Changes in mental health services in response to the COVID-19 pandemic in high-income countries: a rapid review

Studies were eligible if they reported operational changes in mental health services such as new models of care/services, initiatives, adaption/expansion of existing services or changes in service delivery model in response to COVID-19 to provide support for patients aged 18 and over, published in the English language from January 2019 to present. All empirical studies, regardless of study design, conducted in an Organisation for Economic Co-operation and Development (OECD) country [22] (to ensure a degree of commonality in health system and socioeconomic and demographic context) were eligible for inclusion. We did not include grey literature, editorial commentaries, protocols or conference abstracts, views of the general public and letter of opinion to peer-review journals.

We searched four electronic databases including Cochrane CENTRAL, MEDLINE, Embase and PsycInfo using a range of keywords and subject headings validated by the information specialist representing COVID-19, mental health and OECD countries described in a Supplementary file 1. Searches were carried out in August 2022.

All records were exported to Rayyan [23], an online tool for review screening, for deduplication. All were screened by two reviewers (ES, AT) by title first followed by abstract if unsure. There was a 98% agreement rate. Where it was unclear based on title and abstract, the full texts were retrieved. Following the screening of titles and abstracts, full papers were retrieved and assessed for inclusion independently against the eligibility criteria by ES and AT. Where there was disagreement or uncertainty, studies were retrieved through consensus discussion with the research team.

Data extraction was carried out using two forms: one developed for mental health patients and one for care providers. ES and AT independently piloted the data extraction forms on a sample of four studies with different study designs. Extracted data included geographical location; population group; study design, methodology; description of mental health service provision prior to COVID-19; changes in service during COVID-19; experiences of service change and impacts and outcomes reported.

All remaining data extraction was conducted by ES and verified and appraised by the team through regular reflexive meetings. Studies included were scrutinised to ensure all relevant data were captured and extracted as appropriate. The review process did not incorporate an assessment of risk of bias as per protocol [19]. To ensure quality assessment a tabulated and narrative synthesis was conducted. This approach allowed to report and discuss the results of the included studies in accordance with best practice guidance [24‐26].

Guided by transparent and reproducible evidence synthesis we selected thematic and tabulated analysis of the data [24, 25, 27]. First, thematic analysis was selected to allow considerable latitude to reviewers and enabled the analysis of findings from both qualitative, quantitative and mixed methodologies [28]. Using this approach, we familiarised ourselves with the completed data extraction forms and inductively generated themes. Thematic analysis of each study was carried out to sort the findings into conceptual categories and groups.

Secondly, findings from the included studies were synthesised using tables and a narrative summary to report changes identified and their impacts reported as per protocol following current best practice to conduct synthesis systematically and transparently [24, 25, 29]. Meta-analysis was not possible because the included studies were heterogeneous in terms of the populations, study designs, methods and outcomes reported. Instead, we used effect direction following the Cochrane guidance [30] which allowed us to tabulate the reported heterogeneous changes in services and access the complex set of multiple outcomes. We applied the developed effect direction plot method with some adaptation to the characteristics of our research [30, 31]. A visual representation with arrows demonstrated the effect direction reported where ▲ was seen as an improvement in the outcome; ▼ was a deterioration in outcome and ◀▶ was no change. Due to a variety of study designs included, pooled analysis was not deemed appropriate, therefore we did not indicate study size and statistical significance.

A total of 6969 records were screened and 42 were assessed potentially eligible based on title and abstract. Of these, 33 studies met the inclusion criteria and were included in the review. A PRISMA flow chart is presented in Fig. 1.

Of the 33 studies included, 13 were quantitative, 12 were qualitative, 5 reported case studies, 2 were RCT, 2 were descriptive and 1 used mixed methods. The majority of studies focused on mental health patients (n = 24) followed by mental health service providers (n = 7) or both (n = 2). Nineteen of the studies were conducted in the US, and the rest were in the UK (n = 4), EU countries (n = 6, Ireland (n = 2), Italy (n = 1), Spain (n = 1), Austria (n = 1) and the Netherlands (n = 1), Canada (n = 3) and Japan (n = 1). Table 1 provides an overview of the included studies.

There was a high level of diversity of study populations included. The majority of studies (n = 23) were undertaken with a wide range of service users and staff including: general population [33, 40], patients with psychiatric disorder [31, 35, 54], patients with social anxiety disorder [38, 46, 56], patients with substance abuse disorder [37, 48, 61], patients with serious mental health illnesses (SMI) [31, 34‐36, 40‐42, 44, 48, 50, 51, 58, 59], and the rest of 10 studies had a narrow and specific focus on homeless people [32], prison inmates [39], patients in psychiatric centres [42], veterans [43, 47, 63], patients of specific treatment units or clinics (a tertiary care anxiety disorders clinic [45], (The Salamanca Psychiatry Department [49], patients at risk of suicide [60] and acute psychiatric patients who require a higher level of care [53]. Several studies included various groups of patients or focus of care provision (for example: a study includes patients with SMI and substance abuse disorder [37].

Most studies reported a transition of service delivery from in-person to a range of telehealth (n = 29) with only four studies reporting a hybrid method of service delivery. The most common teleconference method was through video calls (n = 15), multiple sources of contact through video, phone, texts, emails and electronic health consultations (n = 5), telephone (n = 5), both telephone and video calls (n = 4), face-to-face and phone or videoconferencing (n = 2), text messaging (n = 1) and use of immersive virtual reality reminiscence (n = 1).

The changes in mental health services were grouped into seven categories based on the description of key adjustments in service delivery and provision in response to COVID-19. They related to technological support when using or delivering care, accessibility and flexibility to services, remote diagnostics and assessment, privacy of patients and care providers, safety measures and operating hours of services (Table 2).

The changes in services in this study are found to focus on the procedural mechanisms of care delivery which influenced ways of service delivery (from face-to-face care delivery to telehealth or mixed), longevity of service delivery (reduction or increase of operating hours), characteristics of service delivery (flexibility/accessibility/privacy/safety). Despite the changes often being an ad-hoc response to lockdown restrictions, it produced diverse and diametrical impacts on both service users and staff. Studies have reported mixed experiences of receiving care remotely in a diverse range of service users. For example, service users who were homeless could not access the services and did not feel that their safety or privacy were maintained when using the services remotely [32]. The existing pre-pandemic research did not identify any specific groups of patients who experienced deterioration in access of telehealth and broadly referred to them as patients who live in isolated areas or those “wanting total anonymity for personal reasons” [64]. Conversely, new emergency contact line which was available 24/7 enabled more accessibility in service provision [36]. Combined with a variety of clinical settings (e.g. inpatient or outpatient care, closed units such as prison settings), models of service delivery (e.g. intensive compulsory therapy, optional regular text messages, voluntary group therapy), diversity of characteristics and sociodemographic factors of service users and staff specialities, evidence has been limited in the effectiveness and feasiblity of remote care implemented throughout the current health system, which is consistent with mixed impacts identified in this review. The existing evidence on use telemental health in pre-pandemic settings goes in line with some of our findings. Accessibility is found as the central feature of telemental health for the predominant number of patients which is supported by high satisfactory rates [64]. Among other changes, the existing literature reported cost implications [64] which was not identified in the current review.

Ten outcomes domains were reported including (1) care access; (2) satisfaction with telehealth; (3) comparability of telehealth with face-to-face care; (4) treatment effectiveness; (5) continuity of care; (6) relationships between patients and care providers (includes relationships and interactions between patients and care providers); (7) remote detection and diagnostics in patients; (8) privacy; (9) treatment length and (10) work-life balance. The most common three outcome areas reported in the included studies were access to services, satisfaction with telehealth and comparability of telehealth with face-to-face care. Table 3 demonstrates a list of outcomes reported followed by examples of positive and negative outcomes.

The effect direction plot (Fig. 2) was developed to visualise how the changes in mental health services impacted the experiences and outcomes reported in the included studies.

The effect direction plot illustrates a diversity of outcomes and often diametrically different experiences. Although the impact and experiences were very mixed across all of the included studies, there were two categories which demonstrated solemnly positive impact. It included improvement in stress management when accessing care and positive experiences of remote detection and diagnostics in patients.

Changes in flexibility, accessibility, operating hours of service delivery and methods of care delivery improved care access and reduced stress in accessing care [31, 33, 35, 37‐40, 42, 46, 49‐51, 56, 61, 62]. This was driven by four changes in service delivery including increased flexibility of service delivery, removing barriers to accessing care, positive changes in operating hours and care delivery via technology. Patients reported easier access (for example: [31, 35] less congested waiting rooms [39] and an increased likelihood of receiving same-day primary care compared to before the pandemic [43]. Appreciation of receiving support along with reducing risks of getting COVID-19 was also seen as an important feature of remote accessibility.

Remote diagnostics and assessment were reported in three studies where patients experienced improvement in accessing assessment services [31, 36, 42].

Conversely, three studies [32, 44, 58] reported reduced hours of service provision, lack of private space for meetings, absence of appropriate equipment and internet connection, and lack of IT skills, resulting in difficulties in receiving treatment. Although support was provided such as IT training and equipment provision, it was not always perceived as inclusive and feasible (for example: homeless [27], veterans [43] and elderly [52] did not have any access or skills to use technology [32]. It was also highlighted that Women and Black veterans were less represented among telehealth services when compared to in-person appointments [43].

Twelve papers found positive outcomes in patients’ continuity of care and extended length of care use [34, 35, 38, 42, 45, 47‐49, 53, 54, 58, 62]. Seven studies related positive outcomes of remaining in care to improved accessibility to care [34, 35, 47, 53, 54, 58, 62] including technology-based care delivery and flexibility of care by both care users and care providers [58, 62].

Patients and clinicians were able to increase hours of care provision due to improved capacity for delivering mental health services [38, 45, 53, 54].

Eight studies showed no significant changes in treatment outcomes between face-to-face and remote care delivery [37, 38, 47‐49, 52‐54]. Five studies reported that remote diagnostics and evaluation, accessibility to services, privacy of care providers, safety of patients and care delivery mode via technology contributed to improvements in patient’s mental well-being [31, 33, 45, 46, 60].

Relationships between patients and care providers were influenced by changes in the mode of care delivery. Five studies reported negative outcomes of having remote care due to the transition to telehealth [31, 32, 42, 55, 56]. Care providers reported that changes in service delivery mode together with changes in privacy and remote diagnostics positively impacted their relationships with patients and were seen as a comfortable way of providing care [58, 59, 61, 63]. An included study [59] found that 70% of service providers felt that video appointments were the same or even better than in-person appointments.

Remote diagnostics and assessment were reported to be an effective approach to evaluate patients’ conditions [31, 33, 37‐40, 58, 59, 66]. In addition to standardised assessment procedures, evaluation could also include such beneficial factors as real-time observations of home environment of a patient. As a negative impact, a fear of missing subtle signs of body language, nonverbal cues or physical signs of disease was seen as a possible barrier to using remote diagnostics [67].

A small number of studies reported that the privacy of patients and care providers was impacted by care delivered remotely. Lack of private and quiet spaces for an appointment, confidentiality challenges and inability to identify acute stress and agitation were all seen as barriers to maintaining private and safe care provision. One study reported improved experiences in privacy among care providers due to the transition to telehealth [32]. A combination of changes in privacy, safety and mode of care delivery all contributed positively privacy of patients. Three studies with patients [32, 35, 44] highlighted that changes in services delivered remotely jeopardised patients’ privacy due to the lack of private space for receiving care for patients.

Twenty-one studies [31, 35, 37‐42, 44, 49, 52‐63] reported patients’ and care providers satisfaction with care delivered in a remote format (patients = 15 studies; care providers = 6 studies). Twelve studies demonstrated high satisfaction rates among patients and care providers highlighting that the care delivered was of appropriate standard followed by positive overall experiences [31, 37, 39‐41, 52, 56, 58‐62]. Changes in service delivery mode via technology, accessibility and flexibility contributed to patients’ care satisfaction. Nine studies reported no differences in mixed satisfaction in experiences compared to pre-pandemic depending on individual circumstances.

Fourteen studies reported their assessment of service comparability [32, 38, 45, 47‐49, 51‐54, 57, 59, 60, 62]. Four reported preferences for telehealth compared to face-to-face provision [45, 51, 59, 62] due to remote care delivery method and accessibility to services. Nine studies suggested no difference [38, 47‐49, 52‐54, 57], and one study showed a preference for face-to-face care [32].

As a response to COVID-19, mental health services faced immediate challenges and opportunities to adapt the delivery of service provision underpinned by the need of increased physical distancing [68]. This often included telehealth referring to telemental health, telepsychiatry, teletherapy, or telepsychology [68]. Although telehealth has been in use for several decades, it has often been optional and has never been followed by a context of pressing and increased need for mental health services from wider population.

In the context of national reliance on mental health services’ transition to new safe formats of care delivery,, the review aimed to synthesise the changes to mental health services in response to COVID-19 and the impact of those changes. It identified changes to service delivery mainly from face-to-face provision to telehealth formats, primarily delivered through telecommunication technologies and videoconferencing supported by phone calls, text messages and hybrid modes of service provision. The service users in the included studies presented a range of population groups from broad catergories (general population, individuals with MCI or SMI) to narrow groups (veterans, prison inmates, eldery, homeless people, people at risk of suicide or at higher level of care). The existing research found that telemental health was predominantly used by women [69]. This has not been identified in our review due to lack of information on gender in included samples.

During a significant shift to telehealth the key changes in services included the transition to remote forms of care followed by technical aspects of using equipment among different patient populations, accessibility and flexibility of services, practicalities of diagnosing mental health conditions remotely, safety measures and privacy of remote services. Accessibility and flexibility were often identified as positive changes by a broad range of service users and care providers. The accessibility issue has also been evidenced in pre-pandemic use of telehealth where treatment facilities would otherwise be not available for underserved areas [70]. For some specific groups of patients (homeless people, current and ex-substance users) who heavily relied on mental health services providing space for therapy, accessibility was jeopardised by limited availability of safe and private space for remote meetings.

Where the operating hours of service delivery were increased either by introducing a new on-call member of care providers or opening a 24/7 emergency phone line, patients were also satisfied by the change. Conversely, in one patients’ group of homeless people operating hours were reduced which had a negative impact on their treatment. Implementation of remote mental health services was hindered by a series of barriers including identification of appropriate equipment, use of suitable applications and identification of telehealth strategies which would be patient-friendly and cost-effective. All of the mentioned features of telehealth require additional education for both patients and care providers which is also evident in other research [71‐75]. Where remote diagnostics was introduced, it appeared as a useful element to maintain assessment of patients’ state and conditions whilst following COVID-19 lockdown regulations. Remote assessment was particularly effective in advancing severe mental illness including schizophrenia-spectrum disorders [76]. Our review demonstrates a variety of different patient groups including the general population who had digital evaluation applied successfully. Although changes in privacy and safety were rarely mentioned in included studies, they caused a lot of concern among both patients and care providers. It was suggested that specifics of maintaining privacy and safety during telehealth required training, awareness and knowledge which was not often possible due to limited resources and time.

The review demonstrated that experiences and outcomes for patients and care providers were mixed and influenced by the telehealth changes identified. Experiences related to care access, care effectiveness, continuity of care use, relationships between care providers and patients, privacy and remote diagnostics were reported. The strong interconnections between telemental health changes and their impact on patients’ and service providers’ experiences of care were also echoed in pre-, during and post- pandemic studies demonstrating varied outcomes [68, 74]. For example, the previous research found that telephone services were the most affordable and easily accessible form of care in comparison with other ways including face-to-face services [68]. However, the study also indicated strong preference of patients to use videoconferencing over telephone-based services which raised a dilemma of most effective service modality. In contrast our review did not identify clear connections of service mode and experiences of accessibility of patients. Instead, access to care was shaped and supported by changes in flexibility, accessibility, operating hours of service delivery and digital methods of care delivery and was improved in nearly half of the included studies. This validates other studies [77, 78] demonstrating that accessibility and flexibility are significant cues to attend telehealth treatment despite the absence of face-to-face appointments. Not only were opportunities to reach care affected, stress levels of patients were also impacted. Where mental health services could not accommodate easy access to treatment and its round-the-clock availability and had to reduce hours of service provision, patients reported negative mental well-being.

The evidence on mental health interventions and service changes has focused on increasing attendance and reducing non-attendance of patients [79]. This was echoed in this review that changes in improved access and flexibility of care were able to accommodate extended treatment length and reduced the number of dropouts from care.

Mixed interpersonal relationships between patients and care providers were identified by a rapid shift to telehealth. The existing pre-pandemic research suggested that the relationships between patients and care providers need to maintain a subtle balance between in-person and remote interactions to avoid risk of losing and openness [80]. In contrast to gradual transition to telehealth [81], a critical factor in a balanced care provision between in-person and remote care, a rapid shift to virtualization of both patients and care providers identified in this review subsequently led to some negative impact on patient-care provider relationships.

Positive outcomes in relationships which were reported by care providers were associated with accessible and regular appointments often easier to achieve in digital care as opposed to face-to-face provision. Conversely, due to care providers’ turnover patients experienced disjoined care. The pre-pandemic research reported that telehealth was “manageable” for patients when they received support from a care team they previously met face-to-face [82]. Our findings showed that rapid changes in staff in delivering telehealth often resulted in care provision by someone a patient did not know or met, leading to discontinuity of care.

Telemental health was considered to be “unquestionably effective” across many population [64] pre-pandemic and consistently useful in reduction of severe sysmptoms in patients with SMI [83]. The more recent research focusing on effectiveness of telemental health referred it as “not inferior to face-to-face services” [84] and found it to be a feasible mode of treatment and equally acceptable when compared with face-to-face services [68, 85‐87]. In this review, although the predominant number of studies suggests that remote mental health care is a viable alternative to face-to-face service delivery with associated improved treatment outcomes we cannot draw the conclusion that telemental health is a desiarable form of mental health services due to heterogenious accessibility, privacy, safety and flexibility needs of patients. By synthesising the available evidence and a mixed range of population groups, patients’ mental health conditions, experiences and treatment arrangements in the included studies, it is important to consider the individual circumstances and patients’ needs when telehealth is applied.

This rapid review was designed and reported following recommendations by the Cochrane guidance for rapid reviews. A review protocol was designed and published [19]. A rapid yet thorough search of multiple databases was conducted. Despite the rigorous search, screening, and analysis process, we acknowledge the existence of some limitations of this study. Firstly, grey literature was not searched which may provide an important forum for disseminating studies that might otherwise not be disseminated. However we aimed to draw on conclusions based at rigorously designed and conducted studies but no all grey literature is subject to a similarly rigorous pre-publication review process [88]. Second, the selection criteria for this review centred on studies in English and limited to OECD countries to ensure a degree of commonality in context. In addition, a formal risk of bias assessment was not undertaken due to the design being a rapid review. However, this is the first rapid review addressing what and how mental health service changes impacted patients’ and care providers' outcomes during COVID-1.