This study employed a single-center, cross-sectional design. The study focused on patients who had experienced a following SARS-CoV-2 infection and sought care at an outpatient post-COVID-19 clinic after being referred by a general practitioner or medical specialist. Initially, referral criteria included patients with persistent fatigue, diminished fitness, and pulmonary symptoms such as exertional dyspnea, thoracic symptoms, or other symptoms, suspected of a previous SARS-CoV-2 infection. Subsequently, this was revised to include only those with a confirmed SARS-CoV-2 infection more than 3 months ago. Inclusion criteria for the study involved having a proficient command of the Dutch language to complete questionnaires and a willingness to complete the questionnaires. Exclusion criteria were age under 18, hospital admission following SARS-CoV-2 infection, and time since SARS-CoV-2 infection less than 3 months.

Data were collected during visits to the outpatient post-COVID-19 clinic at a prominent clinical teaching hospital in the Netherlands, serving approximately 280,000 inhabitants. After referral by the general practitioner or medical specialist, patients concurrently visited the pulmonologist and internist. Standardized cognitive and psychological questionnaires [Hospital Anxiety and Depression Scale (HADS), Post-traumatic Stress Symptoms Checklist 14 (PTSS-14), and Checklist for Cognitive and Emotional Consequences (CLCE-24)] were provided during or after consultation (Zigmond and Snaith, 1983; van Heugten et al., 2007; Twigg et al., 2008). Patients were requested to return completed questionnaires using an enclosed envelope to the Medical Psychology department. Patients not proficient in Dutch or unwilling to complete the questionnaires did not receive them. Following questionnaire return, a psychologist contacted patients, conveyed the results, and assessed the need for additional psychological treatment for coping with cognitive and/or emotional symptoms. Treatment was recommended for patients experiencing difficulties dealing with such symptoms, initiated or exacerbated after SARS-CoV-2 infection, impacting daily life functioning. Over time, clinical practice prompted the inclusion of the Four-Dimensional Symptom Questionnaire (4DSQ) and Cognitive Emotion Regulation Questionnaire (CERQ) to gain additional insight into physical symptoms and coping styles (Terluin, 1996; Garnefski et al., 2001).

Study data were retrieved from medical records and stored in a COVID-specific hospital database in Castor EDC (Castor, 2019). The study was conducted in accordance with the Medical Ethical Committee of Maastricht University Medical Center+, falling beyond the scope of the Medical Research Involving Human Subjects Act (WMO) (2021–3,059). The study was approved by the local institutional review board. Patients were informed about potential research use of their clinical data and had the option to opt out. Data collection occurred between July 1st 2020, and January 1st, 2023.

Demographic characteristics included age, sex, and time between SARS-CoV-2 infection and the post-COVID-19 clinic visit.

The HADS is a self-assessment tool assessing depression and anxiety symptoms (Zigmond and Snaith, 1983). It comprises anxiety (HADS-anxiety) and depression (HADS-depression) subscales, each with seven items. Responses, on a four-point Likert scale (0–3), yield subscale scores ranging from 0 to 21; higher scores indicate more symptoms of anxiety or depression. A cut-off score of ≥8 per subscale indicated clinically relevant symptoms of anxiety or depression (Herrmann, 1997).

The PTSS-14 is a screening questionnaire identifying patients at risk of post-traumatic stress disorder (PTSD) (Twigg et al., 2008). It features 14 items on a seven-point Likert scale (1–7), with scores ranging from 14 to 98; higher scores indicate more PTSS. Possible traumatic events were assessed by evaluating nightmares, anxiety and panic attacks, severe pain, and breathing difficulties and feelings of choking during the period of SARS-CoV-2 infection. The questions were slightly modified to specify the illness period of the SARS-CoV-2 infection. A cut-off score of ≥45 indicated high levels of PTSS (Twigg et al., 2008).

The CLCE-24 comprises 24 questions screening for cognitive and emotional symptoms, with 13 questions specifically assessing the absence or presence of cognitive symptoms (CLCE-cognition) and 1 item assessing the presence or absence of fatigue (CLCE-fatigue). Higher scores indicate more cognitive problems experienced in daily life (range 0–13) (van Heugten et al., 2007). We used a cutoff >2 based on mean of 1.9 (standard deviation = 1.9) in healthy controls (van Rijsbergen et al., 2015).

The 4DSQ measures the tendency to experience distress, somatization, depression, and anxiety (Terluin, 1996). For this study, only data from the somatization subscale, assessing physical symptoms over the past weeks, were used. The somatization subscale comprises 16 items (scores range 0–32), with scores between 11 and 20 indicating moderate physical symptoms and scores >20, used as cut-off in this study, indicating severe physical symptoms (Terluin et al., 2008).

The CERQ measures patients’ generally used cognitive coping strategies (Garnefski et al., 2002). It is a 36-item questionnaire featuring nine conceptually distinct subscales (Garnefski et al., 2001). Items are scored on a five-point Likert scale (1–5), with subscale scores ranging from 4 to 20. Higher scores indicated more frequent use of the coping strategy. Normative data was obtained from the CERQ manual (Garnefski et al., 2002).

The study population was described using means (SD) and medians (IQR). Numbers and percentages were reported for binary variables and variables with cut-off values. We presented both raw scores and z-scores of the CERQ.

Bivariate analyses, using Spearman correlation coefficients with Bonferroni correction for multiple comparisons (0.05/10 personal factors), were used to calculate associations between personal factors (i.e., psychiatric history and CERQ subscales) and cognitive and psychological outcomes. Four hierarchical multiple regression analyses were conducted to assess associations between demographic, physical, and personal factors and psychological and cognitive outcomes. Dependent variables were scores on HADS-anxiety, HADS-depression, PTSS-14, and CLCE-cognition. Age, sex, and time since SARS-CoV-2 infection were added as independent variables to the first block, and physical factors (4DSQ score) to the second block. Personal factors (CERQ subscales) found significant in bivariate analyses were added to the third block. Psychiatric history was omitted in the multivariate analyses due to missing data and lack of power. After applying the Bonferroni correction (0.05/4), the alpha level was set at 0.01 for multivariate analyses. Regression analysis assumptions were met for each model. IBM SPSS Statistics, version 26, was used for data analysis.

Figure 1 indicates that 610 patients were referred to the post-COVID-19 clinic during the study period. Of these patients, 310 (51%) returned the questionnaires, and 45 patients were excluded. In total, data from 265 patients were used in this study.

Table 1 presents the characteristics of the patients included in this study. Approximately 1 in 5 patients had a known psychiatric history (22%). One third of the patients (36%) were indicated for psychological treatment for their symptoms and visited the Medical Psychology department.

Table 2 outlines the scores on the physical and personal variables and outcome measures. Of the patients, 95% reported the presence of fatigue symptoms on the CLCE-fatigue. More than half of the patients (53%) who completed the 4DSQ reported scores above the cut-off. Approximately, between 30 and 40% scored above cut-off values on the HADS-anxiety, HADS-depression, and PTSS-14. Over three quarters of patients (79%) reported more than two cognitive symptoms on the CLCE-cognition. The most common cognitive symptoms were mental slowness (72%), difficulties in paying attention (74%), and remembering new information (75%).

Bivariate analyses revealed that six coping subscales were significantly associated with higher anxiety scores: a higher tendency toward self-blame (r = 0.320, p < 0.001), rumination (r = 0.484, p < 0.001), catastrophizing (r = 0.411, p < 0.001), and other-blame (r = 0.298, p < 0.001), and a lower tendency toward positive refocusing (r = −0.315, p < 0.001) and positive reappraisal (r = −0.228, p < 0.001). Multivariate analyses showed that demographic factors explained 5.2% of the variance in anxiety (step 1). Physical symptoms explained an additional 35.8% of the variance in anxiety (F change = 86.222, p < 0.001) (step 2). Personal factors (coping) explained an additional 15.2% of the variance in anxiety (F change = 7.862, p < 0.001). In the final model, more physical symptoms (β = 0.482, p < 0.001) and a higher tendency toward catastrophizing (β = 0.217, p < 0.01) were significantly associated with higher anxiety scores (R² = 0.562) (Table 3).

Bivariate analyses showed that psychiatric history and five coping subscales were significantly associated with higher depression scores: psychiatric history (r = 0.216, p < 0.001), a higher tendency toward self-blame (r = 0.246, p < 0.001), rumination (r = 0.441, p < 0.001), catastrophizing (r = 0.406, p < 0.001), and other-blame (r = 0.232, p < 0.001), and a lower tendency toward positive refocusing (r = −0.406, p < 0.001). Multivariate analyses demonstrated that demographic factors explained 6.8% of the variance in depression (step 1). Physical factors explained an additional 21.7% of the variance in depression (F change = 43.128, p < 0.001) (step 2). Personal factors (coping) explained an additional 16.9% of the variance in depression (F change = 8.477, p < 0.001) (step 3). In the final model, more physical symptoms (β = 0.367, p < 0.001) and a lower tendency toward positive refocusing (β = −0.325, p < 0.001) were the only independent factors significantly associated with higher depression scores (R² = 0.454) (Table 3).

Bivariate analyses showed that psychiatric history and four coping subscales were significantly associated with higher PTSS scores: psychiatric history (r = 0.211, p < 0.001), a higher tendency toward rumination (r = 0.481, p < 0.001), catastrophizing (r = 0.376, p < 0.001), and other-blame (r = 0.266, p < 0.01), and a lower tendency toward positive refocusing (r = −0.354, p < 0.001). Multivariate analyses demonstrated that demographic factors explained 6.4% of the variance in PTSS (step 1). Physical factors explained an additional 30.7% of the variance in PTSS (F change = 61.539, p < 0.001) (step 2). Personal factors (coping) explained an additional 12.8% of the variance in PTSS (F change = 7.779, p < 0.001) (step 3). In the final model, more physical symptoms (β = 0.428, p < 0.001) and a lower tendency toward positive refocusing (β = −0.290, p < 0.001) were the only independent variables significantly associated with higher PTSS scores (R² = 0.499) (Table 4).

According to bivariate analyses, psychiatric history (r = 0.210, p < 0.001), a higher tendency toward rumination (r = 0.230 p = 0.006), and a lower tendency toward positive refocusing (r = −0.267, p = 0.001) were significantly associated with more cognitive symptoms. Multivariate analyses showed that demographic factors explained 3.4% of the variance in cognitive symptoms (step 1). Physical factors explained an additional 21.8% of the variance in cognitive symptoms (F change = 40.463, p < 0.001) (step 2). Personal factors (coping) explained an additional 4.8% of the variance in cognitive symptoms (F change = 4.675, p = 0.011) (step 3). In the final model, more physical symptoms (β = 0.444, p < 0.001) and a lower tendency toward positive refocusing (β = −0.220, p = 0.003) were significantly associated with more cognitive symptoms (R² = 0.330) (Table 4).

One limitation is the use of regular clinical data, resulting in some missing variables. Missing data also occurred because not all variables were collected throughout the complete study period. No drop-out analysis was conducted because, in accordance with Dutch medical ethics regulations, if participants have not actively consented to the use of their data for research purposes—such as by not returning the questionnaires—this data may not be used or reported. Despite this limitation, the study included a large number of patients, ensuring sufficient power. Additionally, the absence of a specific questionnaire to assess COVID-19 symptomatology is acknowledged. However, many symptoms captured by the 4DSQ, including shortness of breath and muscle aches, overlap with physical symptoms post-COVID-19 (World Health Organization, 2022). Finally, we assessed cognitive symptoms using a questionnaire rather than cognitive testing. The data were collected as part of standard care. Cognitive testing has not been part of standard care for patients with post-COVID-19 syndrome.

A notable strength of this study lies in the inclusion of non-hospitalized patients facing challenges in their daily lives due to persistent post-COVID-19 symptoms, actively seeking additional care, thereby representing a substantial patient population. Within this cohort, we explored both physical and psychosocial risk factors—a distinctive approach that opens important avenues for future research and holds implications for clinical practice.

Future research on persistent cognitive and psychological symptoms post-COVID-19 should include psychosocial factors, such as coping in addition to biomedical factors. From a cognitive-behavioral perspective it can be hypothesized that both cognitive coping (thoughts and beliefs) and behavioral coping influence outcomes. Exploring whether adding behavioral coping enhances prediction models of persistent symptoms is interesting, as cognitive coping may precede behavioral coping, with catastrophizing potentially leading to avoidance behavior, decreased activity levels, and lower quality of life (Vlaeyen and Linton, 2000). Additionally, it would be interesting to assess not only symptoms as outcomes but also levels of participation. More specifically, inability to work is common in patients admitted to a rehabilitation clinic, with almost 50% of patients reporting at least 100 days of sick leave in the past year (Kupferschmitt et al., 2023). Inability to work can have a multifaceted impact on individuals, affecting their financial, physical, emotional, and social well-being. Identifying which symptoms, including cognitive, psychological, and physical symptoms, as well as personal and social factors, are associated with participation levels can provide important insights for treatment (World Health Organization, 2001).

The study emphasizes the need to shift from a purely biomedical to a more comprehensive biopsychosocial approach in clinical practice, focusing on alternative risk factors. This approach allows for more precise referrals to psychological treatment. While it is not possible or desirable for everyone to receive treatment, it is crucial that individuals experiencing symptoms disrupting daily life do receive treatment. At present, there is no evidence-based treatment for persistent symptoms post-COVID-19. Tailoring interventions to address individual needs and incorporating modifiable factors such as coping strategies, could be beneficial. Cognitive-behavioral therapies, including second or third-wave approaches such as cognitive-behavioral therapy and acceptance and commitment therapy, have proven effective in enhancing adaptive thinking patterns and behaviors. These interventions contribute to improving psychological flexibility and enhancing the quality of life in patients with chronic illnesses (Gregg et al., 2007; Hind et al., 2014; Feldmann et al., 2021). Such interventions merit consideration for addressing post-COVID-19 symptoms. Notably, for post-COVID-19 fatigue, cognitive-behavioral therapy has shown effectiveness compared to standard care, offering a potential intervention avenue (Kuut et al., 2023).