non-modifiable risk and protective factors
In the current study, female participants had a significantly greater increase in psychopathology than male participants. This finding aligns with a previous study conducted in a LMIC which found that depression and anxiety increased especially among women from before to during the COVID-19 pandemic (Hamadani et al., Reference Hamadani, Hasan, Baldi, Hossain, Shiraji, Bhuiyan, Mehrin, Fisher, Tofail, Tipu, Grantham-McGregor, Biggs, Braat and Pasricha2020). Similarly, research in high-income countries has revealed higher levels of depression, anxiety, and stress among women during the pandemic compared to men (Gloster et al., Reference Gloster, Lamnisos, Lubenko, Presti, Squatrito, Constantinou, Nicolaou, Papacostas, Aydın, Chong, Chien, Cheng, Ruiz, Garcia-Martin, Obando-Posada, Segura-Vargas, Vasiliou, McHugh, Höfer, Baban, Dias Neto, Nunes da Silva, Monestès, Alvarez-Galvez, Paez-Blarrina, Montesinos, Valdivia-Salas, Ori, Kleszcz, Lappalainen, Ivanović, Gosar, Dionne, Merwin, Kassianos, Karekla and Francis2020; Liu et al., Reference Liu, Yang, Zhang, Xu, Cai, Ma, Wang, Cai, Du, Li, Kang, Zheng, Liu and Zhang2021; O’Connor et al., Reference O’Connor, Wetherall, Cleare, McClelland, Melson, Niedzwiedz, O’Carroll, O’Connor, Platt, Scowcroft, Watson, Zortea, Ferguson and Robb2021). This might be explained by an increase in domestic and childcare responsibilities falling on women during lockdown (Pillay & Barnes, Reference Pillay and Barnes2020; Orkin et al., Reference Orkin, Roberts, Bohler-Muller and Alexander2020; Fancourt et al., Reference Fancourt, Steptoe and Bu2021) as well as an increase in intimate partner violence (Hamadani et al., Reference Hamadani, Hasan, Baldi, Hossain, Shiraji, Bhuiyan, Mehrin, Fisher, Tofail, Tipu, Grantham-McGregor, Biggs, Braat and Pasricha2020). Women also tend to ruminate more than men (Johnson & Whisman, Reference Johnson and Whisman2013), which might have led to greater psychological distress, particularly in the context of COVID-19.
Consistent with our expectations, higher SES was significantly associated with a smaller decrease in well-being. Similarly, studies conducted in Japan, the US, and the UK found that poorer health, lower well-being, and psychological distress were more common among lower SES groups (Lee & Singh, Reference Lee and Singh2021; O’Connor et al., Reference O’Connor, Wetherall, Cleare, McClelland, Melson, Niedzwiedz, O’Carroll, O’Connor, Platt, Scowcroft, Watson, Zortea, Ferguson and Robb2021; Nagasu et al., Reference Nagasu, Muto, Yamamoto and Hashimoto2021). Individuals with higher SES were likely able to work from home, have spacious living arrangements, stable internet access, and food security (Reeves & Rothwell, Reference Reeves and Rothwell2020), contributing to better well-being. Other studies, however, had yielded contrasting findings (Wanberg et al., Reference Wanberg, Csillag, Douglass, Zhou and Pollard2020; Banna et al., Reference Banna, Al Sayeed, Kundu, Christopher, Hasan, Begum, Kormoker, Dola, Hassan, Chowdhury and Khan2022). For example, a study conducted in Bangladesh found that having a higher family income was significantly associated with higher levels of stress during the pandemic (Banna et al., Reference Banna, Al Sayeed, Kundu, Christopher, Hasan, Begum, Kormoker, Dola, Hassan, Chowdhury and Khan2022).
A smaller increase in psychopathology was found among participants of Black African descent. A previous South African large-scale survey yielded similar findings, revealing less psychological distress among Black South Africans during the pandemic compared to their White counterparts (Orkin et al., Reference Orkin, Roberts, Bohler-Muller and Alexander2020). Additionally, a study conducted in the UK found that participants of ethnic minorities did not have significantly worsened mental health outcomes during COVID-19 (Lewis et al., Reference Lewis, Lewis, Roberts, Richards, Evison, Pearce, Lloyd, Meudell, Edwards, Robinson, Poole, John, Bisson and Jones2022). This finding is contrary to expectations as Black South Africans predominantly live in disadvantaged communities characterised by financial constraints, inadequate health care, food insecurity, and other inequalities, which were exacerbated during the lockdown period (Pillay and Barnes, Reference Pillay and Barnes2020). However, this population group might exhibit greater resilience given their ongoing exposure to adversity, as documented in previous research (Theron, Reference Theron2016; Jefferis & Theron, Reference Jefferis and Theron2018; Van Breda & Theron, Reference Van Breda and Theron2018).
In the current study, having a PhD was associated with higher levels of well-being and lower levels of psychopathology. Some studies have found that higher education was associated with lower levels of depression, stress negative affect, and higher levels of well-being (Gloster et al., Reference Gloster, Lamnisos, Lubenko, Presti, Squatrito, Constantinou, Nicolaou, Papacostas, Aydın, Chong, Chien, Cheng, Ruiz, Garcia-Martin, Obando-Posada, Segura-Vargas, Vasiliou, McHugh, Höfer, Baban, Dias Neto, Nunes da Silva, Monestès, Alvarez-Galvez, Paez-Blarrina, Montesinos, Valdivia-Salas, Ori, Kleszcz, Lappalainen, Ivanović, Gosar, Dionne, Merwin, Kassianos, Karekla and Francis2020; Wang et al., Reference Wang, Pan, Wan, Tan, Xu, Ho and Ho2020), while other studies have reported an increased risk of mental health problems (Wang et al., Reference Wang, Pan, Wan, Tan, Xu, Ho and Ho2020; Kar et al., Reference Kar, Kar and Kar2021; Banna et al., Reference Banna, Al Sayeed, Kundu, Christopher, Hasan, Begum, Kormoker, Dola, Hassan, Chowdhury and Khan2022).
Consistent with previous research (O’Connor et al., Reference O’Connor, Wetherall, Cleare, McClelland, Melson, Niedzwiedz, O’Carroll, O’Connor, Platt, Scowcroft, Watson, Zortea, Ferguson and Robb2021; Lewis et al., Reference Lewis, Lewis, Roberts, Richards, Evison, Pearce, Lloyd, Meudell, Edwards, Robinson, Poole, John, Bisson and Jones2022), the current study found that participants with a pre-existing mental health condition had significantly poorer well-being compared to those without. This might be due to limited support from clinical services and a lack of psychosocial programmes for those with mental health issues during the pandemic. Health care services in South Africa paid little attention to the mental well-being of those providing or seeking healthcare, as government funding was redirected to the testing and treatment of COVID-19 cases (Rwafa-Ponela et al., Reference Rwafa-Ponela, Price, Nyatela, Nqakala, Mosam, Erzse, Lalla-Edward, Hove, Kahn, Tollman and Hofman2022). Additionally, failure to access services due to fears of contracting the virus might have resulted in exacerbation of symptoms. This suggests again that individuals with a pre-existing mental health condition are at an especially high risk of worsened well-being during times of crises and require appropriate monitoring and support measures (Alonso et al., Reference Alonso, Vilagut, Mortier, Ferrer, Alayo, Aragón-Peña, Aragonès, Campos, Cura-González, Emparanza, Espuga, Forjaz, González-Pinto, Haro, López-Fresneña, Salázar, Molina, Ortí-Lucas, Parellada, Pelayo-Terán, Pérez-Zapata, Pijoan, Plana, Puig, Rius, Rodríguez-Blázquez, Sanz, Serra, Kessler, Bruffaerts, Vieta and Pérez-Solà2021).
Participants exposed to adulthood adversity had a significantly greater increase in psychopathology and decrease in well-being. While trauma exposure is a well-established risk factor of mental health problems (Atwoli et al., Reference Atwoli, Stein, Williams, Mclaughlin, Petukhova, Kessler and Koenen2013; Van Zyl et al., Reference Van Zyl, Nel, Du Toit and Joubert2017), this risk factor may become accentuated in the context of COVID-19. In prior studies, threats to personal safety during lockdown were associated with increased levels of depression and anxiety (Wright et al., Reference Wright, Steptoe and Fancourt2020; Keynejad, Reference Keynejad2023). South Africa has seen a drastic increase in GBV since the start of the pandemic (Nguse & Wassenaar, Reference Nguse and Wassenaar2021). Women were confined to their homes with their abusive partners for a prolonged period, increasing the level of physical, sexual, and emotional abuse (Leburu-Masigo & Kgadima, Reference Leburu-Masigo and Kgadima2020). Additionally, lockdown restrictions and the re-prioritisation of healthcare services to COVID-19 infections disrupted accessibility to services for victims of GBV (African Union Commission - Women, Gender and Development Directorate (AUC-WGDD), United Nations Entity for Gender Equality and the Empowerment of Women (UN Women), Office of the United Nations High Commissioner for Human Rights (OHCHR) and United Nations Population Fund (UNFPA), 2020). Future disaster management plans should include risk mitigation strategies for trauma-exposed individuals to lessen the physical and mental health effects of pandemics.
Modifiable risk and protective factors
The current study found that certain coping strategies endorsed during the pandemic were associated with poorer mental health and well-being. Although significant associations were found, the cross-sectional design limits our ability to establish the direction of these associations and further longitudinal research is needed to determine cause and effect. Firstly, direct personal contact/interactions was significantly associated with a greater decrease in well-being and increase in psychopathology. This finding suggests that those with a higher p-score and lower well-being might have sought direct contact with social support systems to better cope during the pandemic. Alternatively, the reverse may also be true, in that individuals depending on and valuing direct social contact were more deprived of this option during the pandemic. Results from a qualitative study conducted among healthcare workers in South Africa found that being socially isolated was emotionally difficult for those who depended on family and social relationships for support (Rwafa-Ponela et al., Reference Rwafa-Ponela, Price, Nyatela, Nqakala, Mosam, Erzse, Lalla-Edward, Hove, Kahn, Tollman and Hofman2022).
Secondly, substance use endorsed as a very important coping strategy was significantly associated with a greater decrease in well-being and increase in psychopathology. Likewise, a previous systematic review revealed that substance use during lockdown was linked to increased depression, anxiety, and loneliness (Roberts et al., Reference Roberts, Rogers, Mason, Siriwardena, Hogue, Whitley and Law2021). The consumption of alcohol or use of substances were also found to be common coping strategies used during the COVID-19 pandemic in other countries (Budimir et al., Reference Budimir, Probst and Pieh2021; Ogueji et al., Reference Ogueji, Okoloba and Demoko Ceccaldi2022). Previous research suggests that those with greater mental health difficulties use substances to self-medicate psychological distress (Thornton et al., Reference Thornton, Baker, Lewin, Kay-Lambkin, Kavanagh, Richmond, Kelly and Johnson2012). Alternatively, individuals may have used substances to avoid the negative effects of the pandemic. A previous study found that those who engaged in avoidance coping behaviours during the pandemic were more likely to experience greater distress and poorer well-being (Dawson & Golijani-Moghaddam, Reference Dawson and Golijani-Moghaddam2020). Furthermore, a ban on alcohol sales was implemented in South Africa at various stages in 2020, which led to a significant decrease in alcohol-related deaths (Moultrie et al., Reference Moultrie, Dorrington, Laubscher, Groenewald, Parry, Matzopoulos and Bradshaw2021). Future research on the impact of banning alcohol sales on mental health and well-being is warranted.
Studying/learning something new was the only coping strategy found to be significantly associated with higher levels of well-being and lower levels of psychopathology. In prior COVID-19 research, keeping busy with activities during lockdown was reported to be a frequent strategy used to cope (Fuller & Huseth-Zosel, Reference Fuller and Huseth-Zosel2021; Kar et al., Reference Kar, Kar and Kar2021). In a qualitative study conducted in the UK, participants reported that engaging with university work was helpful to remain occupied during lockdown (Ogueji et al., Reference Ogueji, Okoloba and Demoko Ceccaldi2022). Engaging in work or studies may be a proactive approach to help individuals accomplish goals, have a sense of purpose and agency, or maintain normalcy in light of the restrictions imposed. Future interventions should focus on promoting positive coping during infection times. Positive coping strategies were found to be protective against anxiety and depression (Fullana et al., Reference Fullana, Hidalgo-Mazzei, Vieta and Radua2020) and were linked to post-traumatic growth during the COVID-19 pandemic (Willey et al., Reference Willey, Mimmack, Gagliardi, Dossett, Wang, Udeogu, Donovan, Gatchel, Quiroz, Amariglio, Liu, Hyun, Eltohamy, Rentz, Sperling, Marshall and Vannini2022).
Financial loss was significantly associated with a greater decrease in well-being and increase in psychopathology, which is consistent with previous research in high- and low-income countries (Chatterji et al., Reference Chatterji, McDougal, Johns, Ghule, Rao and Raj2021; Mojtahedi et al., Reference Mojtahedi, Dagnall, Denovan, Clough, Hull, Canning, Lilley and Papageorgiou2021; Oyenubi & Kollamparambil, Reference Oyenubi and Kollamparambil2022). A study by Posel and colleagues (2021) found that South African adults who kept their jobs during lockdown experienced significantly lower levels of depression and saw improvements in mental health over time, compared to those who became unemployed. Although unemployment is generally linked to poorer mental health (Paul and Moser, Reference Paul and Moser2009; Picchio and Ubaldi, Reference Picchio and Ubaldi2022), this impact may have worsened during COVID-19 due to added stressors such as illness, loss of loved ones, and restrictions on movement. Finally, studies have found that government social grants were associated with better mental health outcomes during the pandemic (Chatterji et al., Reference Chatterji, McDougal, Johns, Ghule, Rao and Raj2021; Posel et al., Reference Posel, Oyenubi, Kollamparambil and Picone2021). This underscores the importance of providing government aid in tandem with mental health interventions to prevent worsening of mental health and well-being during a pandemic.
Limitations
The current study is not without limitations. Due to the cross-sectional design of the study, conclusions about causality cannot be made. As the majority of the sample were female, White, and university educated, findings cannot be generalised to the broader South African population. Due to the electronic format of the survey, those without access to the internet, such as lower-income groups, were unlikely to be represented in the sample. The survey was administered to various low-income communities by a research nurse, allowing the survey to be slightly more accessible; however, further research is needed on the impact of COVID-19 in lower-socio-economic communities in South Africa. Furthermore, outcomes were assessed using self-report measures that relied on participants’ retrospective recall of the last two weeks before the outbreak, which may be vulnerable to recall bias.
Implications
Findings from the study provide important implications for future research and policy-making decisions. While the current study found a significant deterioration in well-being and mental health from before to during the pandemic, these are likely transient stress responses to a global pandemic and research is needed to determine the long-term impact of the COVID-19 pandemic in South Africa. Future research should assess the role of resilience in improving mental health and well-being outcomes during the pandemic. This information could be used to inform targeted interventions, such as stress-regulation training to build resilience (De Visser et al., Reference de Visser, Dorfman, Chartrand, Lamon, Freedy, Weltman and Mallak2016), among those who are at heightened risk of worsened outcomes. As suggested by our study, adequate mental health care service delivery and psychosocial support are critical during times of a global crisis. In the future, public health responses need to address the psychological and social impacts alongside the physical impact of a pandemic. National government responses to a global pandemic need to be adapted to consider the needs of vulnerable population groups, and balance public health and safety with existing social issues. Moreover, online Cognitive Behavioural Therapies (CBTs) that promote adaptive coping strategies (e.g., relaxation techniques and engagement in activities) should be implemented to help individuals who are unable to cope effectively in times of crisis. Internet-based CBT has been shown to effectively reduce depression, anxiety, and insomnia (Olthuis et al., Reference Olthuis, Watt, Bailey, Hayden and Stewart2016; Sijbrandij et al., Reference Sijbrandij, Kunovski and Cuijpers2016; Soh et al., Reference Soh, Ho, Ho and Tam2020), while also minimising in-person contact and infection risk during a pandemic.
Conclusion
The COVID-19 pandemic adversely impacted the well-being and mental health of South African adults. Various non-modifiable and modifiable factors were associated with worsened mental health and well-being, while others were associated with improved outcomes. In future pandemics, government strategies should be directed at reducing the rate of infection without infringing on fundamental human rights to protect groups that are more severely affected.
Data availability
Data for this study can be accessed via the following URL: https://osf.io/bqmgf/?view_only=84a0fd383cd54ad886d1ffe805a2310c
Acknowledgements
This work was supported by the South African PTSD Research Programme of Excellence and the SAMRC Genomics of Brain Disorders Unit.
Author contribution
-
MS, CC, TT, AE, and SS conceptualised and designed the study.
-
GS contributed to data collection.
-
AA analysed the data and drafted the manuscript.
-
SS contributed to supervision.
-
All authors reviewed and approved the final manuscript.
Financial support
Project seed funding was awarded through Stellenbosch University Special Vice-Rector (RIPS) Fund awarded to GS. This funding provided support for data collection.
Competing interests
MS received honoraria/has been a consultant for AbbVie, Angelini, Lundbeck, Otsuka.
CUC has been a consultant and/or advisor to or has received honoraria from: AbbVie, Acadia, Adock Ingram, Alkermes, Allergan, Angelini, Aristo, Biogen, Boehringer-Ingelheim, Bristol-Meyers Squibb, Cardio Diagnostics, Cerevel, CNX Therapeutics, Compass Pathways, Darnitsa, Delpor, Denovo, Gedeon Richter, Hikma, Holmusk, IntraCellular Therapies, Jamjoom Pharma, Janssen/J&J, Karuna, LB Pharma, Lundbeck, MedAvante-ProPhase, MedInCell, Merck, Mindpax, Mitsubishi Tanabe Pharma, Mylan, Neurocrine, Neurelis, Newron, Noven, Novo Nordisk, Otsuka, Pharmabrain, PPD Biotech, Recordati, Relmada, Reviva, Rovi, Sage, Seqirus, SK Life Science, Sumitomo Pharma America, Sunovion, Sun Pharma, Supernus, Tabuk, Takeda, Teva, Tolmar, Vertex, and Viatris. He provided expert testimony for Janssen and Otsuka. He served on a Data Safety Monitoring Board for Compass Pathways, Denovo, Lundbeck, Relmada, Reviva, Rovi, Supernus, and Teva. He has received grant support from Janssen and Takeda. He received royalties from UpToDate and is also a stock option holder of Cardio Diagnostics, Kuleon Biosciences, LB Pharma, Mindpax, and Quantic.
None of the other authors declared any interest.
Open access
