Mental health improvement after the COVID-19 pandemic in individuals with psychological distress

Abstract

The COVID-19 pandemic and associated countermeasures had an immensely disruptive impact on people's lives. Due to the lack of systematic pre-pandemic data, however, it is still unclear how individuals' psychological health has been affected across this incisive event. In this study, we analyze longitudinal data from two healthy samples (N = 307) to provide quasi-longitudinal insight into the full trajectory of psychological burden before (baseline), during the first peak, and at a relative downturn of the COVID-19 pandemic. Our data indicated a medium rise in psychological strain from baseline to the first peak of the pandemic (d = 0.40). Surprisingly, this was overcompensated by a large decrease of perceived burden until downturn (d = - 0.93), resulting in a positive overall effect of the COVID-19 pandemic on mental health (d = 0.44). Accounting for this paradoxical positive effect, our results reveal that the post-pandemic increase in mental health is driven by individuals that were already facing psychological challenges before the pandemic. These findings suggest that coping with acute challenges such as the COVID-19 pandemic can stabilize previously impaired mental health through reframing processes.

Figure 1

Trajectories of psychological strain as a function of time and gender. Trajectories of psychological strain are shown for females (red, N = 230) and males (blue, N = 77) before, at the peak, and during abatement of the COVID-19 pandemic. All values were z-standardized using the pre-pandemic mean and standard deviation. Error bars indicate 95% confidence intervals of between-subject estimates. Boxplots denote 1st, 2nd, and 3rd quartiles with whiskers extending 1.5 inter-quartile ranges or until the most extreme data point has been reached. Data points beyond the whiskers are plotted individually.

Figure 2

Risk factors exhibiting similar impact on the trajectory of psychological strain. Risk factors include high social anxiety (a,b), low self-efficacy (c), and high maladaptive emotion regulation strategies (d). All risk factors were associated with elevated baseline strain prior to pandemic onset but also with a less severe increase until pandemic peak. Nevertheless, people with elevated risk factors reported consistently greater strain across all time points. Risk factors were analyzed as continuous variables but are depicted as median splits for simplicity. Error bars indicate 95% confidence intervals of between-subject estimates. Boxplots denote 1st, 2nd, and 3rd quartiles with whiskers extending 1.5 inter-quartile ranges or until the most extreme data point has been reached. Data points beyond the whiskers are plotted individually.

Figure 3

Interactive effects of emotion regulation and gender on psychological strain. For adaptive emotion regulation strategies (a), men with higher values experienced a smaller rise in strain until pandemic peak but only if the time gap between first and second assessment was also comparably high (blue line in top right subplot). Acceptance (b) also only had a protective effect on males. Importantly, these gender effects need to be considered with caution due to unequal group size (230 females vs. only 77 males). Error bars indicate 95% confidence intervals of between-subject estimates. Boxplots denote 1st, 2nd, and 3rd quartiles with whiskers extending 1.5 inter-quartile ranges or until the most extreme data point has been reached. Data points beyond the whiskers are plotted individually.

Figure 4

Effect of childhood trauma on psychological strain. Childhood trauma severity was associated with greater baseline strain and less increase until pandemic peak. Notably, compared to other risk factors (see Fig. 2), psychological strain during pandemic peak was independent of childhood trauma. Error bars indicate 95% confidence intervals of between-subject estimates. Boxplots denote 1st, 2nd, and 3rd quartiles with whiskers extending 1.5 inter-quartile ranges or until the most extreme data point has been reached. Data points beyond the whiskers are plotted individually.

Figure 5

Overview of the acquired samples and analytical strategy. Sample 1 was assessed during the pre-pandemic baseline between June 2013 and March 2020 and a relative downturn of the COVID-19 pandemic in September and October 2021. Sample 2 was measured during the first peak of the pandemic in April 2020 and during the relative downturn in fall 2021. During all examinations, psychological strain was measured via a compound measure of the ASI-3, PSWQ, and STAI-T (cf. “Questionnaires” section). During the common measurement at the relative pandemic downturn, statistical twins were created (see Quasi-longitudinal Matching in “Data processing” section). The data of the twin from sample 2 was then used to impute the psychological strain during pandemic peak into the trajectory of the twin from sample 1, thus creating a quasi-longitudinal data set. For the main analysis, changes in psychological strain over time were analyzed relative to pre-pandemic moderators comprised of questionnaire sum scores for social anxiety (SPAI & LSAS), self-efficacy (GSE), emotion regulation strategies (CERQ), traumatic childhood experience (CTQ), and adverse life events (LTE & ALE).