The taxonomy of “digital health interventions” and their subgroups in the therapeutic rehabilitation of stroke patients – A scoping review / Die Taxonomie der “digitalen Gesundheitsinterventionen” und ihrer Untergruppen in der therapeutischen Schlaganfallrehabilitation – Ein Scoping Review

Stroke is a leading cause of disability and death worldwide. Currently there are more than 101 million people living with stroke worldwide (Feigin et al., 2021). Many of these stroke survivors suffer from motor, sensory and speech impairments that limit daily life activities and result in an ongoing or permanent need for rehabilitation (Patel et al., 2006; Sturm et al., 2004). Allied health professionals such as physiotherapists, occupational therapists and speech and language therapists provide rehabilitation and therapeutic treatment for stroke survivors. Rehabilitation is known to mitigate motor, sensory and speech impairments (Dobkin, 2005).

However, many patients do not continue their treatment after leaving the hospital and returning home because access to rehabilitation services is often limited by transport, lack of specialists, cost, geographical barriers, awareness, and inequalities (National Guideline Alliance [UK], 2022). Some of these barriers, however, can be overcome by digital interventions. New technologies such as videoconferencing offer an opportunity to address this problem. There is low-to-moderate evidence that telerehabilitation services have a positive effect on the physical functioning, quality of life and speech of stroke survivors (Laver et al., 2020). They are therefore important not only in patient care, but also in the research landscape.

There are many definitions of digital interventions, with related terms that are inconsistent and often have minor differences (Bashshur et al., 2011). Perednia & Allen (1995) provided one of the first definitions of telemedicine that has been cited to date: “use of telecommunications to provide medical information and services.” Then, in 2009, Brennan et al. established a widely used definition of telemedicine and telerehabilitation. Telemedicine is defined as the transfer or exchange of medical or health information using information and communication technologies, and telerehabilitation as the use of these technologies for the purpose of rehabilitation services (Brennan et al., 2009).

After a period of using various definitions and terms, the WHO introduced a taxonomy in 2018 to promote a shared language around the use of digital technology for health (World Health Organization, 2018). The Classification of Digital Health Interventions (DHI) categorizes the different ways in which digital and mobile technologies are being used, including interventions for clients, healthcare providers, managers of healthcare systems, and data services. Targeted to the primary user, telemedicine is indexed in the category for healthcare providers. However, terminology and definitions around digital stroke rehabilitation are used inconsistently across health professions.

The lack of clarity in the definitions and terms, underlying technological concepts and location contexts makes it difficult to compare the findings and results of different studies and to assess the evidence in these areas. Therefore, the aim of this review is to provide a comprehensive overview of the definitions and terms used in the field of digital interventions for stroke delivered by healthcare professionals. In order to understand the use of terms in relation to the underlying technological concepts and location contexts, which broadly fall under the WHO taxonomy of ‘consultations between remote clients and health care providers’, a concept map of the terms and synonyms used in papers on stroke rehabilitation by occupational therapists, physiotherapists and speech and language therapists will be derived. In a second step, we attempt to develop a definition for the terms used in synchronous remote stroke rehabilitation.

The scoping review has been conducted in accordance with the updated JBI methodology for scoping reviews (Peters et al., 2021) and follows Arksey and O’Malley’s five-stage framework for conducting a scoping review (2005). The protocol of this review was registered at the Open Science Framework (OSF).

The PCC (population, concept and context) framework was followed to establish clear and meaningful objectives and eligibility criteria. For this scoping review, papers about adult stroke survivors (population) were eligible if they had received synchronous rehabilitation services delivered by physiotherapists, occupational therapists and/or speech language therapists using dedicated high-quality, interactive, bidirectional audio-visual systems, virtual reality-based or sensor-based technology (underlying technological concepts). Sources of evidence were included if rehabilitation services were delivered remotely to or from a patient’s home, clinic or rehabilitation centre (location context).

The search strategy aimed to locate both published and unpublished studies. An initial limited search of MEDLINE was undertaken to identify articles on the topic. The text contained in the titles and abstracts of relevant articles, and the index terms used to describe them, were used to develop a full search strategy within the Medline database, the Cumulative Index to Nursing and Allied Health Literature (CINAHL) and the Cochrane Database of Systematic Reviews. The search strategy, including all identified keywords and index terms, was adapted to each included database and/or information source. Following the protocol, papers published after 2000 for human adult stroke survivors were included in the study, as this is considered to be when health-related research using telecommunication technology came into common use. As shown in Figure 1, the first systematic electronic search was carried out in September 2021. A further search was conducted in January 2023.

Figure 1

Following the search, all citations identified were collated and uploaded into CADIMA, a free web tool that facilitates and ensures the documentation of reviews and allows for the automatic removal of duplicates (Julius Kühn-Institut, 2023). Following a consistency check of the keywords and criteria, three independent reviewers for assessment against the review inclusion criteria then screened titles and abstracts. The reviewers then assessed the full texts of selected citations in detail.

Any disagreements that arose between the reviewers at each stage of the selection and extraction process were resolved through discussion. As the aim of this review is to identify trends and gaps in the existing literature, critical appraisal is not necessary to answer the research questions. It is therefore not included.

Data were extracted from the papers included in the scoping review by KS-G with the help of a data extraction tool developed by KS-G and AE. The chart includes authors, years and countries of publication, study methods, technological concepts, location contexts, and key findings relevant to the review questions (see (Supplement 1)).

The results were summarized using MAXQDA software (VERBI Software, Consult, Sozialforschung GmbH) by KS-G. The coding of the segments followed the PCC framework as well as those of the allied health professions. Some publications contained information on more than one allied health profession, location context or underlying technology concept, and thus some segments were coded more than once. Once the coding was finished, a word cloud of the most frequently used terms was created using MAXQDA. Diagrams were rendered in Microsoft Excel 2016.

As stated in the research protocol, this scoping review used a broad approach to assimilate published research literature. It also considered text and position papers and congress papers and posters.

As shown in Figure 2, systematic searches identified 1,200 records. After the removal of duplicates, the remaining 1,090 titles and abstracts were screened, and 843 records were excluded. Of the 247 full-text articles assessed for eligibility, 152 full-text articles were excluded. The remaining 95 papers have been included in the scoping review. An additional reference list of all included papers can be found in the (Supplement 2).

No French-language texts, and only two German-language papers, were included; all other articles from French-speaking countries (including Belgium and Switzerland) and German-speaking countries (including Austria and Switzerland) had originally been published in English. Of the 93 papers published in English, two were from Africa and four from Australia. Forty-three were from North America, with eight from Canada, including one Canadian-Israeli collaboration and one Canadian-Irish collaboration. Thirty-two articles were by European authors and 14 articles were from Asia.

Regarding the study methods, nearly two-thirds of the included papers were primary studies (n = 62), of which 48 had a quantitative study design: -

14 randomized controlled trials (RCTs): Asano et al., 2021; Borstad et al., 2022; J. Chen et al., 2017; Grau-Pellicer et al., 2020; Lin et al., 2014; Linder et al., 2015; Maresca et al., 2019; Øra et al., 2020; Ozen et al., 2021; Palsbo, 2007; Piron et al., 2008; Piron et al., 2009; Thielbar et al., 2020; and Uswatte et al., 2021.

Figure 2:

27 secondary analyses of primary data were included: -

11 Systematic Reviews (SR): Appleby et al., 2019; Cacciante et al., 2021; J. Chen et al., 2015; X. Chen et al., 2022; Johansson & Wild, 2011; Laver et al., 2013; Laver et al., 2020; Rintala et al., 2019; Sarfo, Ulasavets, et al., 2018; and Schröder et al., 2018.

Additionally, five study protocols (Gauthier et al., 2017; Jhaveri et al., 2017; Koh et al., 2015; Øra et al., 2018; Park et al., 2022; and Tousignant et al., 2014), one editorial letter (Keidel et al., 2017), one position statement (AOTA, 2018) and two conference papers/posters (Ivanova et al., 2017; Minge et al., 2017) were included.

Figure 3:

The underlying technological concepts of the included papers were based on a range of technologies, including “information and communication technologies, including video/teleconferencing, remote data collection devices, telemonitoring, computers, mobile phones, robotic devices, exergames and virtual reality (VR)” (Nikolaev & Nikolaev, 2022). According to the ontology-based classification published by Otto (2020), the underlying technological concepts can be divided into three conceptual groups: image-based, virtual-environment and virtual reality-based, and sensor-based concepts.

As shown in Figure 3, the majority of publications used image-based concepts (n = 30). These include information and communication technologies (ICT) that are used to facilitate communication between the healthcare professional and the patient in a remote location using telecommunication devices (e.g. telephone, mobile phone, video conference) that are defined as “high-quality, interactive, bidirectional audiovisual systems” (Schwamm et al., 2009).

Virtual environment (VE) or VR-based concepts including augmented reality were described in 21 publications. Chen (2022) describes a virtual reality platform that uses a simulated environment to provide immersive, intuitive and interactive feedback that stimulates the neuroplasticity of the human brain. Holden (2007) describes a VE-based telerehabilitation for the improvement of upper limb function in patients with stroke. This rehabilitation service can be delivered from a remote location while the patient remains in their home environment.

Sensor-based concepts were used in 23 publications. This conceptual group includes any type of robotic, gaming and kinect-based rehabilitation services delivered remotely, and comprises terms like ‘exergaming,’ ‘serious games,’ and ‘wearables.’ A mix of different technologies were used in 18 publications.

The location contexts of the included papers (see Figure 4) show that the majority of the described projects were conducted at home (n = 63), including a Taiwanese pilot study in a long-term care facility. Lin (2014) tested a bidirectional and multi-user telerehabilitation system for chronic stroke patients aiming at balance and satisfaction through a wireless sensor network (WSN). Sixteen publications were conducted in both the clinic and the home. During their stay in the clinic, stroke survivors were recruited and trained to use the devices. They then took the devices home and continued their therapy via ICT. This approach was mainly presented in Speech and Language Therapy (SLT) publications (n = 8) for stroke survivors with aphasia.

Figure 4:

Figure 5:

Clinic-based interventions using telemedicine were mainly used in experimental pilot studies and case studies with small sample sizes in the developmental phase and for concept evaluation (n=13). Five publications focused on VR-based concepts, and five addressed sensor-based robotic and exergaming, including two constraint-induced movement therapies for the upper extremities (CIMT) with a device called ‘Automated CI Therapy Extension’ (AutoCITE). Lai (2004) conducted a community-based stroke rehabilitation program via videoconferencing that included educational talks, exercise and psychosocial support for stroke survivors attending a community center in China. Rehabilitation centers were also the site of two pilot studies, one in Spain and one in Portugal.

The included papers were searched for terms and synonyms describing ‘consultations between remote clients and health care providers’ according to 2.4.1 of the WHO taxonomy of Digital Health Interventions (DHI). The full variety of terms used in the 95 publications is shown in the Venn diagram (Figure 5 on the previous page).

Of the 95 publications that met the inclusion criteria, 29 dealt exclusively with PT. Interdisciplinary work involving PT and OT was the subject of 14 publications, and collaborative treatment using PT, OT and SLT was the subject of 11 publications. One publication shows a link between PT and SLT. Eighteen publications refer exclusively to digital SLT interventions.

The majority of publications used the term ‘telerehabilitation’ (n = 56) to describe ‘consultations between remote clients and health care providers,’ as shown in the word cloud (Fig. 6). The term ‘telerehabilitation’ (also referred to as TR or ‘telerehab’) was used by all three allied health professions, as were virtual reality interventions: ‘VR-rehabilitation’ (n = 7) and ‘VR-therapy’ (n = 4). The term ‘teletherapy’ (n = 9) was found in PT and SLT publications and ‘telehealth’ (n = 8) was used in PT and OT publications. The terms ‘telemedicine’ (n = 4) and ‘telepractice’ (n = 3) were used exclusively in SLT publications, with ‘telepractice’ being the term suggested by the American Speaking and Hearing Association (ASHA) (2023).

As shown in Figure 7, the terms ‘telesupervision,’ ‘telepractice,’ ‘teletherapy,’ ‘virtual rehabilitation,’ ‘augmented rehabilitation’ and ‘telecoaching’ were all found to be equivalent to ‘telerehabilitation’ and were used in the narrow sense of identity (blue arrows).

‘Telesupervision,’ ‘telepractice,’ ‘teletherapy,’ ‘virtual rehabilitation,’ ‘augmented rehabilitation’ and ‘telecoaching’ have a number of subordinate terms. For example, the term virtual rehabilitation (VR) is superordinate to the terms ‘videogame-based rehabilitation,’ ‘e-visit’ and ‘gaming exercise.’ These subordinate terms have a similar meaning and indicate a relationship to the superordinate term (black arrows).

The relationship between super- and subordinate terms is reversed in some publications. For example, in some papers, the term ‘telespeech therapy’ is also referred to as ‘telepractice.’ However, as ‘telepractice’ was not used synonymously with ‘telespeech therapy,’ so the gradation of the terms is to be understood as subordinate rather than subsidiary.

When reducing the complexity of the above figure, it is obvious that all terms and synonyms focus on telerehabilitation. The number of papers focusing on ‘telerehabilitation’ annually has increased over the last 20 years, peaking in 2020. As shown in Figure 8, 31 of these publications were published after 2016.

Figure 9 is a concept map of location contexts and the underlying technological concepts in relation to the allied health professions providing the Digital Health Interventions (DHI). Heterogeneous uses of technological concepts have an equal distribution across the different location contexts and allied health professions. Home-based digital intervention predominates in all professions, although the technology used varies. In SLT, the focus tends to be on image-based interventions (e.g., for naming objects), in OT on sensors for the promotion of perception (e.g. the correct movement of the upper limb), and in PT on VR-based interventions (e.g., for the promotion of balance).

Figure 6:

Figure 7:

Figure 8:

Within the 95 included publications, 40 provided a definition for ‘telerehabilitation.’ The definitions reference either ‘tele,’ from ancient Greek (τῆλε), meaning ‘remote,’ or ‘at a distance,’ the first part, or ‘rehabilitation,’ meaning the provision of therapeutic services, in the second part.

According to Rosen (1998), who first defined telerehabilitation as ‘the delivery of rehabilitation services at a distance,’ definitions that are more recent seek to incorporate underlying technological concepts and location context. Lai (2004) defined telerehabilitation as ‘the provision of distance support, assessment and intervention to individuals with disabilities via telecommunication,’ and ‘a subcomponent of the broader area of telemedicine,’ while for Pitt (2017), telerehabilitation ‘falls under the broad umbrella term of telehealth.’ Nikolaev (2022) defined telerehabilitation (TR) as ‘the provision of rehabilitation services via telemedicine, i.e. using information and communication technologies, including video/teleconferencing, remote data collection equipment, telemonitoring, computers, mobile phones, robotics devices, exergames, and virtual reality (VR) tailored at individuals with disabilities, their families, clinicians, supervisors, and the community.”

This scoping review used a broad approach to assimilate published literature, providing a wide-ranging picture of the currently-used terms and synonyms for ‘consultations between remote clients and healthcare providers’ in stroke rehabilitation. It highlights the lack of consistency in the use of terminology in the field of digital stroke rehabilitation, and also demonstrates the heterogeneity in the underlying technological concepts and location contexts.

Based on the analysis of the included publications, there are still no universally accepted terms and definitions in relation to countries, languages or professional groups, but there is a clear preference for the term ‘telerehabilitation.’ Though a variety of terms are used, and the ontological location varies, the WHO taxonomy aims to provide a ‘bridging language’ to ensure understanding between different actors. However, among the health professions, the level of conscious awareness of this taxonomy, as well as the uniform understanding of the ‘umbrella term’ ‘telerehabilitation’ for digital stroke rehabilitation, are generally low to non-existent. Exchange and networking among the therapeutic professions, as well as the establishment of specific definitions of DHI, will be necessary to reach a common understanding. Referencing the WHO taxonomy would also ensure a place for telerehabilitation within telemedicine, which in turn would facilitate its definition.

Figure 9:

In this way, it becomes clear that the original definition by Rosen (1998) is sufficient and leaves room for different concepts of technology and different contexts of location. Definitions published before 2009 that include underlying technological concepts are usually limited to image-based concepts, such as audiovisual systems (e.g., videoconferencing solutions); sensor-based and VR-based concepts were not widely available before 2009.

Some limitations to our scoping review are worth noting. First, scoping reviews have inherent limitations due to their broad focus and the fact that they include protocols, papers and reviews, which results in duplication of information. Second, other reviewers may have coded differently. Finally, the results only apply to therapeutic publications on DHI in stroke rehabilitation, and therefore may lack transferability to other areas and clinical diagnoses.

In conclusion, there has been growing use of the term ‘telerehabilitation’ to describe ‘consultations between remote clients and health care providers’ in papers by allied health professionals that discuss stroke. While highlighting the lack of consistent usage and definition within the included papers, this review summarised existing terms, synonyms and definitions and developed a conceptual map of underlying technological concepts across locational contexts.