Background
On 11 March 2020, the World Health Organization declared the coronavirus disease 2019 (COVID-19) crisis to be a pandemic.Reference Bedford, Enria, Giesecke, Heymann, Ihekweazu and Kobinger1 The worldwide administration of the vaccine began on 8 December 2020.Reference Dagan, Barda, Kepten, Miron, Perchik and Katz2,Reference Knoll and Wonodi3 However, considering the production volume and the speed with which vaccines can be administered, achieving herd immunity may take some time.Reference Fontanet and Cauchemez4 Thus, COVID-19 is still being considered an important global health crisis.
Approximately 2.2% of individuals worldwide who contracted COVID-19 died because of the infection,5 meaning that 97.8% of people who contracted COVID-19 are survivors. This shows that quality of life of the survivors and sequelae have emerged as important public health issues.Reference Garrigues, Janvier, Kherabi, Le Bot, Hamon and Gouze6 Psychological sequelae has been reported as an important issue in current literature.Reference Moradi, Mollazadeh, Karimi, Hosseingholipour and Baghaei7–Reference Park, Jung, Park, Lee, Kim and Kim9 In a previous cohort study from the USA, survivors of the diseases who had no previous psychiatric history had a higher risk of developing psychiatric disorders during the 14–90 days follow-up period.Reference Taquet, Luciano, Geddes and Harrison10 In South Korea, a high association of psychological sequelae in COVID-19 survivors has been reported in previous studies.Reference Oh, Park and Song8,Reference Park, Jung, Park, Lee, Kim and Kim9 However, to date, studies have only evaluated the total prevalence of mental illness during a short-term follow-up period of 1 month.Reference Park, Jung, Park, Lee, Kim and Kim9,Reference Oh, Park and Song11 Thus, there is an urgent need to evaluate the overall mental illness levels with a longer follow-up period.
Aims
This study investigates the prevalence of mental illness and the associated factors for its development in COVID-19 survivors. It hypothesises that COVID-19 survivors are more likely to develop mental illness compared with other individuals. Moreover, we hypothesised that COVID-19 survivors who experienced specific treatment for the infection in a hospital might have a higher prevalence of mental illness compared with other survivors.
Method
As a population-level cohort study, guidelines from the Strengthening the Reporting of Observational Studies in Epidemiology have been followed.Reference Von Elm, Altman, Egger, Pocock, Gøtzsche and Vandenbroucke12 The protocol of this study was approved by the Institutional Review Board of the Seoul National University Bundang Hospital; approval number: X-2009-636-902. The National Health Insurance Service (NHIS) sharing service also approved the study protocol and permitted the use of the NHIS COVID-19 database (NHIS-2021-1-070). The need for informed consent was waived because the data in this study were analysed retrospectively in an anonymised form from the NHIS database. The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.
NHIS COVID-19 database
The NHIS COVID-19 database has been created for purposes of medical research in cooperation with the Korea Disease Control and Prevention Agency (KDCA) and the NHIS. The KDCA provided data for all individuals whose polymerase chain reaction (PCR) test indicated a positive result, confirming that the individual had contracted the infection for the period of 1 January to 4 June 2020. These data included date of confirmation and results of treatment in addition to demographic information. The NHIS extracted the control population through stratification methods based on age, gender and location of their home, using ZIP code as of February 2020. Next, the KDCA provided the data on individuals who received a PCR test, and the result was negative. Therefore, the NHIS COVID-19 database contains data for three groups: a COVID-19-positive group; a control population; and PCR negative-tested individuals.
As the only public health insurance system, the NHIS possesses information regarding all disease diagnoses using ICD-10Reference World Health13 codes and any drug prescription and/or procedure information. The NHIS COVID-19 database was first developed on 26 June 2020, and last updated on 1 December 2020. The NHIS COVID-19 database had no missing data except for the annual income level for some individuals.
Study population
The COVID-19 survivors in this study are defined as individuals who were diagnosed with COVID-19 and then discharged from hospital after treatment. Additionally, individuals who were diagnosed with COVID-19 but were not admitted to hospital because they only had mild or no symptoms are also considered as COVID-19 survivors, if they recovered from the infection. In South Korea, individuals who were diagnosed with COVID-19 were admitted to hospital if they had severe symptoms, such as pneumonia. However, if they had no symptoms or only had mild symptoms, they were isolated with close monitoring in specific government-managed centres.
The other population in the NHIS COVID-19 database has been defined as the control group in this study, which consisted of the control group and PCR test-negative individuals. Among the study population, individuals under the age of 20 were excluded because the prevalence and risk factors for mental illness are different in children and adolescents.Reference Park, Kim, Cho, Kim, Shin and Yoo14 In addition, individuals who had a previous history of mental illness up until 31 December 2019, were excluded from the analysis to focus on newly developed mental disorders in 2020.
Endpoint: development of mental illness
The primary end-point of this study was the development of mental illness. To define mental illness using ICD-10 codes, the criteria used by Lee et alReference Lee, Yang, Moon, Yoo, Ha and Kim15 were used. Mental illness includes non-affective psychotic disorders (F20–24 and F28–29), affective psychotic disorders (F25, F30–31, F32.3 and F33.3), anxiety and stress-related disorders (F40–48), alcohol or drug misuse (F10–16 and F18–19), mood disorders without psychotic symptoms (F32–34, F38–39, excluding F32.3, F33.3), eating disorders (F50) and personality disorders (F60–63 and F68–69). The development of mental illness among the study population was determined according to registered ICD-10 codes from 1 January to 1 December 2020. Thus, the diagnosis of mental illness among the study population was determined in the approximately 6-month follow-up time as the data were from 1 January to 4 June 2020. In South Korea, the ICD-10 codes of mental illness must be registered by physicians or psychiatrists in order for patients to receive financial support for the treatment of mental illness.
Confounders
The following information was collected and considered as confounders: demographic variables (gender and age), residence and annual income level in 2020, underlying disability (registered on the NHIS database) and Charlson Comorbidity Index (CCI) score, which was calculated using registered ICD-10 diagnostic codes during the 4 years from 2015 to 2019 (Supplementary Table 1 available at https://doi.org/10.1192/bjo.2021.1001). All individuals were divided into seven groups according to age (20–29, 30–39, 40–49, 50–59, 60–69, 70–79 and ≥80 years), and residence was divided into five areas (Seoul, Gyeonggido, Daegu, Gyeongsangbukdo and other areas). In South Korea, the annual income level of all individuals must be registered in the NHIS database to determine individual insurance premiums, which is divided into quartiles (Q1, Q2, Q3, Q4 and unknown).
Additionally, treatment information for all patients with COVID-19 was collected to reflect severe COVID-19 infection similar to a previous study,Reference Shin, Shin, Moon, Jin, Kim and Yang16 including supplemental oxygen therapy, mechanical ventilator use, continuous renal replacement therapy, extracorporeal membrane oxygenation use and high-flow oxygen therapy. Thus, if the survivors were given any specific treatment based on the treatment information, they were considered to be a part of a specific treatment group.
Statistical methodology
The clinicoepidemiological characteristics of all individuals in the database are presented as mean values with s.d.s for continuous variables (CCI) and as numbers with percentages for categorical variables (all other variables except for CCI). To compare the characteristics between COVID-19 survivors and the control group, a t-test and chi-square test were used for continuous and categorical variables, respectively. Subsequently, uni- and multivariable logistic regression analyses were performed to examine if COVID-19 survivors were associated with a higher prevalence of development of mental illness in 2020, compared with the control group.
In the multivariable model (model 1 as main analysis), all covariates were included in the adjustment. Additionally, the COVID-19 survivors were divided into two groups (no specific treatment for COVID-19 group and specific treatment for COVID-19) and included in model 2 to examine the impact of treatment experience on the development of mental illness. Second, the duration of isolation because of COVID-19 was included in model 3 as a continuous variable. A multivariable logistic regression model was used to analyse the development of mental illness; this excluded the individuals who died in 2020, as the deaths during this period (evaluation period) could affect the development of mental illness in both the COVID-19 survivor group and the control group.
Next, for the competing risk analyses, mental illness was divided into seven types: non-affective psychotic disorders, affective psychotic disorders, anxiety and stress-related disorders, alcohol or drug misuse, mood disorders without psychotic symptoms, eating disorders and personality disorders. Seven multivariable logistic regression models were constructed using all covariates to examine which form of mental illness was mostly observed in COVID-19 survivors compared with the control group. Finally, subgroup analyses were performed using multivariable models according to gender (male and female), age (20–39, 40–59 and ≥60 years old), and CCI (0–2 and 0–3 groups). All multivariable models confirmed that there was no multicollinearity between the variables with a variance inflation factor of <2.0, and Hosmer–Lemeshow statistics were used to confirm that the goodness of fit of the models was appropriate at P > 0.05. The results of the logistic regression analyses are provided as odds ratios (ORs) with 95% CIs. R software (version 4.0.3; R Foundation for Statistical Computing, Vienna, Austria) was used for all statistical analyses, and a P-value of <0.05 was determined as statistically significant.
Results
Study population
A flow chart depicting the individual selection process in this study is presented in Fig. 1. The NHIS COVID-19 cohort initially included 351 377 individuals. We then excluded from the analysis 23 003 individuals aged less than 20 years, 67 376 individuals who had a mental illness before 2020, and 114 patients with COVID-19 who died from the infection during their hospital admission. Therefore, 260 883 individuals were included in the analysis. Among these, the number of COVID-19 survivors was 6148 (2.36%), and there were 254 735 individuals in the control group. The clinicoepidemiological characteristics of the participants are presented in Supplementary Table 2.
Fig. 1 Flow chart for the study. COVID-19, coronavirus disease 2019; NHIS, National Health Insurance Service.
Among the total study population, 20 262 (7.8%) participants were diagnosed with mental illness, and anxiety and stress-related disorders were the most common and were observed in 13 736 (5.3%) participants, followed by mood disorders without psychotic symptoms in 8089 (3.1%) participants. The results of the comparison of clinicoepidemiological characteristics between COVID-19 survivors and the control group are presented in Table 1.
Table 1 Comparison of clinicoepidemiological characteristics between COVID-19 survivor and the control group
COVID-19, coronavirus disease 2019.
The proportion of men in the COVID-19 survivors was higher (39.2%; 2414/6148) than that in the control group (54.2%; 137 977/254 735) with a P < 0.001. The proportion of ≥80-year-olds and 70- to 79-year-olds in the COVID-19 survivors were lower (2.7% (165/6148) and 6.0% (369/6148), respectively), compared with the control group (5.6% (14 264 of 254 735) and 7.5% (19 093 of 254 735), respectively). The prevalence of mental illness in 2020 was higher in the COVID-19 group (738/6148, 12.0%) than in the control group (19 524/254 735, 7.7%, P < 0.001).
Development of mental illness in 2020
The results of the logistic regression analysis for the development of mental illness in 2020 are presented in Table 2. In model 1, the prevalence of mental illness in the COVID-19 survivors was 2.40-fold higher than that in the control group (OR = 2.40, 95% CI 2.21–2.61, P < 0.001).
Table 2 Logistic regression analysis for development of mental illness in 2020a
COVID-19, coronavirus disease 2019.
a. Other variables in the multivariable model 1 are presented in Supplementary Table 3.
In model 2, compared with the control group, the no specific treatment for COVID-19 subgroup (n = 5256) and the specific treatment for COVID-19 subgroup (n = 892) had a 2.23-fold (OR = 2.23, 95% CI 2.03–2.45, P < 0.001) and 3.27-fold times (OR = 3.27, 95% CI 2.77–3.87, P < 0.001) higher prevalence of mental illness among survivors, respectively.
In multivariable model 3, a 1-day increase in the duration of isolation (in government centres and/or days spent in hospital) because of COVID-19 was associated with a 2% higher prevalence of mental illness (OR = 1.02, 95% CI 1.01–1.02; P < 0.001). The ORs with 95% CIs for all other variables in the multivariable model 1 are presented in Supplementary Table 3. A similar trend was observed in the sensitivity analysis, excluding 6467 individuals who died in 2020, as shown in Supplementary Table 4.
The results of the logistic regression analysis for the development of mental illness in detail are presented in Table 3. The COVID-19 survivors showed higher prevalence of affective psychotic disorders (OR = 1.44, 95% CI 1.13–1.85, P = 0.003), anxiety and stress-related disorder (OR = 1.72, 95% CI 1.56–1.89, P < 0.001) and mood disorders without psychotic symptoms (OR = 2.15, 95% CI 1.92–2.41, P < 0.001) than the control group.
Table 3 Multivariable logistic regression analysis for development of mental illness in detail
COVID-19, coronavirus disease 2019.
The results of the subgroup analyses are shown in Table 4. The ORs for mental illness in the COVID-19 survivors were 1.67 (OR = 1.67, 95% CI 1.45–1.92, P < 0.001) in the male group and 2.06 (OR = 2.06, 95% CI 1.87–2.27, P < 0.001) in the female group. The ORs for mental illness in the COVID-19 survivors were 1.60 (95% CI 1.36–1.89, P < 0.001), 2.27 (95% CI 1.99–2.58, P < 0.001) and 1.86 (OR = 1.86, 95% CI 1.63–2.13, P < 0.001) in the 20–39, 40–59, and ≥60 years groups, respectively. Those individuals with 0–2 and ≥3 points on the CCI showed the same ORs for mental illness of 1.93.
Table 4 Subgroup analyses
COVID-19, coronavirus disease 2019.
Discussion
Main findings
This nationwide cohort study in South Korea established that, compared with the control group, COVID-19 survivors are associated with a 2.40-fold higher prevalence of mental illness, at least during the 6-month follow-up period. Among mental illnesses, a higher prevalence of mood disorders, anxiety and stress-related disorders, and affective psychotic disorders were associated with the survivors. Furthermore, this association was more evident in COVID-19 survivors who experienced specific treatment for the infection in hospitals. The study suggests that newly developed mental illness is a significant issue observed in COVID-19 survivors.
Comparison with findings from other studies
Using the same database (NHIS COVID-19), we recently reported that COVID-19 survivors were at higher risk of psychological sequalae.Reference Oh, Park and Song8 However, the results of the previous study differ from those of the current study in several ways. First, the previous study was based on the database from 1 January to 4 June 2020, whereas the current study used the updated data from 1 January to 1 December 2020. Therefore, we were able to evaluate in the current study the development of mental illness across a longer period of 6 months. Second, different criteria were used for determining the primary end-point. Whereas the previous study focused on the development of depression, psychosis, alcohol misuse and drug misuse, we included more mental illnesses in this study, using the criteria used in the study by Lee et al.Reference Lee, Yang, Moon, Yoo, Ha and Kim15 Therefore, anxiety and stress-related disorders, eating disorders and personality disorders were evaluated as end-points in the current study. Finally, whereas the previous study only included data on COVID-19 survivors and a control population, extracted using stratification methods, the current study included test-negative individuals as well. Thus, we examine a significantly larger sample size compared with the previous study.
In South Korea, the association of mental illness with susceptibility and outcomes of COVID-19 have been sufficiently reported,Reference Lee, Yang, Moon, Yoo, Ha and Kim15,Reference Lee, Yang, Moon, Kim, Ahn and Kim17 but information regarding the new development of mental illness among survivors was lacking. Previous studies suggest that psychological sequelae are an important issue among COVID-19 survivors.Reference Moradi, Mollazadeh, Karimi, Hosseingholipour and Baghaei7,Reference Park, Jung, Park, Lee, Kim and Kim9–Reference Oh, Park and Song11 In the USA, survivors of COVID-19 have been found to experience psychiatric disorders 14–90 days after a COVID-19 diagnosis, and anxiety disorders, insomnia and dementia were mostly observed.Reference Taquet, Luciano, Geddes and Harrison10 In South Korea, the development of psychological sequelae has been recently reported in COVID-19 survivors.Reference Park, Jung, Park, Lee, Kim and Kim9,Reference Oh, Park and Song11 The present study was conducted during a 6-month follow-up of COVID-19 survivors to evaluate psychiatric diagnosis and is different from previous studies that evaluated mental illness during a relatively short period.Reference Park, Jung, Park, Lee, Kim and Kim9–Reference Oh, Park and Song11 Furthermore, unlike previous studies,Reference Park, Jung, Park, Lee, Kim and Kim9–Reference Oh, Park and Song11 the impact of COVID-19 severity on the development of mental illness includes information about treatment received for COVID-19.
In the competing risk analyses, a higher prevalence of mood disorders, anxiety and stress-related disorders, and affective psychotic disorders were observed in COVID-19 survivors. Depression is a major form of mood disorder, and post-traumatic stress disorder (PTSD) and anxiety disorders are major types of anxiety and stress-related disorders. A previous cohort study reported that a high prevalence of psychiatric sequelae such as PTSD, depression and anxiety were observed among 402 COVID-19 survivors at a 1-month follow-up after admission to hospital.Reference Mazza, De Lorenzo, Conte, Poletti, Vai and Bollettini18 The prevalence of mental illness among COVID-19 survivors is comparable with psychiatric sequelae for other traumatic experiences. A previous cohort study in Australia reported that 22% of people experienced a new psychiatric disorder after traumatic brain injury.Reference Bryant, O'Donnell, Creamer, McFarlane, Clark and Silove19 Specifically, patients were more likely to develop PTSD, panic disorder, agoraphobia and social phobia significantly at 12 months after traumatic brain injury.Reference Bryant, O'Donnell, Creamer, McFarlane, Clark and Silove19 Compared with the survivors from traumatic brain injury,Reference Bryant, O'Donnell, Creamer, McFarlane, Clark and Silove19 we showed that a higher prevalence of mood disorder was observed in the COVID-19 survivors in our study.
Another previous cohort study in South Korea reported that 42.9% and 27.0% of Middle East Respiratory Syndrome (MERS) survivors reported PTSD and depression, respectively, at 1 year after the MERS outbreak.Reference Park, Park, Lee, Kim, Lee and Lee20 In a multivariable analysis, anxiety and recognition of stigma during the MERS outbreak period were significant risk factors for developing PTSD at 12 months after the MERS outbreak.Reference Park, Park, Lee, Kim, Lee and Lee20 Compared with the MERS survivors,Reference Park, Park, Lee, Kim, Lee and Lee20 the COVID-19 survivors showed a significant increase in the prevalence of psychotic disorders in our study. In 2021, Taquet et al reported that COVID-19 survivors in the USA had a 1.81-fold higher prevalence of mood, anxiety or psychotic disorder compared with those diagnosed with influenza infection at 6-months follow-up.Reference Taquet, Geddes, Husain, Luciano and Harrison21 Considering the results in our study and those of Taquet et al, the COVID-19 survivors were at a higher risk of developing mental illnesses, which includes mood, anxiety, or psychotic disorders.Reference Taquet, Geddes, Husain, Luciano and Harrison21
Interpretation of our findings
This study establishes that COVID-19 survivors who did not require any specific treatment for infection were 2.23-fold more likely to have a mental illness than those in the control group. In South Korea, patients with COVID-19 who had mild or no symptoms were isolated with monitoring in government-managed centres until the results of the PCR test were negative. Individual stress could be an important factor leading to this phenomenon. First, the COVID-19 pandemic leads to changes in interpersonal relationships and lifestyles because of prolonged periods of solitude and loneliness. These psychosocial stressful conditions have detrimental effects on fragile individuals and affect their ability to modulate emotions.Reference Li, Wang, Xue, Zhao and Zhu22,Reference Berg-Weger and Morley23 Second, the higher prevalence of mental illness among COVID-19 survivors might be caused by social stigma among individuals, as reported in previous studies.Reference Bagcchi24,Reference Singh and Subedi25 Among survivors in South Korea, social stigma during the MERS outbreak was an independent risk factor for developing PTSD after 12 months.Reference Park, Park, Lee, Kim, Lee and Lee20 Furthermore, a recent study in Germany reported that patients who had COVID-19 were at a higher risk of unemployment because of the social stigma attached to the infection.Reference Dragano, Rupprecht, Dortmann, Scheider and Wahrendorf26 Such unemployment could lead to depression according to a previous epidemiological study.Reference Dooley, Catalano and Wilson27 Therefore, this study establishes that mental health in COVID-19 survivors is a significant public health issue not only for those who were severely ill but also for patients with mild or no symptoms. However, given that we could not demonstrate the direct impact of emotional stress on mental illness development in those COVID-19 survivors with mild or no symptoms in this study, more research is needed regarding this issue.
Limitations
This study has some limitations. First, the variables body mass index, history of alcohol consumption and smoking were not included in the adjustment as the NHIS database did not include this information. Second, the difference in sample size between the control group and the COVID-19 survivor group is of a large order of magnitude: 254 735 individuals in the control group and 6148 individuals in the COVID-19 survivor group. Third, the multivariable logistic regression is known to adjust only measured confounders, and there might be unmeasured confounders that might affect the results. Fourth, although mental illnesses should be registered in the NHIS database for individuals to receive proper financial support from the government, some data for some patients may be missing because of the lack of accessibility to medical resources among the national population in South Korea. Finally, we included information about patients who required and received specific treatment for their COVID-19 infection and used this to reflect the severity of COVID-19; an accurate measure of the severity of the COVID-19 infection was not considered in this study.
To conclude, during the approximately 6-month follow-up period, COVID-19 survivors were at a higher risk of mental illness compared with the rest of the population of South Korea. Moreover, this association was more evident in COVID-19 survivors who experienced specific treatment for the infection in hospital. To enhance the quality of life of COVID-19 survivors, proper treatment of such mental illnesses should be an important health consideration in future.
Supplementary material
Supplementary material is available online at https://doi.org/10.1192/bjo.2021.1001.
Funding
This research was supported by a grant of Patient-Centered Clinical Research Coordinating Center (PACEN) and COVID-19 Public Health Research funded by the Ministry of Health & Welfare, Republic of Korea (grant number : HI19C0481, HC20C0013).
Acknowledgements
None.
Data availability
Data are available upon reasonable request. Anonymised data used in the present study may be available upon reasonable request to the corresponding author.
Author contributions
T.K.O. designed the study, analysed the data, interpreted the data and drafted the manuscript; H.Y.P. and I.-A.S. contributed to the study conceptualisation, acquisition of data and review of manuscript.; S.H.L., M.y.S., H.S.O., K.-H.S., E.-S.Y. and H.Y.P. contributed to data collection, project administration and review of manuscript. All authors have given final approval for the final version of the manuscript.
Declaration of interest
None
Open access
eLetters
No eLetters have been published for this article.