During the COVID-19 pandemic mental health and wellbeing was a concern, especially for the most vulnerable (World Health Organization, 2020; Yu et al., 2021). Global mental health was rapidly studied during the early pandemic, documenting considerable suffering (Mukherjee et al., 2021). Years following the pandemic the question becomes what does the picture look over time? Did mental health and wellbeing worsen, or did people demonstrate resilience and adjust? Tentative evidence suggests that some recovery took place (Manchia et al., 2022; Yarrington et al., 2021). On the other hand, there is also evidence of persistent impacts 6 and 12 months later in COVID-19 survivors (Mazza et al., 2022; Taquet et al., 2021) suggesting long lasting effects. Thus, more careful study is needed to determine what has happened and what to do about it (Gruber et al., 2021; O’Connor et al., 2021; Tegnell, 2021).

During May and June 2020, we conducted a survey in Sweden with the purpose to assess mental health and wellbeing and factors related to these (McCracken et al., 2020). Based on responses from 1,212 adults we found high rates of depression, anxiety, and insomnia, at 30%, 24%, and 38%, respectively. We further found that these outcomes were associated with poor self-rated overall health, a history of mental health problems, presence of COVID-19 symptoms, and specific worries around health and finances (see also Rondung et al., 2021). It is not known whether circumstances improved or declined for mental health, in Sweden or globally, during the years that followed.

An unexpected result from the pandemic is a high rate of Post COVID-19 Condition or PCC (Fernández-de-las-Peñas et al., 2021; Nalbandian et al., 2021). This includes persisting symptoms potentially involving multiple organ systems, including fatigue, breathing problems, cognitive disturbance, problems with the gastrointestinal tract, joints, skin, problems with mood and sleep, and other problems, that follow infection by the coronavirus (Carfì et al., 2020; Hayes et al., 2021; SBU, 2020). The prognosis of persisting symptoms remains unclear, and it appears that additional research and new models of care are needed (Nalbandian et al., 2021). It is unknown how these symptoms correlate with, or perhaps contribute to, mental health outcomes.

The purpose of this study was to repeat our study conducted in Sweden during June 2020 and reassess the state of mental health and wellbeing 18 months following the start of the pandemic using an independent population sample. The outcomes were levels of depression, anxiety, and insomnia as they are sensitive to stressful events and serve as important predictors of both emotional and physical health. The second purpose was to examine the prevalence of PCC (i.e., symptoms lasting for at least six weeks following infection (e.g., physical, and mental dysfunctions and cognitive impairments – see method section for full description) and their relations with mental health outcomes. Additional questions addressed included the identification of risk factors previously associated with health outcomes, such as a previous history of mental health problems and COVID-19 related worry.

Initially, 1,657 people provided their consent and entered the survey. In general, there was an increase in missing data with each subsequent measure in the survey, due to dropout, as opposed to skipped or missed items. There was a 10.6% non-completion rate of the three mental health outcome measures and the sample size retained was n = 1,482. However, sample size varied slightly from analysis to analysis due to small amounts of missing data on other individual variables.

Participant characteristics are included in Table 1. Notably, the percentage of women who participated was 89.5%. Mean age of the total sample was 47.6 years, SD = 11.7, and range was 18 to 81. The participants were well educated with 61.1% having completed university or post graduate education. The vast majority of participants were from Sweden or another Scandinavian country, 90.1%. Most were married or in a relationship, 75.0%, or single, 18.3%, and a little less than half had young children, 45.9%. Most participants were working either full or part time, 80.3%, and the next largest category was retired, 6.8%. A substantial proportion reported a history of mental health difficulties, 43.2%. Fewer, 27.9% of the total sample, were currently experiencing these, and 23.9% reported they had a formal diagnosis.

Demographic variables included those detailed in Table 1. To further clarify, participants were also asked whether they (a) lived in suburbs, city, or countryside; (b) had children living at home; and (c) the number of people living in their home in total; whether (d) they had received a formal diagnosis of a mental health condition; (e) were experiencing one of these currently; or (f) were suffering with any of the identified conditions placing people at risk of a poor outcome from COVID-19, including age over 70, hypertension, angina, stroke, heart disease, diabetes, cancer, smoking, respiratory disease, and immune suppressant. Specific worries about COVID-19 were assessed with items used in our previous study (McCracken et al., 2020). Here participants rated their worries about their own health, others’ health, their personal finance, world economy, and the future, on a 5-point scale from 1 (Not at all worried) to 5 (Extremely worried). The respondents’ answers to the five worry items were summed up to calculate a total worry score (α = .77).

Data were analyzed using IBM SPSS version 26.0. First, sample characteristics including rates of infection, vaccination, other health-related descriptors, and rates of cases meeting clinical cutoffs for depression, anxiety, insomnia, and persistent physical symptoms were analyzed. Next Pearson correlation coefficients were calculated to identify background or health status factors significantly associated with depression, anxiety, and insomnia scores. Categorical background variables, such as work and relationship status, were dichotomized as labelled, for example, into “out of work” versus not, and “in a relationship” or not. A second set of correlation analyses included depression, anxiety, and insomnia with each of the individual persistent physical symptom reports and the summary of the total number reported, excluding sleep, depression, anxiety, quality of life, and daily functioning, to avoid inflating the correlations. Additional correlation analyses examined relations between participant attributions of symptoms to a COVID-19 infection, this time analyzing the full symptom set again examining whether the symptom was regarded as a direct result of infection with the COVID-19 virus or not. Finally, hierarchical multiple regression analyses were conducted, with depression, anxiety, and insomnia scores as the criterion variables. In these analyses, age, being in a relationship, being out of work, and having above average financial status were included as background variables in the first block of predictors, based on the correlation analyses. The second block of predictors included relevant health status variables, including self-rated physical health, reported mental health history, and a summary score of risk factors for poor COVID-19 outcome. The third block included reported infection with COVID-19. The fourth block included the summary score for COVID-19-related worry. And finally, a selected set of persistent physical symptom reports plus the summary total of symptoms reported was included in the final block of predictors. The included individual symptoms were selected based on having achieved a medium sized correlation or larger with either depression, anxiety, or insomnia in the correlation analyses.

Table 1 includes descriptive statistics on COVID-19 infection and vaccination related data.

Table 1 includes summary results from the measures of depression, anxiety, and insomnia. Mean values fall in the mild or subthreshold range, and most people fall below the clinical cutoff in each case. On the other hand, 27.4%, 16.2%, and 45.0%, met criteria for clinically significant depression, anxiety, and insomnia, respectively, based on cutoff scores > 10. Further, 13.6% reported some frequency of thoughts of being better off dead or self-harm (PHQ-9 item 9), and 67.3% reported that depression symptoms were associated with some level of difficulty in work, home, or social activities (PHQ-9 item 10). For potential comorbid presentations of these three conditions, 50.6% met criteria for at least one, 25.6% for at least two, and 12.7% for all three. Naturally these measures are intercorrelated, with depression and anxiety scores correlated at r = .82, depression and insomnia at r = .67, and anxiety and insomnia at r = .65.

Data on worries about own health, health of friends or family, own finances, national or international finances, or the future, was provided on a scale from 1 to 5, from not at all worried to extremely worried. The highest area of worry was around health of friends or family, M = 2.85, SD = 1.25, followed by national or international economy, M = 2.68, SD = 1.19, and the future, M = 2.68, SD = 1.30. Worry about own health, M = 2.43, SD = 1.91, and own finances, M = 1.99, SD = 1.23, were lowest ranked.

Overall, 84.5% of survey participants reported at least one persisting COVID-19 symptom from the set of 25 and the mean number of symptoms reported was M = 6.5, SD = 5.3. At least one symptom was reported by 78.0% of those who reported they were never infected with COVID-19 and 90.5% of those who reported that they had been. These percentages, although both high, were significantly different, χ²(1, N = 1,552) = 46.3, p < .001. With the removal of the symptom items related to sleep, depression, anxiety, quality of life, and daily functioning, high percentages of participants reported at least one of the remaining 20 symptoms, including 73.1% of those who reported never having COVID-19 and 88.2% of those who reported having it, χ²(1, N = 1,552) = 57.0, p < .001.

Figure 1 includes the percentages for persons reporting each of the 25 persistent symptoms potentially associated with COVID-19 infection. The figure includes both the percentage reporting the symptom and the percentage of those who attributed the symptom to COVID-19. The most frequent persistent symptom overall was fatigue, reported by 55.4% and 48.0% of those (24.9% of the overall sample), attributed the fatigue directly to COVID-19. A total of nine symptoms were reported by 30% or more of respondents, and these included, in addition to fatigue, sleep disturbance, difficulties with attention, reduced quality of life, joint pain, difficulties with memory, depression, headache, and reduced daily functioning. For about half of the respondents who reported decreased quality of life or daily functioning these effects were attributed to COVID-19. For all the other symptoms the percentage was lower. There were several symptoms where the majority of those who reported the symptom also attributed it to COVID-19, at rates all greater than 67%. These included shortness of breath, reduced sense of smell, reduced sense of taste, and reduced lung function.

Three separate hierarchical multiple regression analyses were conducted with depression, anxiety, and insomnia as dependent variables (see Table 4). These were to examine the combined role of background, health status, COVID-19 infection, worry, and persistent post-COVID-19 symptoms in accounting for variance in these outcomes.

Amongst these blocks of variables, history of COVID-19 infection was the least informative, accounting for less than 1% of variance in all cases. The four background variables, including age, relationship status, employment status, and self-rated financial status, accounted for 8.4%, 7.6%, and 4.1% of variance in depression, anxiety, and insomnia, respectively. Older age significantly predicted lower scores for depression and anxiety, and being single, unemployed, and, unexpectedly, being well off financially, each contributed significantly to higher score for insomnia, although the variance accounted for was modest.

The health status block including self-rated health, mental health history, and the summary score of physical conditions representing risks for poor COVID-19 accounted for 27.0%, 21.0%, and 17.0% of variance. Here the health rating and mental history were significant predictors, but the physical risk factors were not. The single best variable in the equations was the total COVID-related worry scores, accounting for 9.2%, 10.0%, and 7.5% of variance in the outcomes. With all other potential predictors included, on average the twelve persistent physical symptoms with medium effect sizes and the summary score for the number reported (adjusted to exclude depression, anxiety, sleeping problems or reduced quality of life or daily functioning) were relatively good predictors of outcomes as a set, accounting for 14.0% or variance in depression, 9.0% for anxiety, and 7.7% for insomnia. The specific symptoms most strongly related to the three outcomes were fatigue and headache, which both significantly predicted all three, and attention problems, which was a relatively strong predictor of both depression and anxiety, but not insomnia. Finally, we reran the regression analyses selecting only the participants who had reported a history of COVID-19 infection and the results did not change appreciably. For example, the variance accounted for by the persistent physical symptoms remained similar or the same, 14.0% for depression, 8.4% for anxiety, and 7.5% for insomnia.