Study setting {9}
This study takes place in different academic and rehabilitation hospitals in the Canton of Zurich, Switzerland.
Eligibility criteria {10}
Inclusion criteria
Participants fulfilling all the following inclusion criteria are eligible for the study:
-
Adults (≥ 18 years) with chronic stroke (≥ 6 months post-stroke, ischemic or haemorrhagic, (45))
-
Able to stand for 3 minutes and walk 10 meters, Functional Ambulation Category (FAC) ≥ 3
-
Able to follow a two-stage command
-
Able to give informed consent as documented by signature
Exclusion criteria
Participants fulfilling one or more of the following exclusion criteria are not eligible for the study:
-
Unable or not willing to give informed consent
-
Having been diagnosed with other neurological diseases (e.g. Parkinson’s Disease, multiple sclerosis), except cognitive deficits or dementia
-
Clinical contra-indications for the study intervention
-
Unable to follow the study intervention or the test for the primary endpoint (MoCA), e.g. due to a neglect, aphasia or other language problems
-
Overlapping enrolment in another clinical trial
Who will take informed consent? {26a}
Written informed consent is obtained by movement scientists and therapists, who are part of the study team and were trained for this task. Eligible, potential participants are provided with detailed study information in oral and written form. After at least 24 hours for consideration, interested potential participants are invited to a first study appointment, where they first have an informed discussion with a member of the study team to clarify remaining questions and again outline the benefits, risks and their rights associated with the study. Written informed consent is then obtained from those willing and able to participate in the study.
Additional consent provision for collection and use of participant data and biological specimens {26b}
The consent includes collection and use of individually identifiable participant data for the study procedures (through members of the study team only) and for audit trials by the cantonal ethical committee and other local authorities. Moreover, participants consent that their de-identified data is used for the analysis and publication of the study results. This study does not involve collecting biological specimens for storage.
Interventions
Explanation for choice of comparators {6b}
Participants in the control group continue with their usual care and receive no additional intervention. They are called once a week to align for contact to the study team and to gather their physical and cognitive activity data (see ‘Intervention and further activity outcomes’). This comparator was chosen as the aim of this study is to determine the effect of additional motor-cognitive exergame training. This aim was based on systematic reviews, which recommend the application of exergame training in addition to usual care, aiming at increasing the amount of rehabilitation offered to patients (33, 38, 46). Exergames that contain aspects of virtual reality have shown to be able and significantly effect on Body Structure/Function and Activity level outcomes, including improvements in cognitive function and there is, therefore, evidence supporting the use of such interventions as an adjunct for stroke rehabilitation (47, 48).
Intervention description {11a}
The intervention group receives concept-guided, personalized, motor-cognitive exergame training additionally to usual care. The personalized motor-cognitive training for stroke (PEMOCS) concept, which guides the intervention, determines on the one hand the training dosage with regards to frequency, intensity, time, type, and volume of the exercises, which is based on recommendations from scientific literature for motor-cognitive and exergame trainings in chronic stroke and healthy older adults. Participants train twice a week for twelve weeks (intervention period, see Fig. 1). Training sessions last between 30 (beginning) and 40 (end) minutes, progressing in duration for two minutes every second week and resulting in 840 minutes total planned training time (see Table 1). The training is performed using the exergame device Dividat Senso (Dividat AG, Schindellegi, Switzerland, for a detailed description of the device see (49)). The Dividat Senso consists of a TV screen and a pressure-sensitive plate, as well as a handrail on three sides to provide security to trainees (Fig. 2). Participants perform whole-body movements including steps, body weight shifts and running movements on the plate to control the games. The system provides real-time feedback on the gamer’s performance, including visual, auditory, and tactile cues facilitating the interaction of the participant with the video games. These video games target different cognitive functions (incl. attentional, executive, memory, and visuospatial functions). Via an online platform, personalized training programs can be created within the Dividat training system.
The PEMOCS concept is designed to provide personalized progression and variability in training considering principles for neuroplasticity, motor learning, and training (50–52). It was developed based on Gentile’s Taxonomy for motor learning (53) and tested for its feasibility in the target population (49). In accordance with the Modified Consensus on Exercise Reporting Template (CERT) for Therapeutic Exercise Interventions (54, 55), the PEMOCS training concept will be published elsewhere (PEMOCS: Theory Derivation of a Training Concept for Personalized Motor-Cognitive Training in Chronic Stroke, Huber et al., manuscript in preparation). In short, motor and cognitive tasks of the exergame training are allocated to different difficulty levels along a skill-progression scheme with three dimensions. Based on the participants’ subjective ratings of their perceived motor-cognitive task difficulty and their perceived performance (see ‘Intervention outcomes’), progression through the difficulty levels is determined individually for each participant. Therefore, the training is personalized within a standardised progression scheme. Additionally, variability rules ensure variation in training tasks.
Table 1
Intervention – training variables
Time point | Frequency [per week] | | Time [min / session] | | Volume / Week [min] | |
Week 1–2 | 2 | x | 30 | = | 60 | |
Week 3–4 | 2 | x | 32 | = | 64 | |
Week 5–6 | 2 | x | 34 | = | 68 | |
Week 7–8 | 2 | x | 36 | = | 72 | |
Week 9–10 | 2 | x | 38 | = | 76 | Total Volume [min] |
Week 11–12 | 2 | x | 40 | = | 80 | 840 |
Table 1: Training variables of the intervention, describing the frequency, time, volume per week, and total volume of the concept-guided, personalized, motor-cognitive exergame training.
Criteria for discontinuing or modifying allocated interventions {11b}
The intervention is discontinued in case of withdrawal of the informed consent, participant request or if the health status of the participant or any harm does not allow the continuation of the intervention. Minor individual modifications of the planned intervention are possible (e.g. using a training aid for additional stability) if following the training concept is possible.
Strategies to improve adherence to interventions {11c}
All sessions are supervised one-to-one by a trained movement scientist, who engages the participant in the study procedures and encourages complying with the set appointments. Additionally, game preferences of each individual participant are considered for applying variability in the intervention while considering the training concept to foster motivation and fun during the gaming sessions. Adherence to training sessions and compliance with scheduled training time are recorded.
Relevant concomitant care permitted or prohibited during the trial {11d}
Concomitant care (usual care) including other therapies is allowed in both groups during the whole study. The content, time, and intensity of usual care as well as of general physical and cognitive activities are recorded through structured interviews every week in both groups (see ‘Intervention and further activity outcomes’).
Provisions for post-trial care {30}
N/A, as there are no disadvantages likely to arise from the intervention.
Outcomes {12}
Primary and secondary outcomes are collected at three time points (T0 – week 0, T1 – week 12, T2 – week 24; see Table 2, Figs. 1&3) by cognitive and motor assessments as well as a health-related-quality-of-life questionnaire. At the baseline measurement (T0), participant characteristics and baseline factors are recorded. Intervention and activity outcomes are collected during the intervention (T0-T1) and the follow-up (T1-T2) periods, respectively. All variables will be aggregated as means or medians, depending on the distribution of the data, and the analysis metric for all will be final values (see ‘Statistical methods’). An overview of all outcome variables is presented in Tables 3 & 4.
Table 2
Overview over outcome recording
Time (weeks) | 0 | + 1–12 | + 12 | + 13–24 | + 24 |
Time point | Baseline (T0) | Intervention Period | Post-Intervention (T1) | Follow-up Period | Follow-up (T2) |
Study staff | Assessor | Training supervisor | Blinded assessor | Training supervisor | Blinded assessor |
Baseline factors: BQ, NIHSS, FAC, mRS, BBS, FM-LE | + | | | | |
1° Endpoint: MoCA | + | | + | | + |
2° Endpoints: SIS 3.0, SRT, TMT, Stroop, NBT, MRT, TUG(-Cogn), 10MWT, OWA | + | | + | | + |
Intervention endpoints: CPL, ADH, PTD, PP | | + (Intervention group only) | | | |
Activity endpoints: UC, PCA | | + | | + | |
Table 2: Overview over outcome recording. ADH: Adherence; BBS: Berg Balance Scale; BQ: Baseline Questionnaire, incl. see ‘Baseline factors’; TUG(-Cogn): (cognitive dual-task) Timed-up-and-go test; CPL: Compliance; FAC: Functional Ambulation Category; FM-LE: lower-extremity component of Fugl-Meyer Assessment; MoCA: Montreal Cognitive Assessment; mRS: modified Rankin Scale; MRT: Mental Rotation Test; NBT: N-back Test; NIHSS: National Institute of Health Stroke Scale; OWA: Outdoor Walking Assessment; PCA: Physical and Cognitive Activities; PP: Perceived Performance; PTD: Perceived Task Difficulty; SIS 3.0: Stroke Impact Scale; SRT: Simple Reaction Test; Stroop: Stroop Interference Test; TMT: Trial-Making Test; UC: Usual Care; 10MWT: 10-Meter Walk Test.
Table 3: Primary and secondary outcomes
Test
|
Outcome variables
|
Definition
|
Primary outcome
|
MoCA
|
Total score (0-30)
|
= Sum of domain scores + 1 point for participants with ≤ 12 years of education and a MoCA score < 30
|
Secondary outcomes
|
SIS 3.0
|
- Total score (0-100)
|
= Sum of domain scores
|
|
- Perceived recovery (0-100%)
|
Collected by visual analogue scale
|
|
- Domain scores (0-100%): strength, memory / thinking, emotion, communication, ADL / IADL, mobility, hand function, participation
|
|
SRT
|
Of all six conditions:
|
|
|
- Reaction time
|
= Logarithmic mean of trial reaction times
|
|
- Missed
|
Number of non-answered stimuli within 1500ms
|
|
- Mistakes
|
Number of reactions upon no stimulus
|
TMT A&B
|
- Time A, Time B
|
Processing time of each sub-test
|
|
- Mistakes A, Mistakes B
|
Number of incorrect touches in each sub-test
|
|
- B:A ratio
|
|
Stroop
|
For reading and naming:
|
|
|
- Interference tendency
|
= Median RT interference – median RT baseline
|
|
Of all four conditions:
|
|
|
- Reaction time
|
= Median reaction time of all trials
|
|
- Mistakes
|
Number of false reactions
|
NBT
|
- Correct
|
Number of correct responses (stimulus inquired a reaction and participant responded).
|
|
- Omissions
|
Number of missed responses (stimulus inquired a reaction but participant did not respond).
|
|
- Mistakes
|
Number of incorrect responses (stimulus inquired no reaction but participant responded).
|
|
- Reaction time (correct)
|
= mean reaction time of correct responses
|
|
- Reaction time (mistakes)
|
= mean reaction time of incorrect responses
|
MRT
|
- Accuracy
|
|
|
- Reaction time
|
= mean reaction time of correct responses
|
TUG
|
- Time TUG
|
= mean time of the 3 motor single-task trials
|
|
- Type of cognitive task
|
Serial subtraction or Verbal fluency
|
|
- Correct response rate single-task (CRRsingle)
|
= mean CRR of the 3 cognitive single-task trials
|
|
|
CRR
|
|
- Time TUG-Cogn
|
= mean time of the 3 dual-task trials
|
|
- Correct response rate TUG-Cogn (CRRdual)
|
= mean CRR of the 3 dual-task trials
|
|
- Motor dual-task effect (DTE%motor)
|
|
|
- Cognitive dual-task effect (DTE%cognitive)
|
|
10MWT,
|
- Time
|
= mean time for 10m of the three trials
|
preferred
|
- Gait speed (preferred)
|
= mean gait speed
|
|
|
Instruction: “Walk at a comfortable speed.”
|
|
- Cadence
|
= mean steps / min
|
|
- Stride length
|
= mean stride length
|
|
- Stride length variability
|
= mean stride length variability
|
|
- Stride time
|
= mean stride time
|
|
- Stride time variability
|
= mean stride time variability
|
|
- Double support time
|
= mean double support time
|
|
- Stance phase affected, unaffected
|
= mean stance phase
|
|
- Swing phase affected, unaffected
|
= mean swing phase
|
|
- Swing width affected, unaffected
|
= mean swing width
|
|
- Asymmetry index*
|
|
|
- Walk ratio**
|
|
10MWT,
|
- Time (fast)
|
= mean time for 10m of the three trials
|
fast
|
|
Instruction: “Walk as fast but safe as possible.”
|
OWA
|
- Gait speed (preferred)
|
= mean gait speed
|
|
|
Instruction: “Walk at a comfortable speed.”
|
|
- Cadence
|
= mean steps / min
|
|
- Stride length
|
= mean stride length
|
|
- Stride length variability
|
= mean stride length variability
|
|
- Stride time
|
= mean stride time
|
|
- Stride time variability
|
= mean stride time variability
|
|
- Double support time
|
= mean double support time
|
|
- Stance phase affected, unaffected
|
= mean stance phase
|
|
- Swing phase affected, unaffected
|
= mean swing phase
|
|
- Swing width affected, unaffected
|
= mean swing width
|
|
- Asymmetry index*
|
s. above
|
|
- Walk ratio**
|
s. above
|
Table 3: Overview of primary and secondary outcomes with definitions. TUG-Cogn: cognitive dual-task timed-up-and-go test; MoCA: Montreal Cognitive Assessment; MRT: mental rotation test; NBT: N-back test; OWA: outdoor walking assessment; SIS 3.0: stroke impact scale; SRT: simple reaction test; TMT: trail-making test; TUG: timed-up-and-to test; 10MWT: 10-meter walk test. *The asymmetry index is calculated according to the presented formula not considering affected body side. Allocating left and right in the formula has an impact on the algebraic sign of the asymmetry index, not, however, on its absolute value. As absolute values will be reported for comparability of the asymmetry indices of left- and right-affected patients, the formula can be used as reported (115). **(138)
Table 4: Intervention and further activity outcomes
Intervention outcomes
|
Compliance rate (overall)
|
|
Compliance rates in each week (1-12)
|
Adherence rate (overall)
|
|
Adherence rates in each week (1-12)
|
Reasons for not attending or aborting a training session
|
|
Perceived motor-cognitive task difficulty (overall)
|
Collected by visual analogue scale
|
Perceived m-c-task difficulty in each week (1-12)
|
Perceived performance (overall)
|
Collected by visual analogue scale
|
Perceived performance in each week (1-12)
|
Further Activities: Usual Care
|
|
- Frequency intense physical therapy (PT)
|
= mean frequency per week of intense PT
|
- Volume / week intense PT
|
= mean volume per week of intense PT
|
- Total volume intense PT
|
= sum of the volume of all intense PT
|
- Frequency moderate PT
|
= mean frequency per week of moderate PT
|
- Volume / week moderate PT
|
= mean volume per week of moderate PT
|
- Total volume moderate PT
|
= sum of the volume of all moderate PT
|
- Total volume PT
|
= sum of the volume of all PT
|
- Types of PT
|
|
- Frequency intense cognitive therapy (CT)
|
= mean frequency per week of intense CT
|
- Volume / week intense CT
|
= mean volume per week of intense CT
|
- Total volume intense CT
|
= sum of the volume of all intense CT
|
- Frequency moderate CT
|
= mean frequency per week of moderate CT
|
- Volume / week moderate CT
|
= mean volume per week of moderate CT
|
- Total volume moderate CT
|
= sum of the volume of all moderate CT
|
- Total volume CT
|
= sum of the volume of all CT
|
- Types of CT
|
|
- Total volume other therapies
|
= sum of the volume of all other therapies
|
- Types of other therapies
|
e.g. massage
|
Further activities: general
|
|
- Frequency intense physical activity (PA)
|
= mean frequency per week of intense PA
|
- Volume / week intense PA
|
= mean volume per week of intense PA
|
- Total volume intense PA
|
= sum of the volume of all intense PA
|
- Frequency moderate PA
|
= mean frequency per week of moderate PA
|
- Volume / week moderate PA
|
= mean volume per week of moderate PA
|
- Total volume moderate PA
|
= sum of the volume of all moderate PA
|
- Total volume PA
|
= sum of the volume of all PA
|
- Frequency intense cognitive activity (CA)
|
= mean frequency per week of intense CA
|
- Volume / week intense CA
|
= mean volume per week of intense CA
|
- Total volume intense CA
|
= sum of the volume of all intense CA
|
- Frequency moderate CA
|
= mean frequency per week of moderate CA
|
- Volume / week moderate CA
|
= mean volume per week of moderate CA
|
- Total volume moderate CA
|
= sum of the volume of all moderate CA
|
- Total volume CA
|
= sum of the volume of all CA
|
- Sedentary time per day
|
= mean sedentary time per day
|
Table 3: Overview of intervention and further activity outcomes. CA: cognitive activities; CT: cognitive therapy; PA: physical activities; PT: physical therapy.
Baseline factors
At baseline, the following information is recorded to describe the study population (see Table 2):
-
Demographics (collected via baseline questionnaire): age, sex, years of education, marital status.
-
Other characteristics (collected via baseline questionnaire): weight, height, handedness.
-
Stroke diagnostic details (collected via baseline questionnaire): number of strokes, time point / type (ischemic, haemorrhagic) / lesion site / side and location of initial paresis of the (most recent) stroke, initial (if available) and current (at baseline) National Institutes of Health Stroke Scale (NIHSS, (56)).
-
Clinical characteristics A (collected via baseline questionnaire): Comorbidities (Charlson Comorbidity Index, CCI, (57)), modified Rankin Scale (mRS, (58)).
-
Clinical characteristics B (collected by the assessor at the baseline measurement): Functional Ambulation Category (FAC, (59)), Lower-Extremity component of the Fugl-Meyer Assessment (FM-LE), Berg Balance Scale (BBS, (60)).
Primary outcome
The primary outcome is global cognitive functioning measured by the total score of the Montreal Cognitive Assessment (MoCA) (61). The MoCA is composed of several tests assessing different cognitive domains including attention, executive functions, working memory, short-term memory recall, visuospatial skills, and orientation (44). The MoCA has been successful in detecting cognitive decline (62), showed good reliability in stroke patients and healthy older adults (63–65), as well as fair to good responsiveness and validity (66) in chronic stroke patients (67). The maximum achievable score is 30 points, where more points represent better cognitive functioning. For individuals with twelve or less years of education and a total score < 30, an additional point is added (61). A MoCA score below 24 points indicates mild cognitive impairment in individuals after stroke (44, 68, 69) and an improvement of 1.22 points was found to be a clinically relevant change (67).
Global cognitive functioning, measured by the total score of the MoCA, covers the cognitive domains typically impaired after stroke. Further cognitive tests examining specific cognitive domains/functions will be implemented as secondary outcomes (70).
Secondary outcomes
Health-related quality of life and perceived recovery are assessed using the total score and the single domain scores of the Stroke Impact Scale [SIS 3.0, (71)] at T0, T1 and T2 (see Table 2, Figs. 1&3). The SIS 3.0 is a stroke-specific questionnaire assessing the self-reported health status on 5-point Likert scales (66). It encompasses eight domains (strength, memory/thinking, emotion, communication, ADL/IADL, mobility, hand function, and participation) and a visual analogue scale, where the perceived state of recovery is rated (0–100%) (66). The final score lies between 0 and 100, where a higher score indicates better health-related quality of life. The German SIS (DE-SIS, translated and cross-culturally adapted) was found reliable and valid for the use in German-speaking stroke survivors (72). The SIS 3.0 is administered as online or paper survey or, in case necessary for the participant to be able to answer the questions, in a structured interview led by the assessor.
Secondary cognitive outcomes are collected by computer-based cognitive assessments performed within the Vienna Test System (VTS, Schuhfried GmbH, Mödling, Austria) at T0, T1 and T2 (see Table 2, Figs. 1&3). All these assessments are performed on a touch-screen computer, using either one button on the keyboard or the touch screen to answer the stimuli.
To assess alertness, a simple reaction test (SRT) (“WAFA” within the VTS) is applied. The SRT is a reliable and valid neuropsychological assessment for alertness (73–75). It composes of six tests with visual and auditory stimuli, of which two evaluate intrinsic alertness (participant has to react to an appearing stimulus as fast as possible), two evaluate crossmodal-phasic alertness (a crossmodal warning stimulus precedes the actual stimulus and the participant has to only react to the actual and not the warning stimulus), and two evaluate unimodal-phasic alertness (a unimodal warning stimulus precedes the actual stimulus and the participant has to only react to the actual and not the warning stimulus) (75). The stimuli are answered by pressing a button on the touch screen of the test computer.
Processing speed and cognitive flexibility, a sub-domain of executive functions, are assessed using the trail-making test (TMT) (“TMT - Langensteinbacher Version” within the VTS), which is a widely used, reliable and valid neuropsychological assessment (76–78). The TMT consists of two parts; TMT-A assesses general information-processing speed, it asks to connect rising numbers (1–25) as fast as possible. TMT-B is used to test cognitive flexibility, the task is to connect rising numbers and letters alternatingly (79). The tasks are completed by pressing the numbers and letters on the touch screen of the test computer.
To assess interference inhibition, another sub-domain of executive functions, the Stroop Interference test (“STROOP” within the VTS) is used (80). The Stroop test is a widely used, reliable and valid neuropsychological assessment testing the ability to inhibit the reaction to a more dominant stimulus in favour of the inquired reaction to a less dominant stimulus (81). This assessment contains four sub-tests: two baseline and two interference conditions. In the first baseline condition, colour words are presented in grey font and the participant must select the correct colour (baseline reading). In the second baseline condition, coloured bars are presented, and the participant must select the correct colour (baseline naming). In the interference conditions, the colour words are shown in coloured fonts and the participant has to either select the correct colour of the word (interference reading) or of the font (interference naming) (82). The tasks are completed by pressing the colour fields or words on the touch screen of the test computer.
To assess working memory and related cognitive functions, the N-back test (NBT) (“NBV” within the VTS) is used, which is a widely used, reliable and valid neuropsychological test (83–85). The participant is presented a row of letters and has to decide upon every letter whether it corresponds to the one shown N letters earlier (86). In this study, the test condition N = 2 is used. The stimuli are answered by pressing a button on the keyboard of the test computer.
Mental rotation ability, a sub-domain of visuospatial functions, is assessed using the mental rotation test (MRT) (“3D” within the VTS), which is based on the paradigm by Shepard and Metzler (87). It determines the ability to mentally rotate abstract objects and has been used in stroke patients before (49, 88, 89). Each item consists of a figure composed of a number of blocks. The participant has to imagine how the arrangement of the blocks looks when viewed from another perspective and choose the correct 2D-view from a selection of four possible solutions (90). The solutions are selected by pressing tick boxes on the touch screen of the test computer.
The Timed-up-and-go test (TUG), a reliable and valid assessment in stroke patients (91, 92), is conducted to analyse changes in mobility and dynamic balance. Participants are instructed to perform the TUG “as fast and safely as possible”. Time is measured from the moment the participant’s back leaves the backrest of the chair until it touches the backrest of the chair again (93). To assess mobility under a dual-task conditions and dual-task effects, the TUG-Cognitive is also performed (94). The TUG-Cognitive (TUG-Cog) is a reliable and valid assessment of dual-task mobility in stroke patients and healthy adults (93, 94). A cognitive task is first executed in single-task mode during 60s while being seated on a chair. After this, both tasks, the TUG and the cognitive task are performed simultaneously, while participants are instructed to not prioritise one over the other. As cognitive task, serial subtraction of 3 from a random number between 50 and 100 (not in the row of three), will be used in participants who are able to complete this task (94, 95). Participants, who are not able to accomplish serial subtraction, perform a verbal fluency task instead; naming nouns from categories (e.g. fruits, animals, cloths) starting with a specific letter (94, 95). For all single- and dual-task trials, a familiarisation trial is performed before executing three test trials each. The TUG and TUG-Cog are conducted under laboratory conditions at the participating study centres.
Participants perform a 10-meter walk test (10MWT) and an outdoor walking assessment (OWA) using inertial gait sensors (Physilog® sensors, Gait Up SA, Lausanne, Switzerland) to analyse temporal and spatial gait parameters. The 10MWT has been found reliable and valid in stroke patients (96). The 10MWT is performed according to the protocol by Cheng et al. (96), where participants walk 14 meters, where only the middle 10 meters are timed. The participant starts walking at the 0-meter mark and the stopwatch is started as soon as the first foot crosses the 2-meter mark and stopped again when the first foot crosses the 12-meter mark, while the participant continues walking to the 14-meter mark. At first, a familiarisation trial is performed followed by three trials at comfortable walking and three trials at fast walking speed. Participants are instructed to “walk at a comfortable speed” and “walk as fast but safely as possible”, respectively. Participants use their usual walking aid if needed.
The OWA is a 400-meter walk following an outdoor route without stairs (97). At each study centre, a suitable route nearby was pre-defined and all participants follow this same route. Participants use their usual walking aid if needed and wear proper footwear for an outdoor walk. They are instructed to “walk at a comfortable speed, as if they were on a stroll”. To ensure safety, participants are accompanied by two investigators.
Intervention and further activity outcomes
Compliance and adherence to the trainings and the reasons for not attending or aborting a training session are recorded during the intervention period (T0-T1, see Table 2, Figs. 1&3). Additionally, participants in the intervention group are asked to rate the motor-cognitive task difficulty of the training tasks (perceived task difficulty, PTD) and their motor-cognitive performance (perceived performance, PP) in every training session (during T0-T1, see Table 2, Figs. 1&3). Visual analogue scales (VAS) in the eCRF based on the cognitive load theory (98) and the NASA-TLX [National Aeronautics and Space Administration-Task Load Index, (99)] are used to collect these ratings (Additional file 1). PTD and PP are expressed as percentage values where higher percentages stand for more difficult tasks and better performance, respectively. The intervention outcomes are summarized in Table 4.
All participants are interviewed weekly throughout the whole study (T0-T2, see Table 2, Figs. 1&3) regarding the dose (frequency and time) and content (intensity and type) of moderate to intense physical and cognitive activities, which they perform as part of their usual care or in their leisure time. Definitions for moderate to intense activities are based on the World Health Organisation’s (WHO) 2020 Guidelines on Physical Activity and Sedentary Behaviour (100). The interviews are done using a structured questionnaire (Additional file 1), which implies the FITT-VP principles (40) and the TIDieR checklist [Template for Intervention Description and Replication, (101)]. The further activity outcomes are summarized in Table 4.
Participant timeline {13}
Each participant is involved for approximately 24 weeks (12 weeks intervention period, 12 weeks follow-up period, Figs. 1&3). Participants are contacted and screened for eligibility by their therapist, physician and / or the study team. Full written and oral study information including aims, risks and benefits associated with this study is provided by investigators of the study team, who are trained for this task. Before any study-related procedures start, all participants sign written informed consent. After that, participants first attend the baseline measurement (T0), where baseline data are collected, and all outcome assessments (primary and secondary outcomes) are performed (Figs. 1&3). Subsequently, the participant is allocated randomly to one of the two study arms. Randomization for each participant is run in Research Electronic Data Capture (REDCap) by an investigator other than the blinded assessor. Participants allocated to the intervention group thereafter attend concept-guided, personalized motor-cognitive training, twice a week for twelve weeks (T0-T1), while continuing their usual care and general activities. Participants allocated to the control group receive no additional intervention during the same twelve weeks (T0-T1) and continue their usual care and general activities. After completion of the intervention period, all participants attend the post-intervention measurement (T1), where all primary and secondary outcome assessments are repeated (Figs. 1&3). During the subsequent 12-week follow-up period (T1-T2), participants in both groups receive no additional intervention and continue their usual care and general activities. At the end of this period, all participants attend the follow-up measurement (T2), repeating the procedures from T1.
Sample size {14}
A sample size of 38 participants (approx. 19 per group) was calculated for this study. The sample size calculation was based on systematic reviews investigating the effects of motor-cognitive training and exergames on cognitive functions. Stanmore et al. included studies with any population (including five out of seventeen studies with stroke or other neurological patients) and found a small to medium effect for global cognitive functioning (SMD = 0.44, p = 0.001) and several small to large effects for cognitive domains including executive functions, processing speed and visuospatial skills (0.26 ≤ SMD ≤ 0.90, p < 0.05) (37). Five further reviews included studies with older adults and found small to large effect sizes for global or overall cognitive functioning and cognitive domains including attention, executive functions, learning and memory, and processing speed (0.30 ≤ SMD ≤ 1.37, p < 0.05) (39, 102–105). Based on this evidence, a small to medium effect on global cognitive functioning is anticipated for the planned study (f = 0.21). The sample size was estimated using G*Power, entering the following parameters into the mask for a two-way mixed ANOVA; α-level = 0.05, power = 0.80, number of groups = 2, number of measurements = 3, correlation among rep measures = 0.5, nonsphericity correction = 1.
Dropouts will be replaced by post-recruitment until the planned sample size is achieved. Based on the recent feasibility study (49) and comparable literature (106–108), a dropout rate of 10–20% can be expected.
Recruitment {15}
The PEMOCS study is a single-blind randomized control trial (RCT) with chronic stroke patients recruited from hospitals and rehabilitation centres in the Canton of Zurich, Switzerland. Participants are recruited by therapists and physicians during therapy sessions and stroke follow-up appointments, by flyers on the ward, and by contact of the study team in case of provided general consent. Regular contact between the recruiters and study team meetings should help to maintain an acceptable recruitment rate.
Assignment of interventions: Allocation
Sequence generation {16a}, Concealment mechanism {16b}, and Implementation {16c}
The randomisation is stratified by cognitive status (cognitive impairment absent or present, determined by a MoCA score ≥ 24 or < 24, respectively (44, 68, 69)) and by sex (female or male, (109–111)). Both stratifying variables are allocated 1:1 to both groups. Participants are randomized using REDCap, the same tool as utilised for eCRF keeping (see ‘Data management’, (112, 113)). To perform the randomisation in REDCap, a pre-defined randomisation list in the form of an excel document needs to be uploaded onto the platform, which is then being used by the software. Instructions by REDCap show, how this list must be structured to provide the desired allocation ratio and stratification (112, 113). These instructions were followed by the randomisation-list creator, a person otherwise not involved in the study. This way it was ensured that no member of the study team knows the allocation sequence. After creating the list, the randomisation-list creator encrypted the excel document containing the list with a password. Both, the list and the password, are stored in a secure place, where the investigators of the study team have no access. Moreover, REDCap allows setting individual user privileges (112, 113). The user rights to access and view the randomisation setup feature were removed from all investigators of the study team before the final randomisation list was uploaded onto the REDCap platform.
Study investigators other than the blinded assessor enrol participants and allocate them to groups. A separate instrument in REDCap, which is invisible for the assessor through user privileges, is used for randomisation.
Assignment of interventions: Blinding
Who will be blinded {17a}
An assessor blinded to group allocation will perform all assessment. The assessor cannot access the randomisation tool in REDCap and is not involved in intervention procedures. To blind the data analyst, a de-identified data set will be used for analysis, which will not contain unique identifiers such as the study ID.
Procedure for unblinding if needed {17b}
N/A. The assessor is the only blinded member of the study team, who is, however, never alone with the participants. All assessment sessions are accompanied by a training supervisor, who knows the participants and their group allocation. Hence, in case any situation during a measurement session would require knowledge of group allocation, the training supervisor can handle it and the assessor does not need to be unblinded.
Data collection and management
Plans for assessment and collection of outcomes {18a}
All outcome variables are gathered in an eCRF (see ‘Data management’). Data entry for all outcome variables is checked by at least two trained study-team members (four-eyes-principle) to ensure data quality and correctness. All study-team members gathering data receive specific training for the relevant study procedures. Primary and secondary outcomes are collected at baseline (T0), post-intervention (T1) and follow-up (T2) measurements. The MoCA (primary outcome) is executed on paper following the instructions of the providers (mocacognition.com, (61)) and the data is transferred into the eCRF. All assessors obtain a MoCA certificate (mocacognition.com) for trained execution of the test before performing the assessment in the study. The SIS 3.0 is collected via a questionnaire filled out by the participants electronically in the eCRF or, in case not able to do so, on paper, and transferred to the eCRF by the investigators. All cognitive assessments apart from the MoCA are conducted within the Vienna Test System (VTS, Schuhfried GmbH, Mödling, Austria), a valid and reliable software for neuropsychological testing. All these assessments include written instructions, practice sets and are only started if the participant received clarifying responses on any questions regarding the test functionality. All cognitive outcome variables are obtained from the VTS result sheets and transferred into the eCRF. Outcome variables for the TUG and TUG-Cog are collected using a stopwatch and by noting correct answers on paper and directly entered into the eCRF. For the TUG, TUG-Cog and the 10MWT, practice trials are performed before the actual assessment to ensure clarity of the procedure. Outcome variables of the gait assessments are gathered using the Gait Up system (Gait Up SA, Lausanne, Switzerland) with Physilog® sensors (wearable standalone movement inertial sensors, 50 × 37 × 9.2 mm, 19 g). The Gait Up system provides quantitative, objective, and valid assessment of gait movement (114) presented on output sheets, from where the data are transferred into the eCRF. The first and last two gait cycles are excluded from the analysis to eliminate acceleration and deceleration (115). Baseline characteristics (see ‘Baseline factors’) are collected via a questionnaire filled out by the participants electronically in the eCRF or, in case not able to do so, on paper, and transferred to the eCRF by the investigators. The results of FMA-LE, FAC and BBS (see ‘Baseline factors’) are directly entered into the eCRF. ‘Intervention and further activity outcomes’ are collected within the eCRF throughout the intervention (T0-T1) and follow-up (T1-T2) periods, respectively. In all phases, deviations from the protocol are recorded in the eCRF to ensure traceability and the ability to exactly repeat the assessment procedures at T0, T1 and T2 for each individual participant.
Plans to promote participant retention and complete follow-up {18b}
During the follow-up period, participants in both groups are contacted once a week to inquire their usual care, general physical and cognitive activities (see ‘Intervention and activity outcomes’). This keeps them engaged in the study procedures and, therefore, promotes successful retention.
Data management {19}
An electronic case report form (eCRF) is kept for each enrolled participant using REDCap electronic data capture tools hosted at ETH Zurich (112, 113). REDCap (Research Electronic Data Capture) is a secure, web-based software platform designed to support data capture for research studies, providing 1) an intuitive interface for validated data capture; 2) audit trails for tracking data manipulation and export procedures; 3) automated export procedures for seamless data downloads to common statistical packages; and 4) procedures for data integration and interoperability with external sources (112, 113). This eCRF has been validated before enrolment of the first participant. Study team members who are authorized to enter or edit data in the eCRFs, receive a login to the REDCap study platform, and are listed with signatures in the trial master file (TMF) and the investigator site file (ISF). To assure that any authorized person, who may perform data entries and changes in the eCRF, can be identified, all entries / edits are recorded with name, date, and time. Data entry of the primary and secondary outcomes in REDCap is performed by one and double-checked by another investigator (verification, four-eyes-principle). Should any previously entered data need to be changed (e.g., because a mistake was identified during the verification), a reason must be given to proceed. eCRFs are kept current to reflect participant status at each phase during the study.
Study and participant data will be handled with uttermost discretion and are only accessible to authorized personnel who require the data to fulfil their duties within the scope of the study. Participants are coded and not identifiable in the eCRF or on any other study-specific documents. Appropriate coded identification (study ID) is used. Each study ID composes of four random letters or numbers, which are not related to any characteristics, or the time point of inclusion of the participants. The Sponsor will store the participant identification list in a secured and locked location. All study data are archived for ten years after study termination or premature termination of the study.
Confidentiality {27}
Personal information of potential and enrolled participants is kept confidential and only accessible for involved study-team members for study-related purpose. Data protection is kept according to current guidelines of the Swiss law. Participants, who have withdrawn from the study, can ask the deletion of their personal information at any time. All participants receive a study ID (a random sequence of four letters and numbers) not associable with their personal data. All study data is stored only with this ID and never related to any personal data. The key to decode study data is kept locked and only accessible for involved study-team members for study-related purpose(s). After completion of the study procedures, study data are archived according to Good Clinical Practice (GCP) guidelines for at least ten years.
Plans for collection, laboratory evaluation and storage of biological specimens for genetic or molecular analysis in the trial / future use {33}
N/A as no biological specimens are collected in this study.
Statistical methods
Microsoft Excel (Microsoft Corporation, 2016) will be used to aggregate and tabulate the data. All statistical analyses will be performed using RStudio open-source software (Bosten, USA, (116)) or SPSS Statistics (version 26 for windows; IBM, Chicago, IL, USA).
Statistical methods for primary and secondary outcomes {20a}
Distributions of all baseline factors, primary and secondary outcome variables will be checked with the Shapiro-Wilk test (117). Appropriate descriptive statistics will be obtained for all baseline factors and outcome variables (means and standard deviations for normally distributed data, medians and inter-quartile ranges for non-normally distributed data, frequencies for categorical data). Differences between groups in baseline factors will be evaluated using an independent t-test if assumptions for parametric testing are met or a non-parametric alternative otherwise. For categorical data, a chi-square test or a Fisher’s exact test will be used as appropriate (118).
Assumptions on the residuals of primary and secondary outcomes will be checked using the DHARMa package in R (119). Appropriate actions will be taken if one or more assumptions are not met. All primary and secondary outcomes will be analysed following the standard intention-to-treat (ITT, (120, 121)) principle using linear mixed-effects models (LMEM, lme4 package in R). Subject-specific random intercepts will account for within-subject correlations between time points. Follow-up scores (T2) of the outcomes will be the dependent variables of the models, while group (intervention vs. control, control being the reference), time (T0, T1, T2), and group x time interactions at T1 and at T2, respectively, will be included as independent variables (fixed effects). Baseline factors such as age, sex, and time since stroke will be considered as covariates. Missing data of whole measurement time points (i.e. of dropouts) will be accounted for with the “last observation carried forward” method. For single missing data points (e.g. if a participant did not perform an assessment at one time point or a technical issue produced missing data at one time point), however, no data imputation will be performed as LMEMs can be fitted even if some outcome data are missing (117, 122). For the outdoor walking assessment (OWA), high occurrence of missing data is expected, as ability to walk 400 meters is not covered by the eligibility criteria and, varying weather conditions on the measurement days may interfere with the assessment. Therefore, for the OWA outcome parameters, data sets of participants who did not perform the OWA at one or several time points will be excluded from the analysis. Significance will be set to p < 0.05. Effect sizes will be calculated as r (Bravais-Person correlation coefficient), and interpreted as small (r < 0.3), medium (r < 0.5), and large (r ≥ 0.5) (117).
Methods for analysis of Intervention and further activity outcomes
Mean / median adherence and compliance rates with standard deviations / inter-quartile ranges will be reported overall and for each week of the intervention period. Additionally, reasons for not attending or aborting a training session will be summarized. Mean / median ratings of perceived motor-cognitive task difficulty and perceived performance with standard deviations / inter-quartile ranges will be reported overall and for each week of the intervention period. These will be compared to the targeted ranges for task difficulty and perceived performance to establish if an optimal training load was achieved. The volume of usual care (physical, cognitive, and other therapies) and general physical as well as cognitive activities will be descriptively summarized and considered as covariates in the LMEMs of the primary and secondary analyses.
Interim analysis {21b}
N/A. No interim analyses are planned.
Methods for additional analysis (e.g., sub-group analyses) {20b}
As the ITT analyses may underestimate a present treatment effect ((120)), the analyses of the primary and secondary outcomes will be repeated with only those participants, who did not withdraw from the study within the intervention period (T0-T1), and with adherence rates of ≥ 85% (per protocol analysis). This cut-off was chosen based on systematic reviews covering comparable interventions, outcomes and populations (26, 123), where a minimum of 720 minutes in at least twelve weeks or an intervention duration of at least eight weeks were recommended. Moreover, 85% of 24 training sessions results in an “acceptable” absence of 3.6 sessions, which seems practical to account for sickness and conflicting schedules. The results of the ITT and per-protocol analyses will be compared in the discussion of the study report.
Furthermore, to assess clinical meaningfulness of possible treatment effects, the following analyses will be performed on outcomes that revealed a significant between group effect at either T1 or T2 (124, 125). On the one hand, the difference in change score between the two groups will be compared to clinical meaningful change scores (e.g. 1.22 points in the MoCA, (67)). On the other hand, frequencies of “responders” (individual change score above clinically meaningful change) and “non-responders” (individual change score below clinically meaningful change) between the two groups will be compared (124, 125).
Plans to give access to the full protocol, participant level-data and statistical code {31c}
The full protocol of this study was published on https://clinicaltrials.gov (NCT05524727). De-identified participant-level data and the statistical code will be available from the corresponding author on reasonable request.
Oversight and monitoring
Composition of the coordinating centre and trial steering committee {5d}
Recruitment and screening are performed at all study cites by the (local) principal investigators or their delegated staff. All other study-related procedures are performed by trained members of the sponsor / principal investigator team, who are in daily contact regarding the organisation of the trial.
Composition of the data monitoring committee, its role and reporting structure {21a}
Data monitoring in this study is performed by a senior researcher not otherwise involved in the study procedures. On at least three monitoring visits, the monitor reviews the study (team) documents and regulatory aspects, the enrolment process, the participant data, safety aspects, and protocol deviations. Upon monitoring visits, the monitor generates a report including any findings that must be resolved. Resolution of these findings is performed by the principal investigator. Additionally, the clinical trial centre (CTC) of the University Hospital Zurich performs a quality visit on the protocol and monitoring reports. Both, the monitor, and the clinical trial centre are independent of the sponsor.
Adverse event reporting and harms {22}
Adverse events are recorded in the eCRF throughout the study and managed according to GCP guidelines.
Frequency and plans for auditing trial conduct {23}
The monitor performs at least three monitoring visits on the study documentation and procedures; an initiation visit before the study start, at least one routine monitoring visit (further visits are planned if indicated due to inconveniences), and a closure visit upon the end of the study.
Plans for communicating important protocol amendments to relevant parties (e.g., trial participants, ethical committee) {25}
Necessary protocol amendments are submitted to the ethical committee before implementation. All study-team members are informed about all amendments in real-time. In case of an amendment, which changes study procedures or conditions for participants, participants are informed immediately upon approval of the ethical committee.
Dissemination plans {31a}
The findings of this study will be published in scientific journal articles and scientific presentations. All publications will be authored by the study team, following established authorship guidelines. Participants will receive a copy of their individual study data upon request.