Participants
Participants who have contracted the new coronavirus were surveyed using a web-based cross-section analysis from July to September 2021. The study sample was from the pooled panels of an internet research agency (Rakuten Insight, Inc.), which had approximately 2.2 million panelists in 2019. All participants provided web-based informed consent at registration. Only the participants who responded “yes” to the first question, “Have you ever been infected with COVID-19?” were asked to complete the questionnaire. We excluded participants who completed the questionnaire (n = 7760) who (1) incorrectly answered the dummy question (n = 1195); (2) disclosed not having been infected during the survey (n = 6); (3) gave inconsistent answers about physical symptoms (n = 454); (4) gave improbable answers about the postinfection period (not within 0–20 months) judging from the date of confirmed the first case of COVID-19 infection in Japan (January 15, 2020)16 (n = 84); (5) answered in the open-ended section, which could not categorize existing choices of the questionnaire (n = 5); and (6) identified some fault in data (n = 1). The participant flow chart is shown in Fig. 1. Finally, 6015 individuals (response rate = 77.5%) were included in the analyses. The study was in accordance with the Declaration of Helsinki. All methods were performed in accordance with the relevant guidelines and regulations. The participation was anonymous. A credit point that could be used for internet shopping and cash conversion was provided for participants with an incentive. We used the term gender according to the SAGER guidelines17. This study was approved by the Ethical Board of the National Center of Neurology and Psychiatry in Japan (A2021-34).
Measurement
Outcome variables
Post-COVID-19
Post-COVID-19 was dichotomized into “yes” and “no.” We identified post-COVID-19 based on the WHO definition2. We asked the question: “What symptoms do you have?” Participants could choose multiple answers from the options of symptoms based on the WHO definition of post-COVID-19. Responses were categorized as “yes” when participants chose any one of these considered physical symptoms (menstrual and period problems, altered smell, altered taste, blurred vision, chest pain, cough, dizziness, fatigue, [intermittent] fever, gastrointestinal issues [diarrhea, constipation, or acid reflux], headache, muscle pain or spasms or neuralgias, shortness of breath, tachycardia or palpitations, and tinnitus and other hearing issues). Post-COVID-19 that participants described in the open-ended section were reviewed by four researchers and allocated the type of symptoms they should be applied to.
Exposure variables
Preexisting psychiatric disorders
Preexisting psychiatric disorders were categorized by the answers to the question: “Have you ever been diagnosed with or experienced psychiatric problems before the COVID-19 pandemic?” Participants chose multiple answers from the following options of symptoms: “Nothing,” “Depressive disorder,” “Bipolar disorder,” “Panic attack or panic disorder,” “Anxiety disorder or anxiety-related problems (e.g., hypersensitivity, worry, fear, obsessive–compulsive symptoms),” “Alcohol use disorder or alcohol abuse/dependence,” “The use of illicit substances or psychotropics without prescription,” “Burnout syndrome,” and “Others” (with an optional comment field). Preexisting psychiatric disorders were dichotomized into “yes” and “no” based on the Diagnostic and Statistical Manual of Mental Disorders Fifth Edition (DSM-5)18. Those who chose any one of these disorders (depressive disorder, bipolar disorder, panic attack or panic disorder, anxiety disorder, or anxiety-related problems [e.g., hypersensitivity, worry, fear, and obsessive–compulsive symptoms]) were categorized into the group “yes” preexisting psychiatric disorders. Those who chose options only “Nothing,” “The use of illicit substances or psychotropics without prescription,” or “Burnout syndrome” were categorized into the group “no” preexisting psychiatric disorders. We considered that “The use of illicit substances or psychotropics without prescription” did not meet the diagnostic criteria for alcohol use disorder without the information on duration, amount of use, and symptoms. Moreover, the DSM-5 does not include Burnout syndrome. Among those who chose only “Others,” those whose comments in the comment field matched the DSM-5 diagnosis were categorized into the group “yes” preexisting psychiatric disorders.
Covariates
Covariates were selected based on the basis of previous studies of associated factors of post-COVID1918,19,20,22 and psychiatric disorders.
Kessler6 (K6)
Psychological distress was measured using K623. It consists of six items assessing the frequency of psychological distress occurring in the last 30 days. The response choices are from 0 (none of the time) to 4 (all of the time), and the total score ranges from 0 to 24. The total scores of K6 were categorized as follows: no (≤ 4), slight or moderate (5–12), and severe (≥ 13).
Generalized Anxiety Disorder-7 (GAD-7)
Anxiety symptom was measured using GAD-724. It consists of seven items assessing the frequency of symptoms of anxiety occurring in the last 2 weeks. The response choices are from 0 (not at all) to 3 (nearly every day), and the total score ranges from 0 to 21. The GAD-7 total scores were categorized as follows: no (≤ 4), slight (5–9), moderate (10–14), and severe (≥ 15).
Other covariates
We measured other covariates as follows: postinfection period (< 1 month, ≥ 1 to < 3 months, ≥ 3 to < 6 months, ≥ 6 to < 12 months, or ≥ 12 months), early dyspnea of COVID-19 (yes or no), more than five early symptoms of COVID-19 (yes or no)21, treatment of COVID-19 (no hospitalization, hospitalization without the ICU stay, or hospitalization with the ICU stay), age (20–29, 30–39, 40–49, 50–59, or ≥ 60 years), gender (male, female, or other), educational attainment (high school or lower, some college [e.g., junior college], or university graduate or higher), work status (self-employed, permanent employment, temporary employment, unemployed, or student), cohabitation (yes or no), and medical history (yes or no). The responses were categorized as “yes” when participants answered that they had any of the following medical histories during the survey: hypertension, diabetes, asthma, bronchitis or pneumonia, atopic dermatitis, angina pectoris, cardiac infarction, chronic obstructive pulmonary disease, or cancer.
Statistical analyses
The post-COVID-19 prevalence ratios (PRs) were estimated using Poisson regression analysis with a robust error variance. We used this model because the post-COVID-19 prevalence is over 10%, and the odds ratio could overestimate the PRs25. Variance inflation factor (VIF) was used to check for multicollinearity. Most of the VIF values were less than 2, and the mean VIF of the model was < 2. We entered K6 and work status as covariates because we considered that these were important covariates, although they were over 2 (K6: ≥ 13 = 2.18; work status: self-employed = 2.28 and permanent employment = 2.01). We also estimated the interaction between preexisting psychiatric disorders and postinfection periods. We then conducted a subgroup analysis by postinfection periods because the interaction between preexisting psychiatric disorders and postinfection periods was significant. All analyses were performed using Stata 17.0 (Stata Corp, College Station, TX, USA).
