Skidmore 2017.

Study characteristics
Methods	Design: parallel RCT Duration of trial: 29 months (August 2012 to December 2014) Unit of randomisation: adult inpatients with acute stroke and cognitive impairment undergoing rehabilitation Recruitment and allocation: of 77 potentially eligible people, 27 did not meet inclusion criteria and 7 withdrew prior to randomisation, leaving 43 who were randomised by random generator to 2 interventions, IG (n = 22) or CG (n = 21)
Participants	Setting: inpatient academic health centre Country: USA Sample size: 43 adults, 51% men; IG: 22; CG: 21 Inclusion criteria: diagnosis of first‐time stroke within 30 days of rehabilitation admission and demonstrated cognitive impairments (as indicated by 14‐item Executive Interview ≥ 3) Exclusion criteria: severe aphasia, prior cognitive impairment or dementia, current major depressive disorder, recent substance abuse or psychosis Age: mean: IG: 65.86 (SD 11.67) years; CG: 66.73 (SD 14.25) years Time since stroke onset: mean: IG: 16.29 (SD 18.24) days; CG: 22.36 (SD 30.97) days Types of stroke: ischaemic, n: IG: 14 (67%); CG: 14 (64%) Hemisphere: left, n: IG: 10 (62%); CG: 13 (59%) NIHSS stroke severity: mean: IG: 7.00 (SD 3.90); CG: 7.86 (SD 5.10)
Interventions	Intervention group Brief name: guided training Recipients: inpatients with acute stroke and cognitive impairment undergoing rehabilitation Why: to maximise the expertise of participant by training them to actively engage in the direction and focus of their treatment. Therefore, they learned to identify and prioritise problematic daily activities, identify barriers to performing activities, generate their own individualised strategies for addressing these barriers, and apply these processes through iterative practice. In doing so, guided training is designed to equip people with the ability to generalise knowledge and skills in problem identification and problem‐solving skills to address new but similar problems over time. What (materials): COPM (Law 1998); guided training standardised intervention protocol What (procedures): in addition to usual care; using the COPM, the therapist asked the participant to describe a typical routine prior to the stroke, focusing on a typical weekday, a typical Saturday, and a typical Sunday. The therapist then asked the participant to identify 4–6 activities that were important to them and that he or she thought likely to be problematic after the stroke. The subsequent intervention was then focused on these activities. The participant picked the first activity that he or she wanted to practise. The participant performed that activity and the therapist identified barriers to performance, and taught the participant a global strategy, 'goal‐plan‐do‐check', by asking the participant to set a goal to address the barriers (i.e. identify a criterion for performance), develop a plan to address the goal, complete the plan, and check whether the plan worked or required revising. This process was repeated iteratively until the goal was met (participants moved on to the next activity), or until the end of the 10 sessions. The therapist guided participants using prompting questions and workbooks to facilitate learning and aid the participants in implementing the strategy. Who provided: licensed occupational therapists and physiotherapists who were independent contractors and not members of the usual care rehabilitation team How: face‐to‐face and individually Where: inpatient rehabilitation at an academic health centre and home When and how much: 10 sessions of 45 minutes, 5 days per week for 2 weeks (7.5 hours in total) Tailoring: the research therapist asked the participant to describe a typical routine prior to the stroke, focusing on a typical weekday, typical Saturday, and typical Sunday. Therapist then asked the participant to identify 4–6 activities that the participant thought were important and likely to be problematic after the stroke. Therapists focused subsequent intervention programme on these activities. Modifications: none reported How well (planned): fidelity to each protocol and differentiation of elements between protocols assessed using standardised checklists applied to a random 20% of video‐recorded sessions in each condition. They also assessed the degree to which elements of direct skill training and guided training protocols were present in usual care sessions (1 occupational therapy session, 1 physical therapy session, and 1 speech therapy session, if being followed by speech therapy, for each study participant). How well (actual): adherence to intervention procedures was 85% in the guided training group. The 2 protocols were sufficiently differentiated, with guided training intervention elements present in none of the sampled direct skill training sessions. On average, direct skill training participants and guided training participants received a similar number of sessions. Comparator group Brief name: direct skill training Recipients: inpatients with acute stroke and cognitive impairment undergoing rehabilitation Why: to promote independence of participants in daily activities; direct skill training maximises the skill of the rehabilitation practitioner who directs treatment. What (materials): COPM (Law 1998); direct training standardised intervention protocol What (procedures): research therapist selected an activity to practice in the research session that addressed 1 of the participant‐identified goals. The direct skill training therapist analysed the activity using an interview and performance‐based assessment, and identified barriers to performance of that activity. The therapist then set criteria for performance, developed strategies to improve performance, and taught the participant those strategies. Participants practiced the identified strategies until the performance criteria were met (research therapist moved on to the next activity), or until the end of the 10 sessions. The direct skill training therapist documented each of these steps using pre‐established forms in a workbook shared with the participants. Who provided: licensed occupational therapists and physiotherapists who were independent contractors and not members of the usual care rehabilitation team How: face‐to‐face Where: inpatient rehabilitation at an academic health centre and home When and how much: 10 sessions of 45 minutes, 5 days per week for 2 weeks Tailoring: activities and performance criteria used for the skill training were based on the individual assessment. Modifications: not reported How well (planned): fidelity to each protocol and differentiation of elements between protocols was assessed using standardised checklists applied to a random 20% of video‐recorded sessions in each condition. They also assessed the degree to which elements of direct skill training and guided training protocols were present in usual care sessions (1 occupational therapy session, 1 physical therapy session, and 1 speech therapy session, if being followed by speech therapy, for each study participant). How well (actual): analysis of the sampled usual care sessions indicated that on average 98% of training in usual care sessions was consistent with direct skill training and 2% (usual care training) was consistent with guided training.
Outcomes	Primary FIM Secondary: none Methods of data collection: FIM administered by 4 independent (blinded) trained evaluators Data collection time points: study admission (baseline); inpatient rehabilitation discharge; and 3, 6, and 12 months after study admission
Notes	Funding: yes Conflict of interest: none Published trial protocol: no Trial registration: yes, NCT02766400 Ethics approval: none reported Data: we obtained additional mean and standard error data from the Clinical Trials Registry (NCT02766400); We converted standard errors to SDs within Review Manager 5.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "… assigned to an intervention by the research coordinator using a simple randomization scheme (1:1 allocation ratio) developed with a random number generation program …"
Allocation concealment (selection bias)	Low risk	Quote: "… maintained in an electronic file accessible only to the research coordinator …"
Blinding of participants and personnel (performance bias) All outcomes	High risk	*For personnel only: quote: "… research therapists were independent contractors who were not members of the usual care rehabilitation team. To avoid cross contamination, the research therapists were trained in only one protocol and remained naive to the opposite protocol …"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "… FIM was administered by 4 trained independent (ie, blinded) evaluators".
Incomplete outcome data (attrition bias) All outcomes	Low risk	Linear mixed model analysis used. Quote: "all participants regardless of retention were included in the remaining analysis".
Selective reporting (reporting bias)	Low risk	As per protocol.
Other bias	Low risk	No other identifiable bias.