Task-oriented training in rehabilitation after stroke: systematic review


  • Marijke Rensink,

    1. Marijke Rensink MD Clinical Researcher Faculty Chair, Care for the elderly and chronically ill, University of Applied Sciences, Utrecht, the Netherlands
    Search for more papers by this author
  • Marieke Schuurmans,

    1. Marieke Schuurmans PhD RN Professor Faculty Chair, Care for the elderly and chronically ill, University of Applied Sciences, Utrecht, the Netherlands
    Search for more papers by this author
  • Eline Lindeman,

    1. Eline Lindeman PhD MD Professor of Rehabiliation Department of Rehabilitation and Sports Medicine, Rudolf Magnus Institute of Neuroscience, University Medical Center, Utrecht, the Netherlands
    Search for more papers by this author
  • Thóra Hafsteinsdóttir

    1. Thóra Hafsteinsdóttir PhD RN Assistant Professor Faculty Chair, Care for the elderly and chronically ill, University of Applied Sciences, Utrecht, and Division of Neuroscience, Rudolf Magnus Institute, University Medical Center, Utrecht, the Netherlands, and Faculty of Nursing, University of Iceland, Reykjavik, Iceland
    Search for more papers by this author

M. Rensink: e-mail: marijke.rensink@hu.nl


Title. Task-oriented training in rehabilitation after stroke: systematic review.

Aim.  This paper is a report of a review conducted to provide an overview of the evidence in the literature on task-oriented training of stroke survivors and its relevance in daily nursing practice.

Background.  Stroke is the second leading cause of death and one of the leading causes of adult disability in the Western world. The use of neurodevelopmental treatment in the daily nursing care of stroke survivors does not improve clinical outcomes. Nurses are therefore exploring other forms of rehabilitation intervention, including task-oriented rehabilitation. Despite the growing number of studies showing evidence on task-oriented interventions, recommendations for daily nursing practice are lacking.

Data Sources.  A range of databases was searched to identify papers addressing task-oriented training in stroke rehabilitation, including Medline, CINAHL, Embase and the Cochrane Library of systematic reviews. Papers published in English between January 1996 and September 2007 were included. There were 42 papers in the final dataset, including nine systematic reviews.

Review methods.  The selected randomized controlled trials and systematic reviews were assessed for quality. Important characteristics and outcomes were extracted and summarized.

Results.  Studies of task-related training showed benefits for functional outcome compared with traditional therapies. Active use of task-oriented training with stroke survivors will lead to improvements in functional outcomes and overall health-related quality of life.

Conclusion.  Generally, task-oriented rehabilitation proved to be more effective. Many interventions are feasible for nurses and can be performed in a ward or at home. Nurses can and should play an important role in creating opportunities to practise meaningful functional tasks outside of regular therapy sessions.

What is already known about this topic

  • • Task-oriented training is effective for stroke survivors.
  • • There is a shift from training at impairment level to training on activity level.
  • • Healthcare professionals spend little time providing therapeutic activities and treatment to stroke survivors.

What this paper adds

  • • Training needs to be repetitive, task-specific and meaningful for the individual.
  • • Many interventions in task-oriented training that have proved to be effective are usable in daily nursing practice, such as walking on the ground, moving from sitting to standing from different chairs, and sitting and reaching.
  • • Active use of task-oriented training in the daily nursing care of stroke survivors, will lead to improvements in functional outcomes and overall health-related quality of life.

Implications for practice and/or policy

  • • Nurses should collaborate with other professionals to create opportunities for patients to practise outside of regular therapy sessions.
  • • The effects of task-oriented training programmes provided by nurses in collaboration with other healthcare professionals need to be measured.


Stroke is the second leading cause of death and one of the leading causes of adult disability in the Western world today. In the Netherlands it is estimated that the incidence of stroke will increase from 1·8 per 1000 inhabitants in 2000 to 2·8 per 1000 in 2020 (Struijs et al. 2005).

Stroke rehabilitation is an organized endeavour to help patients to maximize all opportunities for returning to an active lifestyle (Gresham et al. 1997, Aichner et al. 2002). Neuro-rehabilitation is a method for relearning a previously learned task in a different way, either by compensatory strategies or by adaptively recruiting alternative pathways (Matthews et al. 2004). Rehabilitation nursing focuses on assisting people with a disability or chronic illness to attain maximum functional ability, maintain optimal health and adapt to an altered lifestyle (Barker 2002). While stroke is an important cause of disability, there is no generally accepted method for rehabilitating stroke survivors. Commonly-used treatment approaches that focus on impairments and seek to regain a ‘normal’ movement pattern, such as neurodevelopmental treatment (NDT), have proved ineffective (Pomeroy & Tallis 2000, Paci 2003, Hafsteinsdottir et al. 2005, Kollen et al. 2006, Lennon et al. 2006).

During the past two decades, major progress in neuroscience has resulted in novel concepts for rehabilitation interventions after stroke. Various studies support the choice of task-oriented training. Neuro-imaging studies in animals and humans have provided strong evidence for changed activation patterns in many parts of the damaged brain (Turkstra et al. 2003, Kleim et al. 2004, Nudo 2007). In addition, movement and experience-dependent reorganization patterns have been observed in both the damaged hemisphere and the contralateral hemisphere (Johansen-Berg et al. 2002, Jang et al. 2003, Liepert et al. 2004, Lindberg et al. 2004, Luft et al. 2004, Ward 2007). There are strong indications that functional recovery is not only the result of restoration from impairments. Adaptation strategies to compensate for the impairments also contribute (Kwakkel et al. 2004a).

Definition of task-related training

Unfortunately, no conclusive definition of a task-oriented approach exists in the literature. In the task-oriented approach, movement emerges as an interaction between many systems in the brain and is organized around a goal and constrained by the environment (Shumway Cook & Woollacott 2001). Task-oriented training includes a wide range of interventions such as treadmill training, walking training on the ground, bicycling programmes, endurance training and circuit training, sit-to-stand exercises, and reaching tasks for improving balance. In addition, use is made of arm training using functional tasks such as grasping objects, constraint-induced (movement) therapy (CIMT) and mental imagery. Such training is task and patient focused and not therapist focused.

As well as the content of the therapy, the optimal amount of therapy needed for patients after stroke is not exactly known. The time spent in exercise programmes is often decided pragmatically and is not based on the time necessary to learn a given skill (Kwakkel 2006). In a systematic review of the effects of therapy intensity it was concluded that augmented therapy input of at least 16 hours a week has a favourable effect on the improvement of activities of daily living (ADL) (Kwakkel et al. 2004b). However, nurses provide continuous and coordinated care to stroke survivors 24 hours a day, 7 days a week and could play an important role in the relearning process. A well-coordinated and organized multidisciplinary rehabilitation programme, beginning as soon as possible after stroke, is important for an effective stroke unit (Indredavik et al. 1999, Langhorne & Pollock 2002). Therefore, nurses are key members of the stroke team and their input is essential for achieving set rehabilitation goals (Long et al. 2003). In other words, nurses play an essential role in coordinating care and bridging the gap between disciplines (Strasser et al. 2005).

The review


The aim of the present study is to provide an overview of the evidence in the literature on task-oriented training of stroke survivors and its relevance in daily nursing practice.


Literature was systematically reviewed following the steps of the QUORUM statement (Moher et al. 1999) and the Cochrane Handbook for Systematic Reviews (Higgins et al. 2005).

Search methods

First, the databases Medline, Cumulative Index to Nursing and Allied Health Literature (CINAHL) and Embase were searched using the following key terms: stroke (MeSH term) and cerebrovascular accident (MeSH term), combined with task-related training and task-oriented training. Those two terms were combined with activities of daily living, posture, arm, walking and physical condition (MeSH terms) and with the non-MeSH terms balance, balance exercise, sit to stand, gait, constraint-induced movement therapy and mental practice.

Inclusion criteria

  • • Types of participant: stroke survivors in the acute phase, the rehabilitation phase and the chronic phase.
  • • Type of event: ischemic and haemorrhagic stroke.
  • • Types of outcome measure: outcomes of interventions are focused on functional performance and/or activities of daily living. Studies comparing clinical outcome measures with laboratory tests are also analyzed.
  • • Publication date: published in English between January 1996 and September 2007.
  • • Review design: meta-analysis, systematic reviews and randomized controlled trials.
  • • Types of intervention: Only studies with interventions aimed at task-oriented exercises that are feasible and suitable for daily nursing practice in a stroke ward or at home with minimal technical equipment were selected. The intervention needed to be congruent with the definition of the Nursing Interventions Classification (McCloskey & Bulecheck 2000): a nursing rehabilitation intervention is ‘any treatment based upon clinical judgment and knowledge that a nurse performs to enhance patient outcomes’ (p19). Interventions aimed towards the following domains were considered:
    • oBalance exercise, i.e. use of specific activities and movements to maintain, enhance or restore balance (McCloskey & Bulecheck 2000). Task-oriented training involves reaching towards objects placed across a table, and providing an implicit exercise to improve symmetrical weight distribution over both legs (Mudie et al. 2002, Dean et al. 2007) .
    • oAmbulation exercise: promotion and assistance with walking to maintain or restore autonomic and voluntary body functions (McCloskey & Bulecheck 2000). Walking training is often performed on a treadmill, with or without body weight support. In the present survey, only studies in which one of the interventions was walking on the ground were analysed.
    • oStrength training: facilitation of regular resistive muscle training to maintain or increase muscle strength (McCloskey & Bulecheck 2000).
    • oExercise promotion: facilitation of regular physical exercise to maintain or advance to a higher level of fitness and health (McCloskey & Bulecheck 2000).
    • oArm-training: This includes CIMT. CIMT has three components: restraint of the less impaired arm, training functional tasks repetitively and a package of behavioural techniques with emphasis on immediate encouraging feedback when patients make even a small gain (Uswatte et al. 2006). Use of functional tasks makes CIMT a good example of task-oriented training.
    • oMental practice: Mental practice or mental imagery is a technique in which a skill is mentally rehearsed in a repetitive manner without any visible movement or muscle activation (Mulder 2007).

Search outcome

The initial search strategy generated 1506 papers. The first author evaluated the titles and 563 articles were found to fit the inclusion criteria. Of these, 147 abstracts were included for further examination. At the final stage, studies included in published reviews were not counted separately, with exception of three papers on walking training.

Quality appraisal

Inclusion in the final sample was guided by the methodological quality of the 147 studies selected, which were exclusively randomized controlled trials (RCT). This quality was evaluated independently by two authors, using the Delphi Criteria List (Verhagen et al. 1998). The quality of the systematic reviews was evaluated using the criteria described by Grimshaw et al. (2003). In the final dataset, nine systematic reviews (including only RCTs) and 33 RCTs were included (Figure 1). Papers included in the systematic reviews were not analysed separately, with exception of three papers on walking training (Dean et al. 2000, Blennerhassett & Dite 2004, Salbach et al. 2004), which were also included in the review by Van de Port et al. (2007).

Figure 1.

 The screening process.

Data abstraction

The following study characteristics were recorded on a data extraction form: setting and phase, study design and population, intervention, outcome and measurement. For the systematic reviews the intervention, the number of studies included and the conclusion were extracted (Tables 1 and 2).

Table 1.   Characteristics of the systematic reviews included in the review
ReferencesInterventionNumber of studiesConclusion
  1. RCT, randomized controlled trial; CCT, controlled clinical trial; SR, systematic review; SES, summary effect size; CI, confidence interval; FAC, functional ambulation categories; (I)ADL, (instrumental) activities of daily living; QoL, quality of life.

van der Lee et al. 2001Exercise therapy for arm function in stroke survivors15 trials out of 72.
13 RCTs n = 939
No statistical pooling. Insufficient evidence to draw conclusions about the effectiveness of exercise on arm function. There is a suggestion that more therapy is beneficial
Van Peppen et al. (2004)Impact of physical therapy on functional outcomes after stroke151 studies, 123 RCTs and 28 CCTStrong evidence in favour of task-oriented training to restore balance and gait and for strengthening the paretic limb. SES high intensity exercise training 0·13 95% CI 0·03–0·23. Insufficient evidence in functional outcome for traditional neurological treatment approaches
Saunders et al. (2004) Cochrane reviewPhysical fitness training for stroke survivors12 trials, Total n = 289Statistically significant improvement was observed only in FAC scores and max. walking speed after walking training, standardized mean diff. 0·42 m/s 95% CI 0·04–0·79. Any training induced benefit appear to be associated with specific or task-related training
Pang et al. (2006a)Exercise training using as an outcome peak oxygen consumption and walking velocity and enduranceSeven studies (RCT) out of 29, n = 13–157Statistically significant effect size in favour of aerobic exercise to improve peak VO2 . (SES 0·42 95% CI 0·15–0·69, P = 0·001) and in favour to improve walking velocity and endurance (SES 0·30 95% CI 0·06–0·55 P = 0·008)
Pollock et al. (2007), Cochrane reviewPhysiotherapy treatment approaches for the recovery of postural control and lower limb function following stroke21 studies out of 265,
 Total n = 1087
Insufficient evidence to conclude that one approach is more effective in promoting lower limb function or postural control than another approach Limited evidence that using a mix of components from different approaches is more beneficial than no treatment or placebo. Standardized mean diff. SMD 0·94 95% CI 0·08–1·80
Ada et al. (2006)Strengthening interventions including progressive resistance exercise15 trials out of 102, Total n = 359Strengthening interventions had a small pos. effect on both strength (standardized mean diff. SMD 0·33 95%CI 0·13–0·54) and activity (SMD 0·32 95%CI 0·11–0·53). No effect on spasticity.
Van de Port et al. (2007)Exercise training programmes on walking competency after stroke12 studies out of 246 studies, Total n = 501Strong evidence was found for improved functional mobility after gait-oriented training (SES fixed 0·45 CI 0·27–0·63) Findings provide strong evidence that standing balance, (I) ADL or QoL not statistically significant more improved than by conventional care.
Bohannon (2007)Muscle strength and muscle training of the lower limbs after strokeSearch identified 3 SRs, 5 RCTs and 7 other studiesNo statistical pooling. The ability of strengthening to enhance the performance of functional activities or participation remains uncertain, except perhaps regimens involving repeated sit to stand or step up manoeuvres.
Bonaiuti et al. (2007)The constraint induced therapy on adult stroke survivors9 out of 13 RCTs Total n = 243Statistical differences could not been measured Minimal clinical importance was defined as a change at least 10% change of the maximum score of a test. Studies suggest an effectiveness but samples are too small and there is no homogeneity
Table 2.   Characteristics of the studies included in the review
ReferencesSetting/phaseSampleStudy designInterventionOutcome/measuresResults and conclusion
  1. RCT, randomized controlled trial; CI, confidence interval; BBS, berg balance scale; TUG, timed up and go-test; ABC, activities specific balance confidence scale; NHP, Nottingham Health Profile; FM, Fugl-Meyer assessment; PASS, Postural Assessment Scale For Stroke Patients; CoP center of pressure, MAS, Motor Assessment Scale; RM, Rivermead Motor assessment; TEMPA, upper extremity test for older people; 6/10, 6/10 minutes walking test; FIM, Functional Independence Measure; BI, Barthel Index; MAL, Motor Activity Log; ARA, action research arm test; MP, mental practice; Exp, experimental group; CIMT, constraint-induced (movement) therapy.

Marigold et al. (2005)Community chronic phasen = 61RCT 2 groups, 2 interventions1. Agility training
2. Stretching/weight shifting
3 times/week/1 hour/10 weeks 1 supervisor/3 participants
BBS, TUG, step reaction time, ABC, NHP, induced falls on a platformExercise led to improvement in all clinical outcome measures in both groups (trend towards statistical significance). Group-by-time interaction, = 0·04. Group exercise improves functional balance
Katz-Leurer et al. (2006)Rehabilitation Department Sub acute phasen = 24RCT Exp. n = 106 weeks usual rehabilitation  Experimental group:  cycling programme,  during 3 weeks dailyStanding balance PASS
Fugl Meyer
Both groups improved statistically significantly on PASS and Fugl Meyer with a statistically significant group-time interaction effect (P < 0·01) on the PASS and FM score
Bayouk et al. (2006)Discharged home ≥ 6  months post stroken = 16RCT Exp. n = 88 week task oriented  exercise on balance  and mobility. Exp.  plus. manipulation  poprioception of feet  and ankles and/or vision1. Measurement  COP displacement  under 4 sensory  conditions. 2. 10 m  walking testBoth groups improved statistically significantly on the 10 m walking test (<P 0·05). Exp. group performed better in standing double legged with eyes open on a normal and soft surface (P < 0·05)
Balance/perception training
Morioka and Yagi (2003)Hospital Sub-acute phasen = 28RCT Exp. n = 12 Hardness discrimination  using three different  levels of hardness  (5 mm–15 mm) of  sponge rubber placed  under the sole of the  foot (10 days)Postural sway by a stabilometer, eyes open and eyes closed.Incorrect answers in testing the hardness, decreased (P = 0·001). No differences in two point discrimination was found between groups. Difference in postural sway parameters was statistically significant
Bonan et al. (2004)Discharged home chronic ≥  12 months post stroken = 20Single blinded RCT Exp. n = 10Rehabilitation  programme for both  groups, except that  the eyes of the vision  deprived group were  blinded with a maskBalance (laboratory test 6 sensory conditions). Gait velocity, timed stair climbing and walkingIn all six sensory conditions the gain in the experimental group was greater (sign. P = 0·01 and P = 0·04, in 2 conditions). Gain in balance correlated with gait velocity (P = 0·03), timed stair climbing (P = 0·01)
Sitting and reaching
Dean and Shepherd (1997)Community ≥  12 months after stroken = 20RCT Exp. n = 1010 sessions in 2 weeks  at home, reaching  beyond arm length  Control: sham  trainingMuscle strength limb, sit to stand, walking, ‘reach to grasp and drink a glass of water’Experimental group reached further and faster, increased activation of the affected leg muscles (P < 0·001)
No difference in walking distance
Pollock et al. (2002)Hospital sub-acute phase <  6 weeks poststroken = 28 RCT Exp. n = 9Exp. special designed protocol: construction tasks and stacking tasks, during 4 weeks Both groups BobathAchieving normal symmetry of weight distribution measured with a force platformNo clinically significant differences between the groups. The regime of independent practice had no measured beneficial effect on the weight distribution
Mudie et al. (2002)Rehabilitation Unit sub-acute phasen = 40RCT 4 groups n = 4 × 101. task specific reach, 2. Bobath 3. balance, 4. no specific training. 2 weeks daily sessions. Follow-up after 12 weeksSymmetry of weight distribution measured with the balance perfomance monitorAt 12 weeks: 83% of the balance group, 38% of the task specific reaching group, 29% of the Bobath group, 0% of the untrained group distribute their weight to both sides
Thielman et al. (2004)Home 5–12 months discharged from rehabilitationn = 12RCT n = 2 × 6Training paretic limb 1. Training task related 2. progressive resistance4 weeks/ 12 × 35 min/150–180 movements/sessionKinematic analysis arm and trunk, MAS and Rivermead (RM)No statistical significant effect for the MAS. Low level patients showed better performance
Michaelsen and Levin (2004)Home chronic 7–94 months poststroken = 28RCT Exp. n = 14Grasping a cilinder with no trunk moving. Exp: restriction of trunk movements by a harnass. 1 single session1. Fugl Meyer Arm 2. Upper Extremity performance test for the elderly (TEMPA)Trunkrestraint group: more elbow extension, less anterior trunk displacement (group main effect = 5·14, P < 0·05) and better interjoint coordination
Howe et al. (2005)Hospital acute phasen = 35RCT Exp. n = 17Usual care plus exp: Exercises lateral weight transference in sitting 12 sessions/4 weeksDynamic reaching, sitting and standing. (timed standing up) Static standing balanceNo differences between groups. Lateral weight transference did not appear to enhance the functional rehabilitation of patients in the acute phase
Michaelsen et al. (2006)Community chronic 6–48 months after stroken = 30RCT Exp. n = 15Supervised home programme Exp: progressive reach to grasp with prevention of trunk movements (TR) 3×/week/5 weeks1. Fugl Meyer, 2. movement kinematics 3. TEMPA. Follow-up after 1 monthExperimental group: greater improvement in impairment FM P < 0·035, and function TEMPA P < 0·05. Increased joint range . More severe patients, more effect
Dean et al. (2007)Hospital <  3 months poststroken = 12RCTExp. sitting training protocol as in the study in 1997
Control: sham training
Same tests as study in 1997: Movement time, reaching test, muscle strength affected leg, sit to stand and walkingMaximal reach distance increased 0·17 m (95% CI 0·12–0·21). Force foot increased 21%. Sitting training early after stroke improves sitting ability and carries over to standing up but not to walking
Sit to stand
Cheng et al. (2001)Rehabilitation unit sub- acute phasen = 54RCT Exp. n = 30Exp. Conventional stroke rehabilitation and symmetrical sit to standing trainingAssessing sit to stand. Occurrence of falls in the two groupsStatistically significant improvement in training group in the sit to stand performance. After 6 months less fallincidents (< 0.05)
Dean et al. (2000)Rehabilitation Center ≥ 3  months poststroken = 12RCT convenience sample Exp. n = 6Exp. strengthening affected limb and practicing functional tasks in a circuit training. Control: upper limb tasks 4 weeks/3×/week, 1 hourWalking speed and endurance, peak vertical ground reaction force and step testStatistically significant improvement in the experimental group (= ≤0.05) on all the outcome measures.
Blennerhassett and Dite (2004)Rehabilitation Center sub-acute phasen = 30RCT single blinded clinical trial Exp. n = 15Exp. extra training functional tasks in a circuit of ten tasks. Control: upper limb training functional tasks 4 weeks/1 hour a dayTUG, Step test and 6 Minute Walking tets
Arm: the Jebsen Tayler Hand Function test and the MAS
Mobilitygroup: 6MWT trend to walk further P = 0·01. TUG P = 0·02. The mean difference between the groups was 116 m (95% CI 31–201 m) in the 6MWT. Task related training effects were found
Salbach et al. (2004)Community 1 year post stroken = 91RCT Exp. n = 44Exp: 10 functional tasks. Control: arm activities. 6 weeks/ 3×/week6MWT, 5-meter speed, TUG, BBSBetween group difference was 35 m on the 6MWT (95% CI 7–64),TUG no difference
Salbach et al. (2005)Community Up to 1 year after stroken = 91RCT Exp. n = 44 Follow-up after 1 year (2004)Exp: 10 functional tasks to strengthen lower extremities and walking balance. Control: armactivities 6 weeks/3×/week6MWT, 5-meter comfortable walking speed, TUG. BBS, ABCWalking intervention group; change in balance efficacy correlated with change in functional walking capacity R = 0·45, 95% CI = 0·16–0·68 (2005)
Nilsson et al. (2001)Rehabilitation Center sub- acute phase  < 8 weeks poststroken = 73RCT Exp. n = 36 Follow-up 10 monthsExp: treadmill with Body Weight Support (BWS) 5 days /week/ 30 minute. 3–19 weeks. Control: walking training on the ground, no treadmillFIM, FAC, Fugl Meyer, BBSNo statistically difference between the groups at discharge and in follow-up. Both groups improved on the tests. Walking on the ground and BWSTT are comparable choices early after stroke
Richards et al. (2004)Rehabilitation Unit sub-acute phasen = 63RCT Exp. n = 32Exp: training with a treadmill Control: no technology 1 hour/day/ 5 days/week/2 monthsGaitspeed, Fugl Meyer, BBS, TUG, BI (gait) Laboratory. measures: gait kinematicsAll measures improved (P < 0·01–0·05). No differences between groups (P > 0·05). The efficacy of task oriented training depends not on rehabilitation technology
Peurala et al. (2005)Rehabilitation center >6  months post stroken = 45RCT 3 groups Follow-up 6 months1. Gait trainer plus  Functional Electric  Stimulation (FES)
2. Gait trainer  without FES
3. walking overground
10MWT, 6MWT., strength and spasticity LE, postural sway, Modified MAS, FIMScores in all tests improved (< 0.001). However no differences were found between the groups. Performance remained improved after 6 months
Physical condition
Duncan et al. (1998)Discharged at home 30–90  days after stroken = 20RCT Exp. n = 10Exp: a home based exercise programme, therapist supervised, designed to improve strength, balance and endurance
8 weeks/3×/week
Fugl Meyer upper (UE) and lower extremity. (LE). Gait velocity, 6 m walking, BBS, BIFM UE change 8·4 vs. 2·2, FM LE 4·7 vs. −0·9. gait speed 0·25 vs. 0·9 m/s. 6 m walking 195 feet vs. 114 ft. BI trend in difference. The difference in gait speed was statistically significant in change score 0·05 < 0·1
Duncan et al. (2003)Community Stroke within 30–150 days n = 100RCT Exp. n = 50Exp: exercise programme to improve strength, balance and endurance. Therapist supervised. Control: usual care 12–14 weeks/ 36×/30–90 minutes.Strength, Fugl Meyer LE, BBS, endurance, Wolf Motor Function test UE, 10 m walk and 6 minute walk distanceMultivariate analysis of variance (manova) testing the overall effect, the intervention produced greater gains (Wilk’s λ = 0·64, P = 0·0056)
Katz-Leurer et al. (2003)Rehabilitation unit sub-acute phasen = 92RCT Exp. n = 468 weeks cyclo ergometer training. 2 weeks/5×/week/ 10–20 minute, 6 weeks/ 3×/week/30 minute. Intensity: 60% of heart rate reserveAerobic capacity. Functional tests in stair climbing, walking time and walking distance until fatiguedImprovement in all aerobic parameters. Heart rate at rest (P = 0·02), work load and time (< 0·01) statistically significant. A trend to sign. in functional tests, stair climbing was statistically significant better (P < 0·01)
Pang et al. (2005)Community >  1 year post stroken = 63RCT Exp. n = 32Exp: fitness/mobility exercises stepping, walking, sit to stand, strengthening and balance. 19 weeks/ 3×/week/1 hourO2 consumption, 6MWT, strength, BBS, bone density femur neck, activity scalePeak volume oxygen and 6MWT statistically significant improvement (P = 0·03). No statistically significant improvement in bone density, balance, strength and activity
Yang et al. (2006)Community chronic stroke ≥ 1  year post stroken = 44RCT Exp. n = 24Exp: task oriented strength training in a circuit: standing and reaching, sit to stand, stepping. 4 weeks/3×/ week/ 30 minuteMuscle strength, gait performance on a instrumented walkway, 6MWT, step test, TUG All selected tests improved statistically significant (P < 0·001). With exception of the step test. Strength gain associated with gain in functional tests (P < 0·001)
Winstein et al. (2004)Rehabilitation unit acute phase 2–35  days after stroken = 64RCT 3 groups Follow-up 9 months1. Standard care (SC)
2. Functional task practice (FT)
3. Strength training (ST) ST and FT 20 hours additional hours beyond standard care
Fugl Meyer Strength Functional test of the hemiparetic Upper ExtremityFT and ST groups greater increases in FM scores (P = 0·04) and strength (P = 0·02). Effect was primarily in less severe patients. After 9 months, the less severe FT group continued to make statistically significant gains
Pang et al. (2006a)Community Chronic ≥  1 year after stroken = 63RCT n = 31 arm group n = 32 leg groupGroup training 9–12 patients supervised by 3 therapists. Arm group and leg group. 19 weeks/3×/weekWolf Motor Function test, Fugl Meyer arm, grip strength, Motor Activity Log (MAL)Statistically significant group/time interaction (Wilks λ = 0·726, P = 0·017. WMTF and FM statistically significant higher in post hoc analysis. Patients with moderate arm impairment benefited more.
Higgins et al. (2005)Community  < 1 year poststroken = 91RCT Exp. n = 47Exp: practice of uni- and bilateral tasks. 6 weeks/3×/ week/90 min, 15 minute/day home training. Control walkingBox and Block test Nine hole peg test, grip strength. TEMPABaseline BBT average of 26 blocks in both groups, = 40% of age related values. Postintervention: both groups 28 blocks. No meaningful change in other measures
Taub et al.. (2006b)Community chronic mean =  4·5 year after stroken = 41RCT Exp. n = 21CIMT, 6 hours/day/10 days plus restraint 90% of waking hours. Control: physical fitness and relaxationWolf Motor Function test (WMFT)
Motor Activity Log (MAL)
Patients show large (WMFT) to very large (MAL) improvement in functional use (MAL P < 0·0001). Changes persisted over the 2 years tested
Wolf et al. (2006)Community 3–9 months poststroken = 222RCT Exp. CIT n = 106CIMT: repetetive training restraint 90% of waking hours. Control: no interventionWolf Motor Function test (WMFT), MALBetween group difference WMFT 34% (95% CI 12%–51%, < 0·001) and the MAL between group difference 43%P < 0·001
Wu et al. (2007)Rehabilitation setting 0·5–31  months poststroken = 26RCT Exp. n = 13Exp. mCIMT, restraint 6 hours a day, 2 hours/ 5times/week/3 weeks Control: traditional therapyFIM, MAL Stroke Impact Scale (SIS), HRQOL (quality of life)Statistically significant improvement in motor function, daily function and the physical domain of HRQOL. FIM P = 0·018, MAL amount of use arm P = 0·003
Mental practice
Page et al. (2007b)Community chronic >  12 months poststroken = 32RCT MP n = 16Relaxation plus Mental practice (R + MP) Control: R plus physical practice 30 minutes/2 days/ week/6 weeksAction Research Armtest (ARA) Fugl Meyer UEMP group had statistically significant reductions in arm impairment and statistically significant increase in daily arm function (both P < 0·0001)


The studies included differed markedly with regard to interventions, methodology, outcome measures, patient characteristics and methodological quality. Also, the phase after stroke and the setting of the study differed. Eleven studies were performed in the (sub)-acute phase, the other 22 in the chronic phase. Because of these differences it was not possible to conduct meta-analyses pooling the results of the various task-oriented training interventions, and so the findings are reported using a narrative summary technique.


Material from the literature is organized below in a way similar to previously-published guidelines for stroke rehabilitation (Teasell et al. 2006), starting with balance training and proceeding to sitting and reaching, sit to stand training, gait training and interventions for physical fitness. In the final part, we describe the results of arm training and conclude with mental imagery practice.

Balance training

A systematic review showed that there is insufficient evidence that one intervention or approach is more effective than another in improving balance recovery among stroke survivors (Pollock et al. 2007). In the present investigation, three small studies of good quality focusing on task-oriented exercises and using functional outcome measures were identified and reviewed. The effect of cycling training on balance in sub-acute phase patients was measured by Katz-Leurer (n = 24). The exercise group, who had a daily cycling session in addition to the usual therapy, maintained balance better under different conditions. There was no follow-up after 6 months (Katz-Leurer et al. 2006). The other two studies were performed with patients in the chronic phase. In one study (n = 16) the effects on stroke survivors of a task-oriented exercise programme were compared with altered sensory input. Tasks were performed on different surfaces with eyes closed or open. Those receiving exercise assisted by sensory manipulation improved statistically significantly (< 0·05) in standing balance, but the effect did not extend to walking (Bayouk et al. 2006). Patients (n = 30) in the group trained with an agility exercise programme showed greater improvement in step reaction time and had fewer falls when balance was challenged. The agility exercise programme halved the number of patients who fell during the follow-up after 1 year (Marigold et al. 2005).

Two studies focusing on perceptual learning were identified. Both were well performed but the samples were small and there was no follow-up. Patients in the sub-acute phase (n = 12) were trained to discriminate the hardness of different pieces of sponge rubber placed under the sole of the foot. Balance (body sway) improved statistically significantly (= 0·001) in the experimental group. Improvement of clinical standing balance was not determined (Morioka & Yagi 2003). In the second study, long-term survivors (n = 20) received balance training with visual deprivation. Laboratory measurements of balance under six conditions improved more in the vision-deprived than in the free-vision group. Also, gait velocity and timed stair climbing correlated statistically significantly with improved balance (= 0·01, = 0·04) (Bonan et al. 2004).

Balance training focused on balance parameters did not generalize to functional improvement. Balance training needs to be practised in relation to a task (Bayouk et al. 2006).

Sitting and reaching

Eight studies of moderate quality were found that focused on sitting and reaching. One study performed in the sub-acute phase after stroke showed no differences in symmetric weight bearing between a group of patients practising several reaching tasks independently and a group receiving the usual care (Pollock et al. 2002). A training programme with lateral weight transference exercises did not enhance functional outcome (Howe et al. 2005). Balance feedback training and task-specific training improved symmetrical weight distribution in patients (Mudie et al. 2002). In a small study (n = 12) of patients in the chronic phase, task-related training was more beneficial than resistance training for patients’ functioning on a lower level (Thielman et al. 2004). In two small studies in the sub-acute and chronic phases (n = 20, n = 12, respectively), sitting balance improved after practice, resulting in better and further reaching and faster standing up. There was, however, no carry-over to walking (Dean & Shepherd 1997,Dean et al. 2007). Two studies showed that restriction of trunk movements with reach and grasp training led to further arm reaching with diminished trunk bending while grasping an object (Michaelsen & Levin 2004,Michaelsen et al. 2006). Both studies were performed with patients in the chronic phase after stroke.

Sit to stand

One study (randomization not blinded) focused on patients standing up from a chair. Repetitive sit to stand exercises plus symmetrical standing training resulted in statistically significant improvement in sit to stand performance and a decrease in the number of falls (< 0·05) (Cheng et al. 2001).


Seven studies that focused on walking of patients were included. In four of these, the intervention was task-oriented walking training with outcome measures of balance, gait speed and ADL. One study of good quality showed strong evidence for improved functional mobility after gait-oriented agility training in a circuit, even during the early stages after the stroke (Blennerhassett & Dite 2004). Two studies of good quality were performed in the community (Salbach et al. 2004) and follow-up in 2005 of the same population was reported (Salbach et al. 2005). Task-oriented walking enhanced self-efficacy in walking by statistically significant amounts. One study (Dean et al. 2000) with a small sample (n = 12) showed statistically significant improvement (< 0·05) in the strength of the affected leg during sit to stand after circuit training in which one of the workstations consisted of sit to stand exercises from various chair heights. Moreover, participants’ walking speed increased. Three studies of high quality compared the effects of patients walking on the ground with patients walking on a treadmill (Nilsson et al. 2001, Richards et al. 2004, Peurala et al. 2005). In all these studies, which were performed in both the sub-acute and chronic phases, scores on walking tests improved but no differences were found between the two groups.

Physical condition and muscle strength

Three systematic reviews were found that focused on the improvement of physical condition among stroke survivors. One review, including nine RCTs, showed that aerobic exercise had statistically significant favourable effects on walking velocity and walking endurance (Pang et al. 2006a). The other two systematic reviews drew the same conclusion: functional programmes on cardio-respiratory fitness improved walking performance (Saunders et al. 2004, Van de Port et al. 2007). Five studies used interventions relevant to nursing practice. One showed that training on a cycle-ergometer in the sub-acute phase after stroke resulted in a trend toward functional improvement, and stair-climbing improved statistically significantly (Katz-Leurer et al. 2003). The other four studies, performed in the community, used an exercise programme focusing on strength, balance and endurance. There was good evidence that exercising benefits fitness and walking capacity (Duncan et al. 1998, 2003, Pang et al. 2005). Task-oriented progressive resistance training in a circuit with sit to stand, turning and stepping exercises could improve lower limb muscle strength and functional performance (Yang et al. 2006).

Two systematic reviews of muscle strength training in stroke survivors concluded that such training has uncertain effectiveness on functional performance, except for sit to stand and step-up exercises (Ada et al. 2006, Bohannon 2007).


In a systematic review, van der Lee et al. (2001) found insufficient evidence for the effectiveness of exercise therapy on arm function (van der Lee et al. 2001). In less severe paretic patients, arm function improved more after functional training than after strength training and usual care (Winstein et al. 2004). A functional training programme aimed at improving arm participation proved ineffective: the control group, who received a walking programme combined with ADL tasks such as carrying groceries, performed even better on arm tests. Improvement was only observed in patients who entered the study with better arm performance (Higgins et al. 2005).

Constraint induced movement therapy

In a systematic review including nine small RCTs, CIMT had a positive effect on improving arm function, but it is impossible to draw definitive conclusions because of the methodological variety of the studies (Bonaiuti et al. 2007). Recently, two studies of good quality were published. The Extremity Constraint Induced Therapy Evaluation (EXCITE) trial was a prospective, single blind RCT (n = 222) conducted at seven centres. CIMT showed statistically significant relevant clinical improvement of arm function in patients who had had a stroke within the previous 3–9 months (Wolf et al. 2006). Patients in the mild to moderate chronic phase receiving intensive training following the CIMT protocol, with a restraint on the non-affected arm for 90% of waking hours, showed a statistically significant improvement in the functional use of their arms compared with the control group (Taub et al. 2006a). A placebo controlled trial (n = 41) also showed positive effects of mCIMT on arm function (Wu et al. 2007).

Mental practice or mental imagery

The preliminary results of case studies on the effect of mental practice are positive. Recently, the first RCT of high quality (n = 32) was published. The results show statistically significant reductions in arm impairment and an increase of arm function in daily activities in the mental practice group members (P < 0·0001) (Page et al. 2007b).


This review shows important evidence in favour of task-oriented training in daily nursing care. Balance training is more effective when it is related to a task (Bayouk et al. 2006). Sit to stand exercises result in improved standing-up and may reduce falls (Cheng et al. 2001). Walking on the ground has the same effect as walking with technical assistance such as treadmill training (Peurala et al. 2005). Arm training integrated into tasks has been shown to be effective (Higgins & Green 2005).

There were several limitations to this review. First, since no studies specifically addressing task-related training and nursing were identified, the findings need to be extrapolated to the nursing situation. This may seem arbitrary. However, in view of the strict method and criteria used, the findings are considered to be valid for nursing practice. Second, as we identified many interventions with very different study designs and outcome measures, it was impossible to conduct a meta-analysis of pooled results, and so the findings were described qualitatively. Third, we did not include papers in languages other than English, which may limit the generalizability of the findings.

In this systematic review, we found various interventions that proved to have important effects on patient outcomes. Generally, exercise tasks need to be specific, and should be practised as meaningful tasks (Van Peppen et al. 2004, Van de Port et al. 2007). There is a positive connection between improvement of physical fitness and functional performance. Fitness training should be an important component of stroke rehabilitation, as endurance after stroke is often compromised to a level that limits basic functioning in daily life (Pang et al. 2006a). The reduction in falls after agility training is an encouraging finding (Marigold et al. 2005) but larger studies are necessary. The American Heart Association recommended that rehabilitation training for stroke survivors should include circuit training and balance activities, and emphasised that nurses need to stimulate and provide opportunities to practise (Gordon et al. 2004). Scrutiny of the content of the exercises used in circuit training revealed that many are task-related, such as standing up from a chair (Pang et al. 2006a). As strength training is beneficial in improving functional outcome, it is remarkable that it is not always incorporated into rehabilitation programmes after stroke (Teasell et al. 2006).

It is uncertain whether enhanced therapy improves upper-arm function in patients with little voluntary arm movement (Teasell et al. 2006). Training the arm and lower extremity simultaneously with integrated meaningful tasks (meal preparation and housework) seems to improve specific functional activities (Higgins et al. 2005). When investigating the importance of a meaningful task, use of a favourite drink in an arm-reaching exercise elicited better performance (Wu et al. 2001). The outcomes of a study using modified CIMT with patients with minimal arm function are promising (Page & Levine 2007a). Nurses can play an important role in supporting and guiding patients during this demanding therapy.

Mental practice is a relatively new intervention. The therapy is based on the learned non-use phenomenon found in animal research in 1977 (Taub et al. 2006b). The first published placebo-controlled study of mental practice corroborates the efficacy of exercise programmes incorporating mental practice (Page et al. 2007b), and it should be seen as a promising technique for motor rehabilitation (Mulder 2007). Mental practice certainly offers possibilities for nurses since the movements can be imagined in any context relevant to the patient. Also, observation of an action performed by another person, combined with active execution of the same action, was tested with stroke survivors. With the additional component of observation, the impact of the intervention was statistically significantly greater than physical training alone. The authors suggest that application of this observational component of daily activities could enhance therapeutic effects (Ertelt et al. 2007).

Only 11 of the 33 RCTs included were performed in the (sub)-acute phase, but interventions such as sit to stand exercises, effective in the chronic phase, are feasible in the sub-acute phase as well.

Translation to daily nursing practice

Various authors have explored the roles of nurses in rehabilitation. Long et al. 2002) identified six roles: assessment, co-ordination and communication (also involving family members), therapy integration, emotional support and technical and physical care. Nolan and Nolan (1998)added the important contribution of being present throughout the day. Several researchers have explored the use of time with stroke survivors during the day. Bernhardt et al. (2004, 2007) showed that during the first 14 days after stroke the real time spent in rehabilitation activities to improve mobility and prevent further complications was only 13% of the active part of the day (between 8 AM and 5 PM), and more than 50% of the time patients were resting and waiting in bed. Overall, they were inactive and alone more than 60% of the time (Bernhardt et al. 2004, 2007). Similar results were found in a study comparing therapy time in four centres in Europe: therapy time ranged from 10% in the UK to 27% in Switzerland. Nursing care ranged from 35% of the total therapy time in the UK to 5% in Germany (De Wit et al. 2005). The follow-up study showed that, as a consequence, personal care was better in the UK but functional recovery was better in the other centres (De Wit et al. 2007). Nursing time could be used more efficiently if nurses integrated functional tasks into daily care. The effectiveness of group training is an important finding for nurses (Dean et al. 2000, Marigold et al. 2005, Pang et al. 2006b). Patients may be encouraged to practise simple ADL together under the supervision of nurses and other professionals. Assisting patients with mobility and ADL are nursing interventions that need to be further developed, and the effects on the nursing situation need to be studied. Nurses can implicitly guide balance exercises during standing up and sitting down, for example while helping with dressing. They may train patients in reaching by putting objects further away than arm length, which may improve balance. Further, they may assist patients to exercise from sitting to standing from chairs of different heights. Patients may also be encouraged to practise outside the therapeutic session with dressing, laying the table and eating together.

A well-coordinated multidisciplinary team of professionals is important in rehabilitation after stroke. While each discipline has its own paradigm, most paradigms overlap to some extent. Therefore, multidisciplinary collaboration among healthcare professionals is important. This is because many interventions are undertaken to a greater or lesser extent by different team members (Wade 2005). Various studies have shown positive effects of nurse-led training, including a range of motion exercise programme (Tseng et al. 2007). Therefore, optimal multidisciplinary collaboration is of the utmost importance. If the patient is discharged to home, it is especially recommended that participation in a fitness programme be stimulated (Duncan et al. 2005, Langhammer et al. 2007). Nurses can emphasise the importance of practising functional skills and can endorse participation in physical exercise programmes.

There are no standardised functional tasks, and so nurses are challenged to construct meaningful tasks. The challenge is to increase the therapeutic input of nurses and other professionals in a time of staff shortages and cost restriction (Walker 2007). Possible interventions are summarised in Table 3. Commitment of the leaders is essential if changes in nursing practice are to be accomplished and support is particularly needed in relation to time, education and therapeutic space in a department. After making an inventory, it is sensible to start with a few interventions. The present review has highlighted many factors to consider in future investigations measuring the effects of task-oriented interventions within nursing practice.

Table 3.   Practical recommendations for daily nursing practice
 Nurses may use the following activitiesPositively influence
Balance exercisesAssist patients with implicitly balance exercise during sit-to-stand, dressingImproves balance, self efficacy
Sit and reachPutting objects further away than arm length so that patients need to reach for an objectSitting balance
Sit to standExercise sit-to-stand a few times a day, from chairs with different heightsSit to stand, balance. May decrease number of falls
WalkingWalk patients with or without help (in a safe situation) from patients own room to other places in the unit/ward This may be done a few times a dayMay improve walking ability, walking speed and endurance
Physical fitnessEncourage the patient to practise outside the therapeutic sessions. It would be very helpful to have some cycle ergometers in the ward.Cycling improves functional status and physical condition
Physical fitnessPromote physical fitness classes and social walking for people living at homeImproves walking speed and endurance
Arm trainingActivate the paretic arm and hand. Exercise repetitive functional, meaningful tasks that are chosen by the patient
For example: self-care tasks, reading a book, eating
Improves arm function especially in patients with some arm and hand function. Improves self-care
Constraint induced movement therapyEncourage (chronic) patients with some hand function to take part in a CIMT group. Nurses can performe a part of the treatmentImproves arm and hand function
Mental practice and action observationGain more in-depth knowledge about mental practice and start with a few tasks following the protocol stated in the literature Encourage observation of actions performed by othersMay improve functional outcome
Agility trainingOrganize opportunities for step training (10 cm high)May improve balance
Activities of daily livingPractise dressing a few times a day with the same jacket or T-shirt Observe at which point of the dressing process the patient has difficulties and focus practising the aspect the patient finds difficultImproves self-care, Self-efficacy
Group training May improve functional outcome
Self-efficacy trainingEncourage the patient to do activities that he/she finds difficult to perform. Provide positive feedback and praise every improvement how minor it will beMay improve functional outcome
Activities already done in a wardCreate activities such as lay the table, eating together, care for flowers, distributing postcardsMore, intensive therapy improves rehabilitation outcome

The question is not if, but how nurses can incorporate task-oriented training into daily care.

Rehabilitation therapy can and should be integrated into the daily nursing care of stroke survivors.


Generally, task-oriented rehabilitation after stroke has proved to be effective and relevant for nursing practice. Improvement of impairments has long been seen as a prerequisite for functional movement, but interventions to achieve such improvement do not intrinsically carry over to functional improvement as the correlation between functional and laboratory measures is generally weak. A wide range of interventions, such as functional balance training during reaching and standing up, walking training, arm training and exercises for physical fitness need to be further developed and tailored to the patient’s needs, in close collaboration with other professionals. CIMT and mental practice are relatively new and promising treatment approaches that are appropriate in nursing practice, and they need to be further developed and explored. Moreover, nurses can play an important role in creating opportunities to practise meaningful functional tasks outside the regular therapy sessions. Indeed, the effects of such task-oriented training programmes provided by nurses in collaboration with other healthcare professionals need to be measured.

However, current evidence suggests that by actively using task-oriented training in the daily nursing care of stroke survivors, functional outcomes and overall health-related quality of life will be improved for these people.


This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Author contributions

MR, MS, EL & TBH were responsible for the study conception and design. MR & TBH performed the data collection. MR & TBH performed the data analysis. MR, MS, EL & TBH were responsible for the drafting of the manuscript. MR, MS, EL & TBH made critical revisions to the paper for important intellectual content. MS, EL & TBH supervised the study.