This scoping review was conducted according to the Joanna Briggs Institute guideline for scoping research and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement for reporting scoping (12). The study protocol was pre-registered on the Open Science Framework Platform.¹

Searches were conducted between January 2018 and September 2024 using a specified search string (Table 1; Supplementary material). After an initial search, a comprehensive search strategy was developed and applied to MEDLINE via PubMed, the Cochrane Central Register of Controlled Trials (CENTRAL), the Social Sciences Citation Index (SSCI), and the International Clinical Trials Registry Platform (ICTRP).

All RCTs and CCTs published in peer-reviewed journals enrolled severe stroke patients aged ≥18 years undergoing a rehabilitative intervention (e.g., physiotherapy, speech-language therapy, occupational therapy, nursing, and neuropsychology) were included. Articles were excluded if they used pharmacological and surgical interventions, non-invasive brain stimulation and complementary or alternative medicine interventions. If one or more secondary analyses were published, it was checked whether the primary study had already been included, if not, the first publication of a secondary analysis was included.

The Rayyan management software (Rayyan Systems Inc., Cambridge, MA 02142, United States) was used to select the included articles (13). Two reviewers (KR, LK) independently screened titles and abstracts to exclude those that did not meet the inclusion criteria. In a second step, the same two researchers independently performed a full-text screening of the remaining studies etiology. Disagreements during the entire process were discussed with a third researcher (NS) until consensus was reached.

The data from each included study was extracted by KR and HB, using a data extraction framework. Conflicts were resolved by discussion with NS. According to the definition of scoping reviews, the methodological quality of the included studies was not evaluated. Following the JBI methodology for scoping reviews, a formal appraisal of methodological quality was not required (14).

Stroke severity measurement tools used in the included studies were grouped according to the International Classification of Functioning, Disability and Health into body function and body structures, activities, and participation (15).

Out of 1,004 articles screened, 35 articles were included. Geographically, most of the studies were conducted in Europe (38%), followed by America (24%), Asia (24%), and Australia (14%). Participants were recruited in various settings, which were categorized into three groups. Starting with the clinical setting (n = 15) and the rehabilitation setting (n = 14). The term “non-clinical/rehabilitative setting” (n = 10) was used to categorize various settings—such as laboratory or community settings—that did not align with either of the two primary categories. In two studies (16, 17) participants were recruited in two settings. In the studies by Mulder et al. (18) and Sakakibara et al. (19), participants were recruited from three different settings. Salazar et al. (20) did not provide any information about the setting.

Within these studies, participants were mostly included during the early subacute phase (n = 13), followed by chronic stroke phase (n = 8), hyperacute stroke phase (n = 8), late subacute (n = 3), and acute (n = 2). For two of these studies, the stroke phase was not specified.

Most interventions involved physical activation, including specific exercises, training programs, or early mobilization. Additionally, many interventions incorporated robotic-assisted technologies or other digital health solutions, such as mobile applications, health platforms, or virtual reality. Several studies implemented transcranial stimulation and brain-computer interfaces as part of the intervention. A smaller proportion received video-based education or adherence-enhancing strategies. Most of the control group received conventional therapy, standard hospital care, and home exercise program for the clinic or rehabilitation facility (Table 2).

Eleven different measures were used to assess stroke severity (Table 3). Most studies used the NIHSS (n = 14), followed by modified Rankin Scale (n = 6). The Fugl-Meyer-Assessment for the upper extremity (FMA-UL), and the original Fugl-Meyer-Assessment (FMA), were each used four times to address stroke severity. The Barthel Index (BI) was used in three studies, and the Functional Ambulation Categories (FAC) and the Los Angeles Motor Scale (LAMS) was used in one study. Six studies included two measures to assess stroke severity. The NIHSS was used in seven studies to identify the hyperacute phase.

The original FMA was used in all three settings, mainly during the chronic phase, as well as the combination of FMA-UL and the Fugl-Meyer-Assessment of the lower extremity (FMA-LL) (n = 3) and once during the early subacute phase. In contrast, the FMA-UL was used twice in a rehabilitation setting during early subacute and late subacute.

Most of the measures can be assigned to a single ICF level. The NIHSS, Oxfordshire Community Stroke Project Classification (OCSP), and Los Angeles Motor Scale (LAMS) correspond to body function and impairment. Only the FMA as well as FMA-UL and FMA-LL, and the Functional Independence Measure (FIM) motor items cover both body function and impairment as well as the activity level. The remaining five measures are classified under the activity level.

Figure 2 illustrates the relationships among the three domains—“setting,” “measurement,” and “stroke recovery phase”—using connections of varying widths, where the thickness of each connection reflects the number of shared elements between the domains.

Study protocols provided different cut-off scores to assess severe stroke (Figures 3–5). For the NIHSS, the range for severe stroke was >5 and <20 (21) to 21–24 (22, 23). For the FMA the cut-off score applied was < 25 (24) and <50 (17). For the FMA-UL, the cut-off score was <30 (25) or ≤ 21 (26). For the BI, a cut-off score of ≤ 40 (27) or <30 (28) was used to assess severe stroke. In 11 studies (16, 20, 29–37), the cut-off scores for the evaluation of the severity of stroke were not specified.

This is the first study about the realization of stroke measures focusing on assessing severe stroke and the used cut-off scores. One strength of this review is the comprehensive search strategy specific to non-medical therapeutic interventions in stroke rehabilitation. Furthermore, the research team provides a diverse educational/professional background in treating severe stroke patients.

A reason for the limited number of search results was the inclusion of CCT and RCT study types. Because there is extensive research in the neurorehabilitation of stroke, the quality of these studies also provided the opportunity to include studies that may be of interest for guideline recommendations.

This review’s wide range of measures reflects the diversity of existing tools for assessing stroke severity. These results highlight the variety of measures used in research and those used in clinical practice to evaluate severe stroke.