Depression, but Not Fatigue, Improves as Part of Multimodal Post-COVID Rehabilitation

Abstract

Introduction

Post-COVID syndrome (PCS) can lead to severe restrictions in the ability to work and participate in society and can lead to the development of depressive symptoms. The aim of this study was to investigate how many PCS patients suffer from clinically relevant depressive symptoms and to what extent PCS rehabilitation can modulate psychological symptoms as well as limitations in functional health.

Methods

Prospective multicenter cohort study PoCoRe (N = 1,028). Assessment of depressiveness (PHQ-9, PHQ-2), fatigue (FSMC) and functional health and disability (WHO-DAS2.0). Frequency analyses, mean comparisons.

Results

At the start of rehabilitation, 71.4% (n = 734) fulfilled the core symptoms of depressive disorders in general in the PHQ-2. In the screening-positive patients, the PHQ-9 total score was on average M = 23.74 (standard deviation = 4.43). As expected the depressive symptom burden was reduced more in depressed PCS patients (d = 0.50) than in nondepressed PCS patients (d = 0.32). Fatigue symptoms did not change. Functional health (WHO-DAS) improved above all in coping with everyday life (d = 0.71, or 1.04), and in the nondepressed patients also in mobility (d = 0.27). In the sociomedical assessment, 24.7% of the total samples were assessed as having an impaired capacity for the reference occupation and 19.5% as having an impaired capacity for the general labor market. Depressive patients had the worse sociomedical outcome.

Conclusion

Depressiveness is common and can be well influenced by PCS rehabilitation. Fatigue is less easily influenced, which indicates that they are independent symptoms. Patients with persistent depression and fatigue have a poorer sociomedical prognosis.

Introduction

The World Health Organization defines post-COVID syndrome (PCS) as the persistence or development of new symptoms 3 months after SARS-CoV-2 infection with symptoms persisting for at least 2 months without any other explanation [1]. The symptoms must cause relevant limitations in the ability to function in daily life. Although COVID-19 was originally classified as a respiratory disease, PCS symptoms can range from mild impairment to severe systemic disease. As the majority of patients show symptoms in multiple organ systems [2–4], it is now referred to as a multisystem disease. While the vast majority heal in the first 4 weeks after COVID-19, up to 400 million people worldwide are affected by PCS [5, 6]. Prevalence estimates for post-COVID syndrome range from around 5%–10% of infected people [7, 8]. Of these, around 10% are likely to require rehabilitation. The symptoms can range from fatigue or disproportionate exhaustion, tiredness, and breathing difficulties to neurological symptoms such as memory and concentration disorders. Studies suggest that the risk of developing PCS is higher in women, older people and people with a severe course (e.g., requiring intensive care) or preexisting conditions such as asthma and diabetes [9]. However, the majority of PCS sufferers had a mild course of disease. Other risks of developing PCS include behavioral and psychosocial risk factors, such as lack of exercise, depression or an anxiety disorder due to SARS-COV-2 infection [10–13].

In addition, post-COVID-19 syndrome is a growing problem in occupational medicine that affects the quality of life and productivity of employees as shown in a recent review by Gualano and colleagues [14] (2022). The percentage of PCS patients who were unable to return to work after 6 months ranged from 12.7% [15], over 22.2% [16] and up to 43.3% [17]. A relevant proportion of PCS patients suffer from a long-term burden of disease, which is reflected in their physical and social functioning, employment status, and persistent symptoms. Patients on long-term sick leave warrant sociomedical assessment regarding occupational performance. During rehabilitation, such an assessment is performed regarding both the reference occupation (last employment subject to social insurance contributions) and the general labor market. A work capacity of less than 3 h per day is referred to as reduced earning capacity and often leads to a disability pension if there is no further capacity for the general labor market.

The multitude of persistent and sometimes fluctuating symptoms impairs the ability to perform and function in daily life. Although there are numerous research efforts into the causes and disease mechanisms of PCS, there is still no curative therapy available that addresses the causal relationships. However, symptom-oriented treatment programs [18–21] have proven to be helpful [22–26], whereby the disease behavior (emotional and behavioral) seems to play a decisive role [10].

In Germany, inpatient rehabilitation (“stationäre Rehabilitation”) is a distinct healthcare sector that follows acute treatment and primarily aims to restore functional capacity and work ability [27]. Patients are admitted after stabilization of an acute illness or injury and typically present with neurological, musculoskeletal, cardiovascular, pulmonary, or psychosomatic disorders. This population is characterized by substantial functional limitations, temporary work disability, and frequent psychological distress related to adjustment and vocational concerns. Given the explicit linkage to social insurance systems, response tendencies in self-rating scales may be influenced by patients’ perceptions of health-related benefits and their vocational prospects. Attempts are usually made to use these existing structures and adapt the already well-evaluated treatment programs to the specific needs of post-COVID patients [28, 29]. PCS patients are referred to existing care structures (e.g., pneumology) based on the main symptoms (e.g., dyspnea). They are then treated in rehabilitation clinics that offer PCS rehabilitation and have a pneumological, neurological, cardiological or psychosomatic focus (cf. [30]). Dual rehabilitation is a well-established integrative treatment concept for patients with comorbid mental and physical illnesses (e.g., heart attack and heart-related anxiety) [31, 32]. Post-COVID patients also suffer from pronounced physical symptoms that cause considerable psychological suffering, but cannot yet be treated causally. In order to treat both somatic and psychological factors, the combination of a thorough medical examination and care, psychotherapy and a well-dosed, comprehensive exercise program is the most promising treatment option. In the meantime, concepts adapted to the individual needs of PCS patients, such as multimodal PCS rehabilitation, have been established [33–39]. This rehabilitation concept includes a detailed medical examination and care, a clinical-psychological assessment and treatment as well as a carefully designed, comprehensive exercise program and aims to address both the physical and psychological symptoms of PCS patients and improve their participation in everyday life and work.

Alongside pain and breathing difficulties, fatigue and cognitive disorders are among the most common symptoms of PCS. In diseases that are primarily somatic in origin (e.g., cardiological or orthopedic diseases, chronic pain), psychological comorbidity such as depression usually plays an important role in symptom maintenance, quality of life and functional limitations [39, 40]. However, with regard to the measurement of depressive symptoms in PCS using established screening procedures (e.g., PHQ-9), it should be critically noted that some of the depressive symptoms overlap with physical PCS symptoms. Sleep disturbances, fatigue, reduced drive, and concentration problems can be symptoms of both depression and PCS. As a result, misjudgments can easily occur when PCS patients score positively on common screening instruments due to the high somatic component.

In addition to the PHQ-9, which covers a large proportion of somatic symptoms, the PHQ-2 focuses exclusively on the affective symptoms of major depression: loss of interest and depressed mood. Both screening instruments are used here to differentiate the proportion of PCS patients who actually have depressive symptoms and how many patients may have been misdiagnosed as depressed.

In order to contribute to the state of knowledge about rehabilitation options for post-COVID syndrome, this study addresses the following research questions:

• 1.How many of the PCS patients fulfill the affective depression items in the PHQ-2 (= depressed PCS patients)?
• 2.Are PCS patients depressed to different degrees depending on the indication (dual rehabilitation, psychosomatics, neurology, pneumology)?
• 3.How do depressive symptoms and fatigue develop over the course of PCS rehabilitation?
• 4.How highly do PHQ-2 and PHQ-9 correlate with fatigue?
• 5.How does the functional health of depressed and nondepressed PCS patients develop over the course of PCS rehabilitation?
• 6.Do depressive and non-depressive PCS patients differ in the sociomedical assessment?

Methods

Procedure and Sample

The above questions were addressed as part of the PoCoRe multicenter study [30]. The prospective cohort study is based on standardized assessments of psychological distress (PHQ-9 as a measure of depression), fatigue (FSMC), and functional health and disability (WHODAS 2.0) at the beginning and end of rehabilitation. Screening instruments cannot establish formal diagnoses but do provide valid indications of clinically relevant impairments. This is standard and acceptable practice in both research and clinical settings, and allows meaningful comparison across cohorts. This study utilizes real-world data and was conducted without a control group. Such a design is particularly appropriate and informative in the context of post-COVID research, where the heterogeneity of patient trajectories, variations in healthcare access, and ethical constraints often preclude the feasibility of randomized controlled trials. By analyzing routinely collected clinical and administrative data, real-world data studies provide valuable insights into the long-term outcomes, treatment patterns, and functional recovery of patients in naturalistic settings. This approach allows for the characterization of post-COVID conditions as they manifest in everyday clinical practice, thereby complementing evidence from controlled studies and supporting the development of effective, data-driven healthcare strategies for this emerging patient population. The data collection for this sample took place between March 2022 and August 2023 in six German rehabilitation clinics that have implemented a post-COVID concept: Dual rehabilitation (psychopneumology, psychocardiology), psychosomatics, neurology, and pneumology. The participating six intervention centers agreed to offer a multimodal rehabilitation program including individual and group psychotherapy (2 × 50 min per week CBT), a neuropsychological assessment (TAP, MoCa), cognitive training in a group or individual setting if indicated, individualized endurance training, post-COVID-specific exercise therapy, which has antidepressant effects [41], body awareness exercises, breathing therapy, relaxation techniques and social counseling, among other things. The qualification of psychotherapist is with few exceptions in behavioral therapy. The combination of the different interventions, embedded in the social community of the therapy groups, evokes behavioral activation [42] and well-being (well-being; [43]). However, like in similar real-world studies, individual cases show some variability for specific patients, or institutions. For more detailed information, refer to online Supplement 1 (for all online suppl. material, see https://doi.org/10.1159/000549848), which specifies how many units of which therapy are used per week and at what intensity. As of today, we did not publish standardized therapy manual. Psychotherapy was carried out by cognitive-behavioral psychotherapists (medical and psychological).

Of a total of 1,176 consecutively enrolled PCS patients, 1,028 were included in the study. After taking dropouts into account, the total sample consisted of 1,028 patients. A medically confirmed post-COVID diagnosis was a prerequisite for approval of post-COVID rehabilitation. This usually included a positive PCR test, the corresponding PCS symptoms as well as somatic clarification as a differential diagnosis and exclusion of other diseases that could explain the symptoms (e.g., cardiological, pneumological, neurological diagnostics). The diagnosis was made in accordance with the WHO definition. Eligibility criteria were [27]:

• SARS-CoV-2 infection and following post-COVID syndrome: complaints that are still present more than 12 weeks after the onset of SARS-CoV-2 infection and cannot be explained otherwise.

• As a consequence of the post-COVID syndrome, at the time of the start of rehabilitation, the presence of functional limitations that may threaten the ability to work.

• Written informed consent to participate in the study.

• Aged 18 years or above.

• Sufficient knowledge of the German language to participate in the study.

The measures used in this study were used to determine psychopathology and complaints. The following measurement instruments were used at the start and end of rehabilitation: PHQ-9, PHQ-2, FSMC, WHODAS 2.0. Diagnoses cannot be made using questionnaires, but they are suitable for assessing symptoms, symptom severity, and symptom progression. The Ethics Committee of the Center for Clinical Research of the University Hospital Regensburg (ID 22-2814-101) granted approval for the research in February 2022.

Measuring Instruments

Psychological Stress/Depression (PHQ-9, German Version)

The PHQ-9 questionnaire [44] is a module of the patient health questionnaire and records one of the diagnostic criteria for major depression according to DSM-IV with a total of 9 items. The response scale has a range from 1 to 4 (1 = not at all, 4 = almost every day). Scores are added to obtain a total score, which can be interpreted as follows: 10–14 mild depressive symptoms, 15–19 moderate depression, 20–27 major depressive symptoms. The PHQ-2 is an ultra-short form of the complete PHQ-9 depression module. It focuses on loss of interest and depressed mood as main criteria of depressive disorders in general according to the DSM-IV. Due to the lack of somatic items such as fatigue or loss of drive, it was considered less confounded by somatic aspects of depression and selected as screening instrument for the post-COVID sample. The cut-off for a positive screening result is ≥3 [45]. Administration of the PHQ-2 followed by the PHQ-9 may provide acceptable accuracy for a depression screening [46]. In this study, PCS patients who screened positive on the PHQ-2 were considered “depressed.”

Fatigue (FSMC)

The Motor and Cognitive Fatigue Scale is a self-assessment questionnaire [47] used to assess cognitive and motor fatigue symptoms. The questionnaire contains 20 items, which can be summed up to a cognitive scale and a motor scale. Ratings are given on a 5-point Likert scale. Fatigues are considered as pathological for scores ≥43 on the total scale and ≥22 on the cognitive and motor scales, respectively.

Functional Health and Disability (WHODAS 2.0)

The WHO Disability Assessment Schedule (WHODAS 2.0) [48] records health, functioning, and disability regardless of the underlying disease. With 36 questions, the questionnaire covers the domains of understanding and communication, cognition, mobility (movement, mobility), self-care (personal hygiene, dressing, eating, self-care), dealing with other people (social interaction), activities of daily living (household, leisure, work, education), and participation in social life.

The domains of the WHODAS 2.0 are based on the conceptual framework of the WHO’s International Classification of Functioning, Disability and Health (ICF). It also uses a 5-point Likert scale (0 = none, 5 = extreme).

Work Related/Sociomedical Parameters

The sociomedical parameters relevant at admission and discharge were routinely recorded as part of rehabilitation. People who were able to work at least 6 h a day were considered able to work full time. Performance assessment was carried out separately for the reference occupation (last job subject to social insurance contributions) and the general labor market. An earning capacity of less than 3 h per day is referred to as reduced earning capacity and often leads to a disability pension, if not exceeded by work ability on the general labor market. Assessment of work ability at the time of discharge as well as performance assessment was carried out by the rehabilitation team under the supervision of specialist physicians trained in social medicine.

Statistical Analysis

The statistical analysis was conducted using the statistics program SPSS 28 for Windows. Descriptive statistics (cross-tabulations) were calculated to describe the sample, including an evaluation of previous illnesses and work-related parameters. In a first step, mental stress was evaluated separately for the four indications of dual rehabilitation (psychopneumology, psychocardiology), psychosomatics, neurology, and pneumology. In a further step, the proportion of depressed and nondepressed PCS patients in the overall sample was determined using DSM criteria for major depression (presence of affective items) as classification criterion. Subsequently, depressed and nondepressed patients were examined in more detail. Statistical analyses of the Likert-scaled items and scores included descriptive measures in the form of mean values (M) and dispersion measures (standard deviation [SD]). Group comparisons were calculated using ANOVA. The significance level was set at α = 0.05 on both sides. Multivariate regressions were performed to evaluate the influence of age, gender, course of acute COVID-19 disease and preexisting illnesses on changes in depressive symptoms (PHQ-9), fatigue (FSMC), and general coping with everyday life (WHODAS). In addition, correlation analyses (Pearson) were calculated between PHQ-9, PHQ-2, and FSMC in order to examine the extent to which the PHQ-9 is confounded by somatic items and is therefore less suitable for detecting depression in PCS than the PHQ-2, which is based exclusively on affective items.

Results

Sample Description

In the period from March 2022 to August 2023, a total of 1,028 PCS patients were included: n = 190 dual rehabilitation (psychopneumology, psychocardiology), n = 210 psychosomatics, n = 493 neurology and n = 135 pneumology. Patients who had applied for and been admitted to post-COVID rehabilitation based on a medically confirmed post-COVID diagnosis were included. The gender distribution was consistent with distributions reported in previous studies, with women being predominantly affected by PCS. In this study, 71.8% of PCS patients were female. The mean age in the overall sample was 48.30 years (SD = 10.92), which is slightly younger than the typical age for a rehabilitation population (Rehab Report 2022). During the primary SARS-COV-2 infection, 10.6% of PCS patients were hospitalized; most PCS patients remained at home during acute infection. In the total sample, more than half of PCS patients (59.1%) were unable to work upon admission and 77.8% had already been on sick leave for more than 6 months.

Occupationally, 47.0% of PCS patients were mentally active and 46.1% were both physically and mentally active. Regarding the question of whether patients envisioned themselves returning to work in the future, around 1/4 (24.3%) stated that this was unlikely, 48.0% were not sure and only 25.5% were fairly certain that they would return to work. Descriptive statistics of the subgroups can be found in Tables 1 and 2.

Table 1.

Descriptive statistics

​	Depressed PCS patients	Nondepressed PCS patients
	n = 734 (71.4%)	n = 294 (28.6%)
Demographic characteristics
Gender (% female)	72.1	71.6
Age, years	48.52 (10.58)	49.16 (10.78)
Body weight ♀, kg	77.49 (20.72)	80.36 (52.32)
Body height ♀, cm	167.05 (6.93)	166.93 (8.95)
Hospitalization during acute SARS-COV-2 infection, %	8.3	8.6
Preexisting conditions before SARS-COV-2 infection, %
Heart disease (CHD, heart attack, high blood pressure, arteriosclerosis)	32.0a	21.3a
Pneumological diseases (COPD, asthma)	23.7a	18.0a
Sickle cell anemia, iron deficiency anemia	5.8	4.4
Blood diseases (hemophilia, thrombophilia)	3.3	2.8
Cancer (colon CA, breast CA, prostate CA)	5.7a	2.4a
Chronic intestinal diseases (Chron’s disease, ulcerative colitis)	7.2	5.7
Thyroid or adrenal gland disease	25.9a	20.2a
Diabetes mellitus or hyperlipidemia	10.1a	6.5a
Neurological diseases (Parkinson’s disease, multiple sclerosis)	5.7a	4.0a
Anxiety disorders or depressive disorders	41.4b	14.8b
Occupational characteristics, %
Ability or inability to work on admission	62.8; unable to workb	50.0; unable to workb
	37.2; able to workb	50.0; able to workb
Periods of incapacity for work before the start of rehabilitation	13.3; <3 monthsa	15.1; <3 monthsa
	12.3; 3–6 monthsa	10.2; 3–6 monthsa
	73.8; >6 monthsa	76.1; >6 monthsa

^aSignificant differences p < 0.05. ^bSignificant differences p < 0.001.

Table 2.

Self-assessment of occupational parameters

​	Depressed PCS patients	Nondepressed PCS patients
	n = 734 (71.4%)	n = 294 (28.6%)
Type of work/employment, %
Mainly mentally active	47.6	45.8
Equally mentally and physically active	44.8	47.9
Mainly physically active	4.5	6.3
Self-assessment on admission: able to work again, %
Unlikely	27.8	21.0
Not sure	49.3	48.5
Fairly certain	22.8	29.7
Physical ability to work	1.9; very gooda	2.0; very gooda
	9.2; rather gooda	17.5; rather gooda
	25.8; mediocrea	33.7; mediocrea
	32.9; rather poora	23.4; rather poora
	30.3; very poora	22.6; very poora
Cognitive working ability	0.2; very gooda	2.4; very gooda
	4.0; rather gooda	19.6; rather gooda
	21.6; mediocrea	30.6; mediocre
	34.9; rather poora	25.1; rather poor
	39.3; very poora	21.6; very poor
Self-assessment on discharge, %
No impairment	0.9	3.9
I can do my work, but I have complaints	16.7	22.7
I am sometimes forced to work more slowly or change my working method	18.6	19.6
I am often forced to slow down or change my working method	28.2	23.9
Due to my illness I am only able to work part-time	21.1	19.5
In my opinion, I am completely unable to work	44.3	30.7
Physical ability to work	13.2; very gooda	21.3; very gooda
	14.8; rather good	14.8; rather good
	17.9; mediocre	21.3; mediocre
	26.4; rather poor	26.2; rather poor
	28.3; very bada	13.1; very bada
Cognitive work ability	14.2; very good	13.0; very good
	8.3 rather good	15.9; rather good
	12.5; mediocre	14.5; mediocre
	35.0; rather poor	34.8; rather poor
	30.0; very poora	18.8; very poora

^aSignificant differences.

Psychological Stress: Depressive Symptoms

At the start of rehabilitation, 71.4% (n = 734) of PCS patients reported loss of interest and depressed mood, the core symptoms of depressive disorders in general, in the PHQ-2. Symptom burden did not differ between indications for rehabilitation. In all indications, over 65% of the core symptoms of depressive disorders in general were met in the PHQ-2 (see Table 3 for results). The study of PCS patients who met the criteria for depressive disorders (depressive sample) was able to reduce their depressive symptom burden over the course of rehabilitation with predominantly moderate effects (see Table 4). The significant improvement in fatigue in the corresponding PHQ-9 item is striking. The nondepressed PCS patients were also able to achieve an improvement in depressive symptoms over the course of rehab, but only with small effect sizes.

Table 3.

Depressiveness in the respective indications

​	Total sample	Dual rehabilitation	Psychosomatics	Neurology	Pneumology
	N = 1,028	n = 190	n = 210	n = 493	n = 135
	M (SD)	M (SD)	M (SD)	M (SD)	M (SD)
PHQ-9 total	22.15 (5.34)	22.32 (5.32)	24.29 (5.47)	21.30 (5.09)	21.77 (5.22)
PHQ-2	4.75 (1.83)	4.76 (1.75)	5.51 (1.81)	4.40 (1.80)	4.74 (1.76)
PHQ-2% MJ	71.4	69.2	85.9	65.2	86.9

M, mean; SD, standard deviation; MJ, major depression.

Table 4.

Symptom course: differences within the depressed and nondepressed sample from admission to discharge

​	Depressed PCS patients, n = 734 (71.4%)						Nondepressed PCS patients, n = 294 (28.6%)
	admission	discharge	p value	d	MWD	CL	admission	discharge	p value	d	MWD	CL
	M (SD)	M (SD)					M (SD)	M (SD)
PHQ-9 (scale 1–4)
Loss of joy or interest	2.91 (0.80)	2.43 (0.66)	<0.001	0.64	−0.48	0.39 to 0.57	1.51 (0.51)	1.34 (0.48)	<0.001	0.35	−0.19	0.09 to 0.25
Dejection/melancholy/hopelessness	2.69 (0.84)	2.39 (0.69)	<0.001	0.37	−0.30	0.20 to 0.39	1.14 (0.36)	1.16 (0.38)	0.381	–	0.02	−0.08 to 0.03
Difficulty falling asleep/staying asleep	3.46 (0.78)	3.23 (0.84)	<0.001	0.28	−0.23	0.13 to 0.32	2.81 (1.06)	2.61 (1.07)	0.023	0.18	−0.20	0.02 to 0.37
Tiredness	3.64 (0.63)	3.19 (0.82)	<0.001	0.62	−0.45	0.35 to 0.53	3.13 (0.96)	2.61 (1.01)	<0.001	0.52	−0.52	0.35 to 0.67
Loss of appetite	2.27 (1.06)	2.01 (0.94)	<0.001	0.54	−0.26	0.13 to 0.37	1.55 (0.83)	1.36 (0.63)	0.001	0.26	−0.19	0.07 to 0.31
Self-deprecation and self-doubt	2.24 (1.04)	1.98 (0.93)	<0.001	0.24	−0.26	0.13 to 0.38	1.17 (0.44)	1.16 (0.45)	0.861	–	−0.01	−0.06 to 0.08
Concentration problems	3.26 (0.84)	3.04 (0.89)	<0.001	0.26	−0.22	0.11 to 0.33	2.76 (1.07)	2.37 (0.99)	<0.001	0.38	−0.40	0.22 to 0.56
Psychomotor changes: restlessness or retardation	2.33 (1.03)	2.08 (0.94)	<0.001	0.24	−0.25	0.10 to 0.39	1.83 (1.03)	1.55 (0.77)	0.001	0.31	−0.28	0.11 to 0.45
Suicidal thoughts	1.30 (0.59)	1.19 (0.51)	<0.001	0.18	−0.11	0.03 to 0.17	1.00 (0.13)	1.00 (0.12)	0.932	–	–	−0.02 to 0.02
PHQ-9 total value	24.18 (4.35)	21.47 (4.31)	<0.001	0.62	−2.71	2.06 to 3.35	16.78 (3.75)	15.23 (3.74)	<0.001	0.41	−1.55	0.81 to 2.29
FSMC Fatigue (5-point Likert Scale, max 50 points per subscale and 100 points for the overall scale
Cognitive fatigue	40.74 (7.01)	40.17 (7.59)	0.2.08	–	−0.57	−0.69 to 2.61	34.79 (9.99)	33.98 (10.48)	0.344	–	−0.81	−0.87 to 2.49
Motoric fatigue	40.09 (6.49)	39.63 (7.04	0.277	–	−0.46	−0.32 to 1.46	34.34 (9.22)	33.20 (9.76)	0.154	–	−1.14	−0.43 to 2.71
FSMC total value	80.77 (12.70)	79.81 (14.12)	0.256	–	−0.96	−0.36 to 1.28	60.09 (18.46)	67.21 (19.53)	0.242	–	7.12	−1.27 to 5.04
WHODAS 2.0 Limitations (5-point Likert Scale)
Limitations mobility	2.22 (1.14)	2.13 (1.07)	0.551	–	−0.09	−0.11 to 0.20	2.31 (1.16)	1.28 (0.66)	0.085	–	−0.21	−0.02 to 0.43
Limitations on understanding/communication	3.28 (0.85)	3.03 (0.93)	0.070	–	−0.25	−0.02 to 0.52	1.49 (0.87)	1.83 (1.900	0.036	0.44	−0.48	0.03 to 0.93
Limitations on self-sufficiency	1.23 (0.59)	1.29 (0.62)	0.278	–	−0.06	−0.16 to 0.04	1.04 (0.27)	1.00 (0.13)	0.065	–	−0.04	0.00 to 0.08
Limitations on dealing with people	2.21 (1.02)	2.07 (1.10)	0.335	–	−0.13	−0.14 to 0.42	1.31 (0.67)	1.12 (0.45)	0.010	0.34	−0.19	0.04 to 0.33
Limitations in everyday life	3.81 (0.91)	2.97 (1.44)	<0.001	0.71	−0.84	0.45 to 1.22	3.03 (1.40)	1.49 (1.02)	<0.001	1.30	−1.54	1.03 to 2.04
Limitations in social life	3.50 (0.78)	3.17 (0.76)	0.881	–	−0.32	−0.28 to 0.24	2.38 (1.05)	1.93 (0.93)	0.040	0.45	−0.45	−0.5 to 0.94

M, mean; SD, standard deviation. Cohen-d-effect size, 0.2 small (light green), 0.5 medium (mid green), 0.8 big (dark green).

Fatigue

The FSMC total score as a measure of fatigue was in the severe range at the start of rehabilitation in both the depressed and nondepressed PCS patients. The symptoms changed in the nondepressed PCS patients (small effect), but not in the depressed PCS patients (Table 4). At the start and end of rehabilitation, the depressed subsample had a significantly (p = 0.045) higher overall fatigue score (start of rehabilitation: M = 80.77, SD = 12.70; end of rehabilitation: M = 79.81, SD = 14.12) than the nondepressed subsample (start of rehabilitation: M = 60.09, SD = 18.46; end of rehabilitation: M = 67.21, SD = 19.53).

Multivariate Regression

Multivariate regressions were performed to evaluate the influence of age, gender, course of acute COVID-19 disease and preexisting illnesses (diabetes, psyche, cardiovascular, neurological, autoimmune, pneumological) on changes in depressive symptoms (PHQ-9), fatigue (FSMC) and general coping with everyday life (WHODAS). For the model predicting the change in depression, R² = 0.33 (corrected R² = 0.24), F(10,114) = 4.894, p < 0.001; here only diabetes (β = 0.188 p = 0.035) was significant. The assignment according to PHQ-2 had the greatest influence (β =0.384, p < 0.001). VIF was <10.

For the change in fatigue, there was an R² = 0.094 (corrected R² = 0.054), F(10,225) = 2.333, p < 0.012. In addition to the classification depressed versus nondepressed according to PHQ-2 (β =0.160, p = 0.023), previous neurological disorders (β = −0.169, p = 0.011), and mental disorders (β =0.180, p = 0.013) also appeared to be relevant here. VIF was <10.

The model for the influence on the change in general coping with everyday life was not significant. Neither age, gender, course of the acute COVID-19 illness, preexisting illnesses nor the classification according to PHQ-2 seemed to play a role here (R² = 0.23 [corrected R² = 0.10], F(10,63) = 1.888, p = 0.064).

Correlation Analyses of Depression and Fatigue

Correlation analyses revealed high intercorrelations between depression as assessed by PHQ-9 and PHQ-2 and the FSMC total with its subscales cognitive and motor fatigue, respectively (Table 5). However, fatigue (FSMC total) is more strongly correlated with its subscales (r = 0.957, p < 0.001; r = 0.946, p < 0.001) than depression (PHQ-9) with the core symptoms only (PHQ-2) (r = 0.807, p < 0.001). The PHQ-9, which also contains somatic items, is more strongly correlated with fatigue than the PHQ-2, which only contains affective items, suggesting confounding of the PHQ-9 with somatic factors.

Table 5.

Pearson correlations of depression screening with fatigue

​	PHQ-9	PHQ-2	FSMC total	FSMC cognitive	FSMC motoric
PHQ-9
Pearson correlation	1	0.807a	0.533a	0.488a	0.503a
Significance (2-sided)	​	<0.001	<0.001	<0.001	<0.001
n	1,028	920	904	916	916
PHQ-2
Pearson correlation	0.807a	1	0.403a	0.379a	0.395a
Significance (2-sided)	<0.001	​	<0.001	<0.001	<0.001
n	920	920	859	74	868
FSMC total
Pearson correlation	0.533a	0.403a	1	0.957a	0.946a
Significance (2-sided)	<0.001	<0.001	​	<0.001	<0.001
n	904	859	904	887	889
FSMC cognitive
Pearson correlation	0.488a	0.379a	0.957a	1	0.811a
Significance (2-sided)	<0.001	<0.001	<0.001	​	<0.001
n	916	874	887	916	884
FSMC motoric
Pearson correlation	0.503a	0.395a	0.946a	0.811a	1
Significance (2-sided)	<0.001	<0.001	<0.001	<0.001	​
n	916	868	889	884	916

^aThe correlation is significant at the level of 0.01 (2-sided).

Functional Health and Disability (WHODAS 2.0)

Results on the WHODAS 2.0 as a measure of functional health and disability show that depressed PCS patients are significantly more burdened in all areas than nondepressed PCS patients at both admission and discharge. They were also much more stressed than people in the general population [49]. In depressed PCS patients, limitations in everyday life improved with a medium effect size (d = 0.71), nondepressed patients experienced a strong symptom improvement (d = 1.30). The lack of effect on mobility is consistent with the results in the FSMC (Table 4). Evaluations of the WHODAS 2.0 at individual item level can be found in online Supplement 2.

Sociomedical Assessment and Occupational Self-Assessment

At discharge, 58.6% of PCS patients were considered not fit to work as of their sociomedical evaluation. It was striking that depressed PCS patients were discharged unfit to work significantly more often (64.9%) than nondepressed PCS patients (47.1% unable to work). This poorer outcome was also reflected in the prospective sociomedical assessment. Overall, 22.3% were assessed as being unable to work in their reference occupation and 18.4% as being unable to work in the general labor market (see Table 6). Also, depressed PCS patients reported lower expectations on a self-assessment measure of work ability (Table 2).

Table 6.

Sociomedical assessment on discharge

​	Depressed PCS patients	Nondepressed PCS patients	p value	Phi/Cramer’s V
	n = 734 (71.4%)	n = 294 (28.6%)
Ability (or inability) to work on discharge, %	64.9; unable to worka	47.1; unable to worka	<0.001	0.24
	34.5; able to worka	51.9; able to worka	<0.001	0.24
Sociomedical assessment occupation, %	26.6; <3a	16.1; <3a	<0.001	0.15
	13.3; 3–6a	10.0; 3–6a	<0.001	0.15
	59.9; >6a	73.1; >6a	<0.001	0.15
General labor market, %	20.4; <3a	15.8; <3a	<0.001	0.14
	8.7; 3–6a	6.0; 3–6a	<0.001	0.14
	70.7; >6a	77.2; >6a	<0.001	0.14

<3 = able to work for less than 3 h a day, 3–6 = able to work, 3–6 h per day, >6 = able to work more than 6 h a day.

^aSignificant differences; 0.1 ≤ φ/V <0.3 small, φ/V <0.5 medium, φ/V <0.7 strong.

Discussion

Our data showed that over 70% of PCS rehabilitation patients met the core symptoms of depressive disorders in general – loss of interest and depressed mood. All patients answered the PHQ-9. However, as PHQ-9 scores may be influenced be existing chronic physical conditions, we limited some of our analysis to the affective components covered by two items, representing the key symptoms of depressive disorders “lack of interest, depressed mood” as it has be recommended earlier [46]. In the exclusively somatic indications (neurology, pneumology) of PCS rehabilitation, the symptom burden was somewhat lower, but also here over 60% of PCS patients met the core symptoms of depressive symptoms. In other studies, which examined patients with a recent onset of PCS, i.e., approx. 12 weeks after infection, the proportion of clinically relevant depressed PCS patients was approx. 12% [50]. The clear discrepancy to the numbers presented in this study is likely due to the time aspect. Patients examined here have usually been suffering from severe symptoms for well over 12 months, limiting them in their everyday lives and participation in society. These patients are chronically affected and usually face pressing financial issues due to frequent long periods of inability to work. In addition, there are various other stressors, such as a change in roles in the partnership and in the entire social environment, so that depression may have developed as a reaction to illness. Presumably, the Renaud-Charest et al. [50] (2021) sample had not yet become chronic, and the prognosis was therefore significantly better. The results underline the relevance of depressive symptoms in the chronification of PCS; this is also consistent with other studies [51]. In this sample, depressed PCS patients were significantly more likely to be on sick leave at the start of rehabilitation and had longer periods of incapacity to work than nondepressed PCS patients.

Rehabilitation had a clear effect on depressive symptoms, while fatigue only improved in nondepressed PCS patients. The discrepancy between the clear improvement in “tiredness” in the PHQ-9 and the very slight improvement in fatigue symptoms (FSMC) can be explained by the fact that the tiredness and fatigue recorded in the PHQ-9 are different concepts and items formulated in very different ways. The PHQ-9 only features one item on “tiredness,” whereas the FSMC is a comprehensive questionnaire on cognitive and motor fatigue (see PHQ-9 and FSMC questionnaire). Comparing depressed PCS patients and nondepressed PCS patients in terms of fatigue severity and fatigue progression, it becomes apparent that depressed PCS patients suffer from significantly more severe fatigue symptoms at admission and discharge and do not improve over the course of rehabilitation; this is despite the fact that depressive symptoms improves, which speaks against a simple, direct, and linear correlation. The correlation analyses between the PHQ-9, PHQ-2, and the FSMC indicated that the PHQ-9 is confounded by somatic items and is therefore less suitable for depicting depression in PCS than the PHQ-2, which is based exclusively on affective items. When using the PHQ-2, which omits the somatic items of the PHQ-9, the distinction between depressed and nondepressed PCS patients is clearer. This is because the PHQ-9 is more strongly confounded by fatigue. However, the difference between the PHQ-9 and PHQ-2 is rather small, so either instrument can be used. The PHQ-9 has the advantage of including a suicidality item, which covers a clinically relevant aspect of depressive psychopathology.

In nondepressed PCS patients, both fatigue and mobility improves and there is also a clear effect on everyday function. Overall, nondepressed PCS patients are less impaired in the WHODAS 2.0 than depressed PCS patients and are also more likely to show an improvement in functional health. This may indicate that PCS patients without depressive symptoms presumably secondary to PCS disease respond better to treatment. Acceptance and commitment therapy could be a therapeutic approach to promote disease acceptance in cases of severe chronicity and thus alleviate depression. Overall, PCS rehabilitation has a positive effect on the mental and functional health of PCS patients. This is also consistent with the results of other studies that show a positive effect of PCS rehabilitation [23, 24, 37, 52–55]. PCS rehabilitation contributes to increased participation and better return to work rates, even despite persistent symptoms [56].

In our sample, around 60% of PCS patients were unable to work at the start of rehabilitation, more than 3/4 of them for more than 6 months. About half of PCS patients were not sure whether they would return to work in the future; about 1/4 stated that they were currently completely unable to work due to their illness. Overall, 24.7% of PCS patients were assessed as unable to work in the reference occupation and 19.5% in the general labor market; this is 10% more than for other diagnoses with the same rehabilitation indication. Data from the German Pension Insurance 2022 showed that 15.9% of non-PCS patients were unable to work in the reference occupation and 5.3% were unable to work in the general labor market.

In our study, depressed PCS patients had a ∼10% worse sociomedical prognosis compared to nondepressed PCS patients. In addition to the generally higher symptom burden, the less favorable sociomedical assessment of depressed PCS patients could be explained primarily by the high fatigue values and the lack of improvement in fatigue symptoms.

Our study shows that PCS rehabilitation can contribute to an improvement in mental well-being and functional health. At the same time, however, our data also show that a relevant proportion of PCS patients feel that they are no longer able to cope with their professional future, at least in the intermediate term. This study indicates that a distinction between depressed and nondepressed PCS patients may be relevant, as depression appears to have an influence on the severity of fatigue and the sociomedical prognosis. This is also consistent with the results of other studies [57]. Depressed PCS patients do not experience any improvement in their fatigue symptoms, have a higher overall symptom load and are also assessed less favorably from a sociomedical perspective. Our findings can be interpreted as cautious indications that depression is secondary to PCS symptoms in a large number of cases, which can be seen as a risk factor for chronification. Because our sample consists of patients who are still of working age and well away from regular retirement, this result increases the significance of our findings. According to the estimates of the social physicians in the PoCoRe study, this risk exists for approx. 25% of PCS rehabilitants who leave working life prematurely despite rehabilitative measures, which is of substantial importance for the labor market and the social support systems.

Strengths and Limitations

This study is based on a large sample of 1,028 examined PCS patients in specific PCS rehabilitation settings. To our knowledge, this study was the first to show that most PCS patients, regardless of whether they are assigned to a primarily somatic indication (neurology, pneumology) or psychosomatics, suffer from clinically relevant depressive symptoms. As this is a multicenter study that includes different indications, it can be assumed that the data collected is representative of PCS rehabilitation in Germany. However, the data presented here cannot be seen as representative of PCS patients overall. This study included moderately severely affected PCS patients who have been suffering from their symptoms for a long time and can therefore be considered chronically ill. In the general population, however, there are still many mildly affected people (no need for rehabilitation) and severely affected people with ME/CFS (no need for rehabilitation) who were not included here [58].

Further limitations include the study design as prospective cohort study. Due to the lack of a control group, a randomized controlled study was not feasible, prohibiting any causal statements on the effectiveness of rehabilitation on PCS symptoms. In addition, depression was only assessed using questionnaires and not via clinical rating or structured interview. This should be done in future studies. Nevertheless, our results are of scientific relevance, especially as they are in line with previous research findings on the effectiveness of PCS rehabilitation [25, 26, 36, 59]. The present study also lacks a control group without intervention, so that time effects in the sense of natural regeneration and symptom reduction cannot be investigated. At the same time, spontaneous remission seems unlikely due to the long period of chronicity. For ethical and legal reasons, withholding rehabilitative treatment is indeed prohibited as there is a right to rehabilitation in Germany. Although the data were collected as part of a separate diagnostic procedure for this study, a waiting list control group was not feasible in everyday clinical practice.

Conclusion

The results of this study show that rehabilitation patients with PCS in every rehabilitation indication are predominantly psychologically severely burdened, but that there are good therapeutic effects on depressive symptoms and improvements in functional aspects of health.

In summary, it can be said that more than 70% of post-COVID rehabilitation patients met the criteria for depressive disorders in general. The depressive symptom burden was comparably severe in all rehabilitation indications – even within somatic indications. Over the course of PCS rehabilitation, depressive symptoms decrease significantly and functional health improves in the areas of mobility and coping with everyday life. The symptoms of depressed PCS patients improved with predominantly medium effect sizes except for fatigue. Nondepressed patients showed only small effects; however, where baseline symptoms are minimal, limited improvement is expected. Approx. ¼ of PCS patients rated themselves as no longer able to work at the start of rehabilitation, which was largely consistent with the sociomedical assessment at the end of rehabilitation. Here, depressed patients had a significantly worse sociomedical prognosis: 26.6% were assessed as no longer able to work in their reference occupation and 20.4% as no longer able to work in the general labor market. Nondepressed PCS patients, on the other hand, appeared to have a more favorable sociomedical prognosis.

Prospects

Following rehabilitation, many PCS patients still require psychotherapeutic support to restore occupational performance, activity, and participation. For this purpose, special long-term reintegration models and a very slow increase in stress levels should be considered. The development of specific PCS psychotherapeutic therapy methods may help improve intervention outcomes in both inpatient and outpatient rehabilitation settings.

Statement of Ethics

Address the following aspects in your Statement of Ethics. This study protocol was reviewed and approved by the Ethics Committee of the Centre for Clinical Research, University Hospital Regensburg (identification 22-2814-101) in February 2022. Prior to study participation patients are informed about the study objectives and the study procedure in a personal information consultation and will receive standardized participant information sheets. Then, voluntary written informed consent for study participation and storage, evaluation and transfer of study-related data will be obtained from each study participant by research associates of the respective study center. Withdrawal of written consent is possible at any time, without giving reasons.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

This work was supported by the German Federal Pension Fund (8011 - 106 - 31/31.141). The funding body reviewed the study protocol before granting the funding. The German Federal Pension Fund had no further role in the study design, the data collection, analysis, and interpretation of the data, the writing of the report, reporting of this study, or in the decision to submit the paper for publication.

Author Contributions

All authors contributed to the design of the study with main contributors being V.K., T.H.L., A.A.K., and T.H. A.A.K. was the main contributor in drafting the manuscript. A.A.K.: conceptualization, writing – original draft, planning the methodological approach, and project implementation. T.H.: conceptualization, planning the methodological approach, project administration, and feedback on the initial draft of the manuscript. C.H., M.J., S.K., and G.S.: consulting and project implementation. T.H.L.: conceptualization, supervision, project administration, project implementation, and funding acquisition. V.K.: conceptualization, review and editing, supervision, project administration, and funding acquisition. All authors provided feedback on the initial draft of the manuscript and read and approved the final version.

Funding Statement

This work was supported by the German Federal Pension Fund (8011 - 106 - 31/31.141). The funding body reviewed the study protocol before granting the funding. The German Federal Pension Fund had no further role in the study design, the data collection, analysis, and interpretation of the data, the writing of the report, reporting of this study, or in the decision to submit the paper for publication.

Data Availability Statement

Since the data are on insured individuals of the German Pension Insurance and strict data protection exists, these data are not publicly accessible but can be made available by the corresponding author in a form adjusted for personal data upon justified request.