Introduction

The COVID-19 pandemic began in China in December 2019 [1], and as of December 2022, it had claimed the lives of more than six million people and infected more than 640 million around the world [2]. Most attention has focused on the acute COVID-19 infection [3], with comparatively little devoted to the longer term and ongoing symptoms experienced and reported by many people [4]. These people, who often felt their symptoms were neglected or disbelieved, formed support groups on social media, drawing global attention to this ongoing condition, which was named ‘long COVID’ (LC) [5]. This condition has been described by several other labels, including post-COVID, long haulers, post-acute COVID, and post-SARS-COV-2 [6,7,8]. It has also several definitions [9]. For example, in the UK, according to the National Institute for Health and Care Excellence (NICE), LC represents “signs and symptoms that develop during or after an infection consistent with COVID-19, continue for more than 12 weeks and are not explained by an alternative diagnosis… It includes both ongoing symptomatic COVID-19 (from 4 to 12 weeks) and post-COVID-19 syndrome (12 weeks or more)” [10]. Similarly, the World Health Organisation (WHO) defines LC as follows: “Post-COVID-19 condition occurs in individuals with a history of probable or confirmed SARS-CoV-2 infection, usually 3 months from the onset of COVID-19 with symptoms that last for at least 2 months and cannot be explained by an alternative diagnosis.” [11].

Research on LC shows that following the acute viral infection, damage can occur in different organs in the body including the brain [12]. There are multiple potential symptomatic sequelae including musculoskeletal problems, cognitive impairment, fatigue, and psychological distress [13,14,15,16].

Chronic illness and disability caused by LC represents a major emerging health challenge [4], LC can affect all aspects of an individual’s daily life, including quality of life (QoL), social functioning, and ability to participate in normal activities [17, 18]. In addition, mental health problems appear common in LC [19,20,21,22].

The precise proportion of people who contract COVID-19 and have LC has not been firmly established [23,24,25], though it is estimated that, globally, around 43% of people experience some degree of long-term symptoms following the acute infection [26]. In the UK it has been estimated that more than 1.8 million people have LC [27], though this will likely increase as COVID-19 continues to circulate [3]. Those at most risk of LC are women, elderly people, smokers, obese or overweight people, and those living in socio-economically challenged areas [4, 28].

The impact of LC on mental health has been examined. Studies report that more than 87% of people with LC have mental health symptoms, including anxiety, depression, and cognitive problems [29,30,31,32,33,34,35,36]. As LC is a relatively new illness, it is surrounded by a great deal of uncertainty: it can manifest differently from one person to another, with more than 200 associated symptoms [31] and presentations that do not necessarily fit into established clinical categories [4]. Indeed, some consider LC to be a form of post-viral syndrome [37], which provides some evidence base for treatment and management. However, in the face of uncertainty, patients can be at risk of being either over-investigated and over-treated, or under-investigated and under-treated [4]. As such, no evidence-based treatment options are yet available [38,39,40,41]. Usual care currently provided to people with LC is through LC clinics and services, and general medical practices [42, 43].

A recent systematic review showed that while some research is currently being conducted to examine interventions to support the mental health of people with LC, the breadth and scope of this research is limited [35]. This systematic review was focused on ongoing trials that were conducted between January 2020 to May 2022. In this context, this scoping review aims to more broadly identify what interventions have been examined to support the mental health of people with LC.

Methods

Database search

This scoping review was conducted following Arksey and O’Malley framework [44] and is reported following PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) guidelines [45]. No priori protocol to this scoping review was published or registered. A systematic search was conducted on 3 October 2022 to identify relevant studies using MEDLINE, EMBASE, Web of Science, CINAHL, and PsycInfo databases from January 2020 to early October 2022. Searches were limited to journal articles. Records were exported initially into the reference manager Endnote, where duplicates were identified and removed, and then to Covidence software for data management [46].

Search terms

Between 16 September and 3 October 2022, preliminary searches were initially conducted using the databases mentioned to identify appropriate keywords to use in this review. Search terms were selected based on the target population and the target intervention, and terms were grouped under three headings related to LC, intervention, and mental health problems associated with LC. The search keywords are detailed in Table 1, (see Additional file 1 for an example of the search strategy). Moreover, reference lists of included studies and related reviews were also examined to identify additional potentially eligible studies. Grey literature was also checked to identify any additional relevant studies, using the following resources: Social Care Institute for Excellence (SCIE), Canadian Agency for Drug and Technologies in Health (CADTH), King’s Fund, Open Grey, and the Health Foundation. The search strategy was adapted appropriately when searching the different databases.

Table 1 Search keywords
Full size table

Study inclusion criteria

As is conventional with scoping reviews [47, 48], our inclusion criteria were deliberately broad, namely:

  1. 1.

    Research focused on LC and mental health

  2. 2.

    Research that introduces/discusses an intervention/treatment/support for mental health issues associated with LC

  3. 3.

    Research where assessment of mental health is the primary outcome (since mental health is commonly affected by LC and is also impacted by physical symptoms of LC).

  4. 4.

    Patient population is aged 18 years and above

  5. 5.

    Publications from 2020 to 2022

  6. 6.

    Original research (article/journal article) (including qualitative, quantitative, or mixed research methods, case reports or other study designs that reports an original research).

Studies that did not meet one or more of the above criteria or that met any of the following exclusion criteria were excluded from the review. Research that is focused on:

  1. 1.

    COVID-19 acute infection and/or its treatment

  2. 2.

    Treatment of mental health issues not related to/ induced by LC

  3. 3.

    Treating issues caused by LC but not related to mental health (e.g., respiratory, or cardiovascular problems).

Study selection

Title and abstracts, then full text screening were independently carried out by two reviewers (HA and LDH), to assess article eligibility against the inclusion and exclusion criteria. Any arising disagreements were resolved by discussion between the two reviewers, and if necessary, by consulting with a third reviewer (CFS).

For studies not written in the English language but were found relevant at title and abstract screening, corresponding authors were contacted to request relevant information; when additional information was not provided by the authors, Yandex website (https://translate.yandex.com/doc) was used to translate these studies to confirm eligibility and extract relevant data. Additionally, where full texts of included studies were not retained even after contacting corresponding authors, their abstracts were included and used for data extraction.

Data extraction

A data extraction template was designed using Microsoft Excel to extract the following items-where available- from each study:

  1. 1.

    Author(s) and publication year

  2. 2.

    Study design

  3. 3.

    Study setting and country

  4. 4.

    Participant information (recruited participants, age, gender, and ethnicity)

  5. 5.

    Practitioners involved in delivering the intervention/treatment

  6. 6.

    Description of the intervention

  7. 7.

    Description of the outcome measures and study findings.

Data extraction was carried out by HA and was reviewed and verified by LDH, AC, MS, DC, DRT and CFS to ensure accuracy and completeness. Disagreements were resolved by discussion or through consultation with a third reviewer (CFS).

Data synthesis

A narrative approach was used to synthesize and describe the data extracted from included studies [49].

Dealing with missing data

Corresponding authors were contacted by email to collect any missing data. When no response was received, studies were still included in the review with missing data designated as ‘not reported’.

Results

A total of 940 unique records were identified through the database search. No additional studies were identified from the grey literature or from the reference lists of included studies. After the complete screening and selection process, the final review encompassed 17 studies. The findings of the search process are presented in the PRISMA flow diagram (Fig. 1).

Fig. 1
figure 1

PRISMA flowchart

Full size image

General description of included studies

Table 2 provides an overview of included studies. Of all included studies, one study was retained based on a relevant abstract in the absence of a full text despite multiple attempts to obtain it [50]. All studies were written in the English language except for three, written in Russian [50,51,52]. The included studies were conducted in nine countries, with more than half conducted in Europe (n = 11, 64.7%), five of which were from the UK.

Table 2 Description of included studies
Full size table

Of the 17 included studies, six were case studies [53, 56, 59, 62,63,64] and five were clinical trials [50,51,52, 60, 61], two of the latter were randomised controlled trials (RCTs) [60, 61]. Other designs encompassed four single arm pre-post studies [55, 57, 65, 66], one qualitative study [54] and a single service evaluation study [58]. Several settings were described by included studies including primary and secondary healthcare settings, a university setting in collaboration with a local GP collaboration group, outpatient subspeciality clinic, and hospital outpatient care. Three studies did not clearly report their setting [51, 56, 63].

Eight studies did not report the method used in recruiting participants [50,51,52, 56, 57, 59, 63, 64]. Described recruitment approaches included social media, referral from LC clinics, medical doctors, or a local community network. One study recruited staff with LC opportunistically from specific settings [55].

A total of 940 participants were recruited across all included studies, with individual study numbers ranging from one [56, 59, 62,63,64] to 219 [55]. All case studies (n = 6) recruited only one or two participants [53, 56, 59, 62,63,64]. Participants were recruited at different stages of their LC journey (i.e., at different time points since having the acute COVID-19 infection). This was reported by 13 studies [50, 51, 54,55,56,57, 59,60,61,62,63,64,65], ranging from one month post-acute COVID-19 [57] to over a year post infection [64, 65].

The gender of included participants was reported by all studies except one [50], but in two further studies, gender was reported as a missing variable for some participants [55, 58]. Most recruited participants were women (total n = 353, range 1–121) [51,52,53,54, 56,57,58,59,60,61,62,63,64,65,66, 63]. One study reported the gender of participants pre and post assessment [55].

Age of included participants was reported by all studies, either as a precise age (n = 5) [53, 56, 59, 62, 64], or as an average (mean, median) of all participants (n = 2) [57, 61], or as an average age of groups of participants allocated at different treatment groups or at different assessment time points (n = 4) [50, 52, 55, 60], or as a range (n = 6) [51, 54, 58, 63, 65, 66]. Across all studies, age ranged from a 15–24 age group to 92 years [58].

Ethnicity or cultural background of participants was reported for only nine studies, with half of these including only people identifying as white or Asian and the rest reporting a variety of ethnicities/cultural backgrounds.

Description of study intervention/treatment

Table 3 describes the various interventions tested by the included studies, ranging from a single intervention (e.g., using a pharmacological product, sessions of electromagnetic field, modifications to diet) (n = 7) [50, 51, 59, 61,62,63, 65], to integrated programmes covering diverse elements of health (n = 10) [52,53,54,55,56,57,58, 60, 64, 66]. Healthcare components included rehabilitation programmes, lifestyle interventions, stress management, sleep hygiene, breathing techniques, optimising diet, energy conservation, psychoeducation, maintaining social contacts with close networks, reading, relaxation techniques and cognitive behavioural therapy (CBT). Some interventions were focused on testing the effects of certain pharmacological products or nutritional changes on enhancing outcomes.

Table 3 Description of interventions/treatment
Full size table

Practitioners involved in delivering the intervention were reported by 12 studies, five of which engaged practitioners of a particular speciality, (e.g., medical doctors or clinical psychologists) [53, 60, 62, 64, 65]; one was being supervised by a cardiopulmonary specialist [64]. A multidisciplinary team approach was adopted in the other six studies, including physicians, occupational therapists, psychologists, physiotherapists, dieticians, speech and language therapists, sport medicine specialists and CBT therapists [54,55,56,57,58, 61]. The remaining six studies did not report who was involved in delivering the intervention [50,51,52, 59, 63, 66].

The duration of intervention delivery was reported by all studies except one [66], and ranged from a single dose [62] to 20 weeks [66]. Frequency of intervention delivery was not clearly described by two studies [64, 66], however, it varied widely between the remaining studies from daily [50,51,52, 63], once a week [53,54,55, 58, 61], twice a week [56, 59], or multiple times a week [57, 60, 65].

Some studies used a virtual approach to deliver the intervention (using online platforms for example to deliver sessions or a virtual clinic) [54, 55, 60, 61, 64, 66]. One study used an artificial intelligence-based intervention to deliver a personalised care approach [60].

Description of study outcomes

Outcome measures are described in Table 4. Mental health outcomes included anxiety, depression, stress, sleep, neuropsychological functioning, and wellbeing. The timing of outcome measures was reported by all studies except one. Most studies used pre- and post-intervention measures, with additional follow-ups.

Table 4 Description of outcome measures and findings
Full size table

The most commonly reported questionnaires for anxiety and depression were the Hospital Anxiety and Depression Scale (n = 3), the Generalized Anxiety Disorder-7 (n = 2) and the Patient Health Questionnaire-9 (n = 3). The Short-Form-36 questionnaire was mostly reported for measuring health-related QoL (n = 4). Other studies used a wide array of scales and measures, as detailed in Table 4.

Description of study findings

Table 5  provides a description of study findings. Positive findings were reported by all studies. Nine studies reported significant improvements in different outcomes following the intervention [50,51,52, 55, 57, 60, 61, 63, 65], of which five are trials [50,51,52, 60, 61].

Some of the interventions used by these studies included using drug products i.e., Cholytilin and MexiB 6 for anxiety and depression, Intravenous Cytoflavin in combination with a standard rehabilitation program for depression, Anvifen for anxiety, depression, health-related QoL and sleep. Significant improvements were reported on these outcomes. Other studies examined wellbeing services or rehabilitation programs, such as Multidisciplinary Rehabilitation (MDR) for anxiety, depression and health-related QoL, and Primary Care Wellbeing Service (PCWBS) for general health and health-related QoL, with both showing significant improvement in these outcomes. Neuropsychological programmes included cognitive treatment, together with other respiratory and physiotherapy, for anxiety and depression, that were significantly improved. An English National Opera (ENO) programme that was delivered online and was focused on breathing re-training through singing techniques and utilising lullabies, was used to assess health-related QoL with significant improvement only reported to the mental health composite score. Another intervention included using a high-fibre formula examined anxiety that was reported to be significantly improved. Lastly, a neurofeedback (NFB) program that included locating electrodes on the right and left temporal lobes to examine anxiety and depression, with significant improvements reported. (See Tables 3, 4 and 5 for more details).

Table 5 Description of study findings
Full size table

Discussion

Summary of main results

To the best of our knowledge, this is the first scoping review focused on interventions tested for their impact on mental health issues associated with LC. Seventeen studies were identified, with a wide range of recruitment methods, number of participants recruited, intervention tested, duration of the intervention, practitioners involved in delivering the intervention, outcome measures, and post intervention follow up. Various tools were employed to measure a wide range of outcomes. Results were generally positive, with all included studies reporting improvements in mental health outcomes. However, as most studies were case reports, which are known to be of low-quality evidence compared to RCTs and cohort studies [67], the findings of these studies should be interpreted with caution.

Recruitment and participants

All participants recruited into the included studies were reported to have LC. However, it was not clear in some studies how long participants had been suffering from LC or when they had their acute COVID-19 infection. These participants may have varied with regard to their LC diagnosis, and the duration of time since they had acute COVID-19 infection. The lack of a consistent definition of LC resulted in a heterogeneous sample of participants joining research at different stages of their illness. This might have implications on the impact of the interventions, and the number of symptoms and symptom severity. Therefore, future research needs to take this into consideration and also interrogating whether interventions have different impacts at the different stages of the LC journey.

The over-representation of women and older participants reflects the literature indicating that these are risk factors for LC [26, 28, 68]. However, some studies indicate that people from minority ethnic backgrounds are more likely to acquire a COVID-19 infection compared to the general population [69,70,71].

Thus, the lack of ethnic diversity in studies in this review leaves a knowledge gap and suggests particular attention be given to ethnic and cultural diversity in future studies.

Outcomes and outcome measures

Mental health was assessed in multiple ways, mostly focusing on anxiety and depression, with a number of studies also considering QoL. This reflects the broad nature of mental health and that it encompasses multiple aspects of people’s physical, social and psychological life [72]. A wide range of measures were adopted, making cross-study comparison difficult. Researchers might consider using a standardised core outcome set to enhance the comparability of outcomes in future research, as suggested in a recent Delphi consensus study [73]. Additionally, from a patient and public engagement standpoint [74], people with LC could support future trials throughout the research process, including in selecting the best questionnaires to use for the targeted outcome in manners that resonate with their lived experience. Almost all studies assessed targeted outcomes before and after the intervention, with few studies conducting further follow-up measurements to follow progress over time. The addition of follow-up assessments in future research is of high importance, as this would allow researchers to identify the longer-term impact of their intervention [75].

Interventions and findings

Although this review included only interventions targeting the mental health of patients with long COVID, a wide range of interventions was described. Some interventions were of specific pharmacologic products or looked at dietary change. However, the majority of studies used an integrated approach to try to address the different elements of participants’ mental health and QoL. Most integrated treatments were delivered by a multidisciplinary team and patients needed to take active roles in implementing agreed therapeutic strategies.

Given the heterogeneity of interventions described, no specific ‘standard’ treatment can be delineated. However, our review suggests that a wide variety of interventions have the potential to support mental health of individuals with LC. and a variety of integrated care models have been examined.

All interventions used by studies in this review were associated with positive outcomes. However, considering study designs of most included studies, the findings should be interpreted cautiously. More large-scale, high-quality trials are needed to define effective intervention strategies for LC. As LC presentation varies between individuals, interventions addressing a wide range of symptoms including mental health and QoL could be promising in building the evidence base.

Strengths and limitations

The findings of this scoping review contribute to the body of knowledge on LC. Strengths of the review include: 1) Adherence to the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) guidelines [45], including the widest available evidence through a combination of complementary keywords to search all related databases and was also covering grey literature and the reference lists of all included studies and related reviews; 2) no language restrictions - all retrieved studies in different languages were examined, translated, and included; and 3). Selection bias was minimised by having two independent reviewers conducting the title, abstract and full text screening, with a sample of data extraction being reviewed by a third person.

However, some limitations were encountered with this review. There was missing data from some included studies despite attempts to contact corresponding authors, which precluded rigorous comparisons being made. Another limitation was related to the search strategy, which was limited to studies where mental health was the primary outcome. This may have led to missing some important interventions that may have had positive impacts on mental health as a secondary outcome.

Conclusions

A wide range of interventions designed to support mental health in patients with long COVID has been examined. Overall, these interventions reported positive effects on mental health outcomes, though more research is needed to confirm this. Considering the ongoing nature of LC and treatment recommendations to date, it is important to address patients’ care needs holistically, ideally through using an integrated care approach.