Efficacy of wearable devices for upper and lower limb rehabilitation in stroke patients: A systematic review and meta-analysis of randomized controlled trials

Abstract

Objective

Our aim was to explore the impact of wearable devices on rehabilitation of stroke patients.

Methods

We searched the CINAHL, Cochrane, Embase, PsycINFO, PubMed, Web of Science, China National Knowledge Internet, Wanfang, SineMed, and VIP databases to identify articles from inception through Janu 15th, 2026. The primary endpoint was limbs function, comprising upper limb function, lower limb function, and overall motor function, while secondary endpoints included quality of life and activities of daily living. The standard mean difference was taken as the effect value and the random effects method was adopted when I² was greater than 50%. The risk bias assessment employed the Cochrane Collaboration’s tool, while the sensitivity analysis and publication bias were respectively evaluated through the sequential exclusion method and the Egger’s method.

Results

This systematic review included 20 studies and a total sample of 1007 participants. Significant statistical differences were observed in upper limb function (SMD 0.27, 95%CI 0.09 to 0.45, P=0.004), lower limb function (SMD 1.58, 95%CI 0.31 to 2.85, P=0.01), overall motor function (SMD 4.59 95% CI, 0.80 to 8.38, P=0.02) and activities of daily living (SMD 2.61, 95%CI 1.18 to 4.04, P<0.01) among stroke patients.

Conclusion

Wearable devices are found to have positive impact on the limb function and activity of daily living of stroke patients, but have no significant intervention effect on quality of life.

Keywords

stroke wearable device systematic review meta-analysis

Introduction

Stroke is a common neurological disease characterized by decreased or interrupted blood flow in the brain, which affects brain function and leads to the loss of other bodily functions.^1,2 For example, limb motor dysfunction, dysphasia. Stroke has become the second leading cause of death and the leading cause of disability worldwide.³ The global annual prevalence and incidence of stroke are high. In the past 30 years, morbidity, mortality and the number of disability-adjusted life years have increased significantly. The arm function of 40% of stroke patients continues to decline.^4,5 One half of stroke survivors struggle to readjust to their original social roles, which poses a challenge for social and medical resources. Therefore, the management of the chronic effects of stroke has emerged as an urgent necessity.^6,7 The adoption of a self-management model is considered a way of providing home-based rehabilitation that addresses the challenges facing health care providers.⁸ Owing to the lack of professional rehabilitation guidance, most patients with cerebral infarction have poor self-management ability and rehabilitation compliance; therefore, patients need professional guidance and support during the rehabilitation process after discharge.⁹

With the advancement and development of Internet of Things technology, wearable devices, including activity trackers (Fitbits), pedometers, smart devices and other devices, have emerged as convenient and emerging tools for disease management in recent years.^10,11 The emergence of these devices provides a potential way for patients to self-monitor and manage their diseases.¹² At present, wearable devices are mainly used for rehabilitation improvement and monitoring function, with the aim of improving the quality of life of patients and evaluating the recovery effect of stroke survivors.^13,14 The results of study showed that it is feasible to use wearable hand devices to facilitate functional rehabilitation in stroke patients.¹⁵ Rehabilitation programs in which wearable rehabilitation devices are used can improve stroke patients’ upper limb function and daily living activities, thus effectively promoting their rehabilitation process.¹⁶ Wearable devices have not only made significant progress in facilitating functional restoration but are also increasingly used for monitoring function. Studies have shown that wearable devices can enhance patients’ perceptual ability via vibration stimulation and accelerate the process of recovery.¹⁷ Pedometers and wearable sensors have been shown to effectively promote physical activity through different behavioral modification methods, including setting goals, providing feedback, and monitoring activity levels, which are important for developing an activity plan.^18,19 A study by Hall KS et al. revealed that the use of a pedometer to increase physical activity has a positive effect on decreasing the risk of cardiovascular disease.²⁰ The use of Fitbits is feasible and acceptable in children with chronic diseases and adults with a high risk of cardiovascular disease.^11,21 Smart watches and bracelets are gradually being used in the field of healthcare because of their many advantages with respect to convenience and comfort.²² Two systematic reviews by Parker et al. and Powell et al. investigated the evidence for the use of wearable devices such as electrical stimulation and robotic devices for upper and lower limb rehabilitation, respectively.^2,23 Nevertheless, the above-referenced studies are exclusively limited to exploring the applied effectiveness of wearable devices for upper or lower limb function.

Therefore, it is crucial to conduct a pooled analysis of the application efficacy of wearable devices in stroke patients. The purpose of this study was to conduct a systematic review and meta-analysis of randomized control trials (RCTs) to determine whether the use of wearable devices can promote rehabilitation to provide evidence for clinical medical staff.

Methods

This review was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 statement²⁴ (registration number CRD42024558176).

Information sources

We searched ten electronic databases from inception through Jan. 15th, 2026: the PubMed, CINAHL, Web of Science, Cochrane Library, Embase, PsyclNFO, China National Knowledge Internet, Wanfang, SineMed, and VIP databases. First, included records were uploaded to EndNote V.X21, after which duplicates were removed.

Search strategy

A search strategy was developed and implemented, which included three terms: (1) stroke or ischemic stroke or hemorrhagic stroke, (2) wearable devices, and (3) RCTs. This strategy was adjusted to meet the requirements of different databases. The detailed search strategy is described in Supplemental Material, Table S1. In addition, the reference lists of studies included in previous articles were manually screened to identify additional studies.

Eligibility criteria and study selection

The inclusion criteria were as follows: (1) RCTs; (2) studies including participants diagnosed with stroke; (3) Wearable devices were divided into two main applications: rehabilitation training and monitoring. Rehabilitation training included stimulation-based devices (wearables for functional electrical stimulation) and robotic-assisted devices (exoskeleton robots for targeted training). Monitoring covered physiological parameter monitoring and daily living activity tracking; (4) studies in which the control group received routine nursing; and (5) studies in which the outcomes were clinical outcome indicators.

The exclusion criteria were (1) limb dysfunction caused by other reasons; (2) the inability to obtain complete data; (3) non-English or non-Chinese articles; and (4) qualitative research.

Two researchers (CL and NJ) independently screened the titles and abstracts for the initial assessment. The full-text articles were assessed independently by the two researchers to identify the final articles for inclusion. Disagreements were resolved through discussion and consultation. If the data were published in multiple studies, the most complete and effective analysis was included.

Data extraction

To summarize the effects of wearable device on rehabilitation outcomes related to stroke, two researchers (CL and NJ) independently extracted the data. Disagreements in data collection were resolved through consensus or discussion. Several Chinese articles were included in this study. To ensure the accuracy of data extraction, key study information from Chinese articles was independently translated into English by two researchers. The extracted data of the included studies consisted of the following: (1) study characteristics (first author, year of publication, country, and sample size); (2) population characteristics (number of women, age, stroke type); (3) intervention and control group measures; and (4) outcome measurements. Data on the means and standard deviations (SDs) of upper limb function (including upper motor function and grip strength), lower limb function (including lower motor function and balance ability), overall limb motor function, quality of life and activities of daily living at the follow-up time point closest to the end of the treatment period were extracted. Given that most of the included studies only reported post-intervention means for the outcomes, a fixed-effects or random-effects model based on post-intervention means was employed for meta-analysis to pool the results. This approach ensures the maximal utilization of valid data from the included literature.

Definitions of terms

Stroke included hemorrhagic and ischemic stroke in this meta analysis. Wearable devices mainly referred to devices that can be worn on the body to assist the affected limbs in rehabilitation exercises and monitor patients’ activity information. In this meta-analysis, rehabilitation outcomes mainly included upper limb function (including motor function and grip strength), lower limb function (motor function and balance ability), overall motor function, activities of daily living, and quality of life, as determined by a review of the included studies. Rehabilitation outcome mainly included upper limb function (including motor function and grip strength), lower limb function (including motor function and grip strength), overall limb motor function, activities of daily living, and quality of life in this meta analysis according to reading articles.^23,25,26 In addition, activities of daily livings were defined as the ability of stroke patients to perform a series of foundation activities to meet basic life needs. Motor function was defined as motor quality on the basis of joint movement or functional activity. Grip strength was defined as the degree of a patient’s muscle strength.

Outcomes

Limb function was primary outcome, which was divided into three subgroups: upper limb function, lower limb function and overall limb function. The indicators of upper limb function consisted of the upper limb motor function (Fugl-Meyer Assessment of Upper Extremity) and grip strength. Lower limb function included the lower limb motor function (Fugl-Meyer Assessment of Lower Extremity) and balance ability. Overall improvement was a secondary outcome, which included activities of daily living (ADL) and quality of life (QOL). The scales used to assess motor function included the Fugl-Meyer Assessment, Fugl-Meyer Assessment of Upper Extremity, Fugl-Meyer Assessment of Lower Extremity, Wolf Motor Function Test,²⁷ Action Research Arm Test, or Rivermead Motor Index.²⁸ Grip strength was evaluated and measured using a dynamometer. Balance ability was assessed using the Berg Balance Scale. The assessment of quality of life was performed with the Short Form-36 Health Survey and the European Quality of Life Five-Dimension Scale. The Emotional, Social, and Cognitive Assessment scale, Modified Barthel Index and Barthel index, were used to assess independence in activities of daily living. The outcome after the intervention was selected as the time point closest to the end of the intervention.

Risk of bias assessment

The risk of bias of the included studies was independently assessed by two researchers (CL and NJ) via the Cochrane Collaboration’s tool for assessing the risk of bias (low, unclear or high risk).²⁹ This tool has six domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data and selective reporting. Disagreements were resolved through discussions and consensus. The quality of evidence was evaluated using the Grades of Recommendation, Assessment, Development and Evaluation (GRADE) evidence grading system.³⁰

Statistical methods

Standardized mean differences (SMDs) and 95% confidence intervals (CIs) were used as a summary of effect values due to the outcome indicators being measured by different scales. If I² was > 50%, the random effects DerSimonian‒Laird method was adopted³¹; otherwise, a fixed effects model was used. The findings of the meta-analyses are presented via forest plots. Publication bias was assessed via visual inspection of a funnel plot and a random-effects version of Egger’s regression test. If fewer than 10 primary studies were included, only a funnel plot was generated and visually inspected. To check the robustness of the primary analysis, sensitivity analyses were carried out via the stepwise rejection method if the outcome was included in more than two articles. Heterogeneity was assessed via I² values, which were considered to indicate low (0%-40%), moderate (50%-75%), and high heterogeneity (75% -100%) according to the Cochrane Manual.³² The significance level for all the statistical tests was set at p<0.05. All analyses were performed via RevMan 5.4.1 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration) and Stata 16.0 (Stata Corporation, College Station, Texas, USA).

Differences from published protocols

The final protocol of this review and meta-analysis deviated from the originally registered protocol in the following aspects: Since some studies did not report the sex ratio, subgroup analysis by sex was not performed. Lower limb functional recovery was added as an outcome measure.

Results

A total of 13336 articles were initially obtained by searching seven databases. A total of 834 was excluded because of duplication. The flow chart of the study selection process is presented in Figure 1. There were 20^{13,16,33–50} RCTs included in this study, involving a total sample of 1007 participants. Three original Chinese studies were included. In accordance with population (P), intervention (I), comparison (C), outcome (O), and study (S) principle, studies that did not meet the requirements were excluded (patient population (n=13), type of intervention (n=39), missing or incorrect representation of the outcome (n=22), and type of article (n=28)).

Figure 1.

Flow chart of included studies.

Risk of bias assessment

The risk of bias map was shown in Figures 2 and 3. There were three articles with high risk of bias, 14 articles with moderate risk of bias, and three articles with low risk of bias in terms of the overall risk of bias. In random sequence generation, 18 studies were classified as low risk of bias, and one study was classified as having a ‘high risk’ or ‘unclear’ risk of bias. In terms of allocation concealment, 13 studies reported that involving allocation concealment were ‘unclear risk’. For performance bias, 2 studies were “low risk”. Participant blinding was unachievable in the intervention group due to wearable device, leading to a high risk of performance bias; however, most studies implement adequate assessor blinding, which further resulted in a low risk of detection bias (14 low risk bias). Finally, the studies were rated as a low risk of bias in selective reporting and three studies were rated as unclear risk of bias in other bias. The result of certainty of the evidence was shown in the Supplemental Material, Table S6. For the Fugl-Meyer Assessment, Fugl-Meyer Assessment of Upper Extremity, Fugl-Meyer Assessment of Lower Extremity and Activities of Daily Living, the certainty was rated as moderate.

Figure 2.

Risk of bias judgment.

Figure 3.

Percentages of risk of bias items.

Study characteristics

This review included 20 studies with 2 arms, for a total of 40 arms, including 1007 participants. Among the 1007 participants, 167 were diagnosed with cerebral hemorrhage, 532 had cerebral ischemia, and the remaining 308 subjects’ stroke types were unclear. The included articles were published between 2008 and 2025. The research countries were concentrated in the United States and China. The participants were mainly aged 50 years and older. The intervention duration of these studies varied widely, ranging from three weeks to six months. Seven studies did not report stroke type. The full details of the 20 studies included in the review are shown in Table 1 and Supplemental Material, Table S2-S5.

Table 1.

Characteristics of included studies.

First author (year)	Country	Age	Sample Size (female)	Stroke type (hemorrhage/ischemic)	Intervention measures	Control measures	Duration	Outcome
LIN³³ 2018	Taiwan	52.2±10.2	18(NR)	7/11	Monitoring device +training	Routine training	5 weeks	FMA-UE
Chae³⁴ 2020	Korea	58.3±9.3	25(NR)	NR	Training Based on Smart Watch Monitoring	Routine training	12 weeks	FMA-UE, Grip Strength
Scronce³⁵ 2023	USA	60.0±11.0	12 (6)	3/9	TheraBracelet+rehabilitation training	Routine rehabilitation training	4 weeks	Hand Function Outcome
Lawrie³⁶ 2018	UK	53.0±12.0	30 (7)	NR	Body movement+ smartwatch feedback	Non-feekback body movement	3 weeks	RMI, Grip Strength, BI, EQ-5D-5L
Ruth H³⁷ 2019	UK	73.0 (65.0-80.0)	33 (20)	2/31	Wristband+usual therapy	Usual therapy	4 weeks	ARAT, Motricity Index, BI
Wang³⁸ 2021	China	NR	92(NR)	0/92	smart bracelets+intervention	Routine intervention	3 months	BI
Zhuang³⁹ 2019	China	67.1±4.6	120 (40)	0/120	Bracelet intervention	Routine intervention	6 months	SF-36
Li⁴⁰ 2020	China	46.1±7.5	64 (26)	NR	Sports bracelet +functional training	ADL and functional training	3 months	MMT,BI
Lu⁴¹ 2025	China	61.5± 11.2	79(39)	31/48	Wearable Contextual Intelligent Rehabilitation Training System	Conventional rehabilitation treatment	8weeks	FMA, MBI
Guo¹³ 2023	China	56.3±9.8	120(34)	NR	Remote rehabilitation training system+ wearable devices	Conventional clinical physical therapy	3 weeks	FMA
Yang⁴² 2021	China	44.4± 10.2	25(4)	NR	Wearing custom 3D-printed dynamic wrist splint for rehabilitation	Conventional rehabilitation treatment	6 weeks	FMA-UE
Park¹⁶ 2021	Korea	60.6± 18.1	44(20)	NR	Game-based virtual reality rehabilitation training+wearable devices	Conventional physical therapy	4 weeks	FMA-UE, Grip strength, MBI
Villafañe⁴³ 2017	Italy	67.0 ±11.0	32(11)	8/24	Robot-assisted hand movement training	Conventional physical therapy	3 weeks	BI
Alon⁴⁴ 2008	USA	NR	NR	NR	Customized forearm/hand orthosis stimulates	Conventional physical therapy	12 weeks	FMA-UE
Chang⁴⁵ 2021	Korea	60.2±14.7	75(30)	31/44	Exoskeleton robot-assisted gait training	Conventional physical therapy	4 weeks	BBS, QOL
Wang⁴⁶ 2024	China	62.9±3.1	80(33)	18/62	Internet-based remote home-based rehabilitation guidance + Lower limb wearable device training	Regular post-discharge follow-up guidance	12 weeks	FMA, BBS, MBI
Wei⁴⁷ 2018	China	59.2±11.3	57(12)	15/42	Wearable device vibration alert	Wearing sham wearable device	4 weeks	FMA-UE
Kim⁴⁸ 2025	Korea	64.6± 9.4	60(23)	34/26	Wearable robot-assisted gait training	Conventional physical therapy	8 weeks	FMA-LE, BBS
Barry⁴⁹ 2012	USA	52.7±16.1	19(9)	8/11	Wearing dynamic wrist orthosis for training	Therapist’s manual assistance training	6 weeks	Grip strength
Watanabe⁵⁰ 2014	Japan	67.0±16.8	22(11)	10/12	Single-leg version of HAL on the hemiplegic side	Conventional physical therapy	4 weeks	FMA-LE

ARAT, Action Research Arm Test; BBS, Berg Balance Scale; BI, Barthel Index; EQ-5D-5L, EuroQol Five Dimensions Questionnaire; FMA, Fugl-Meyer assessment; FMA-UE, Fugl-Meyer Assessment of Upper Extremity, FMA-LE, Fugl-Meyer Assessment of Lower Extremity; MBI, Modified Barthel Index; MMT, Manual Muscle Testing; NR, Not reported; RMI, Rivermead Mobility Index; SF-36, 36 item Short Form Health Survey.

Synthesis of results

Upper limb function, lower limb function and overall limb function were the primary outcomes, with 12, 5, 3 studies, respectively. Upper limb function included upper limb motor function and grip strength, with 10 studies (n=307) and 6 studies (n=192), respectively. Lower limb function included lower limb motor function (3 studies, n=142) and balance ability (3 studies, n=215). Overall limb function included 3 studies and 240 participants. Activities of daily living (8 studies and 433 participants)and quality of life (3 studies and 210 participants) were the secondary outcome.

The research found that wearable devices are effective for both upper limb and lower limb rehabilitation, but the statistical significance of the results for upper limb rehabilitation is greater than that for lower limb rehabilitation. The result of upper limb function was statistically significant (0.27,95% CI 0.09 to 0.45, P=0.004, I²=0%), as shown in Figure 4. The upper limb motor function subgroup was statistically significant. (0.28, 95% CI 0.05 to 0.51, P=0.02, I² =0%, GRADE= moderate) However, the SMD in grip strength subgroup was 0.25 (95% CI -0.04 to 0.54, P=0.09, I² =0%, GRADE= low). It means that there was no statistical difference between the intervention and control groups in the grip strength. Therefore, a subgroup analysis for grip strength was performed based on intervention duration, and the results showed that statistically significant outcomes were observed in the subgroup with an intervention duration of less than one month (SMD 0.48, 95% CI 0.04 to 0.91, P=0.03 I²=0, see Supplemental Material, Figure S1). For the upper limb motor function subgroup, the visual inspection of the funnel plot (see Supplemental Material, Figure S2) and Egger’s regression test for funnel plot asymmetry (p=0.29) indicated that publication bias was unlikely to have influenced the results. For the grip strength subgroup, since fewer than 10 studies were included, a funnel plot was used to assess publication bias. As shown in Supplemental Material, Figure S3, publication bias was unlikely to have influenced the results. According to sensitive analysis, the results were stable for upper limb motor function subgroup and grip strength subgroup, as shown in Supplemental Material, Figure S4.

Figure 4.

Forest plot for upper limb motor function and grip power.

The result of lower limb function was statistically significant (SMD 1.58, 95% CI, 0.31 to 2.85, P=0.01 I² =96%), as shown in Figure 5. However, lower limb motor function indicator no statistically significant results (SMD 0.29, 95% CI, -0.10 to 0.68, I² =23%, GRADE=moderate), and limb balance ability demonstrated statistical significance (SMD 3.34, 95% CI, 0.33 to 6.36, I² =98%, GRADE=low). According to sensitive analysis,for lower limb motor function, results were stable after sequential exclusion of Guo et al. and Kim et al., but statistical significance appeared when Hiroki et al. was removed, as shown in Supplemental Material, Figure S5. Subgroup analysis was not performed as only three original studies were included; instead, stratified analysis was adopted. According to the results of stratified analysis, the study by Hiroki et al. was judged to be at high risk of bias (detection bias and performance bias), while the remaining two studies were at low-to-moderate risk of bias. Heterogeneity in methodological quality was identified as the core reason for the lack of statistical significance in the original pooled results. Sensitivity analysis of the balance ability subgroup showed that the results became statistically non-significant when the studies by Chang et al. and Kim et al. were excluded individually (SMD 5.16, 95%CI -4.20 to 12.53, I²=99%; SMD 5.05, 95%CI -4.53 to 14.63, I²=99%), whereas the pooled results remained statistically significant after the exclusion of the study by Wang et al. (SMD 0.29, 95%CI -0.05 to 0.63, I²=0%).

Figure 5.

Forest plot for lower limb motor function and balance ability.

The result of overall limb function was statistically significant (4.59,95% CI 0.80 to 8.38, P=0.02, I² =99%, GRADE=moderate), as shown in Figure 6. Sensitivity analysis results showed that the results became statistically non-significant when the studies by Wang et al. and Lu et al. were excluded individually (SMD 2.14, 95CI% -1.01 to 5.29, P =0.18 I²=99%; SMD 5.11, 95CI% -3.91 to 14.13, P=0.27, I²=99%).

Figure 6.

Forest plot for overall limb motor function.

For the activities of daily living, the implementation of wearable devices was shown to enhance patients’ abilities in performing activities of daily living (SMD 2.61, 95% CI 1.18 to 4.04, P<0.05, I² 97%, GRADE=moderate). According to sensitive analysis, the results were stable for activities of daily living, as shown in Supplemental Material, Figure S6.A. Due to the high heterogeneity, a subgroup analysis was conducted based on the intervention duration. The results showed that when the intervention period was less than or equal to one month, the heterogeneity decreased (SMD 0.25, 95%CI -0.12 to 0.61, P=0.18, I² 0%). Only three studies were included in the quality of life analysis, with extremely high statistical heterogeneity (I²=96%, P<0.001). According to sensitive analysis, when the research by Zhuang et al. was excluded, the results became statistically significant. (Supplemental Material, Figure S6.B). After the removal of this study,³⁹ the heterogeneity of the remaining two studies was significantly reduced (I²=0%, P=0.37). The research conducted by Zhuang et al. indicates that rehabilitation based on wearable devices can significantly improve the quality of life. However, the results of the other two studies showed small differences and negative outcomes.

Discussion

The purpose of this study was to explore the effects of wearable device on stroke patient. This finding suggests that wearable devices improved upper limb motor function, balance ability, and activities of daily living in stroke patients, but had no notable effects on grip strength, lower limb motor function, or quality of life.

Through sensitivity analysis, it was found that several indicators had unstable results. In this study, it was found that the pooled results of lower limb motor function indicators were significantly influenced by the study by Hiraoki et al. A potential reason is that this study was assessed a high risk of bias, which led to biased estimation of its effect size. After excluding this study, the pooled results showed statistical significance, suggesting that the outcome indicator was largely affected by methodological quality. The sensitivity analysis of limb function in this study shows that the studies by Wang et al. and Lu et al. are the key studies influencing the combined effect size. The possible reasons for this are that the sample size weights in Wang and Lu’s studies accounted for a higher proportion.

High heterogeneity (I²> 90%) was observed across multiple pooled outcomes in this meta-analysis, despite the small sample sizes of the included studies. There was high heterogeneity in the balance ability subgroup, the main reasons for this are as follows: firtly, there are significant differences in the intervention durations of the three studies. The intervention duration affects the recovery progress of balance ability, which is a contributing factor to the heterogeneity. Secondly, the effect size of Wang’s single study was relatively high, leading to high heterogeneity in this outcome indicator. For high heterogeneity of quality of life, it was mainly attributed to one study identified by sensitivity analysis. This study had distinct characteristics from the other two, including intervention time. The intervention time of Zhuang et al. lasted for 6 months, while the other two studies only lasted for 3 to 4 weeks. Due to the only three studies included for quality of life, subgroup analyses or meta-regression could not be performed to further verify the impact of these factors, which limited the in-depth exploration of heterogeneity sources. For the indicators of balance ability and quality of life, only three studies were included respectively. The sample sizes were small, and the reliability of the conclusions was limited. For the high heterogeneity of ADL, the results of the subgroup analysis indicated that the intervention time had an impact on this indicator. Therefore, when referring to ADL outcome, a cautious attitude should be maintained. In the future, more high-quality and large-sample RCTs are needed.

Most stroke survivors experience some degree of limb dysfunction due to motor paralysis.⁵¹ Stroke patients can effectively control the recurrence of the disease through self-management and rehabilitation training in their daily lives while actively promoting the recovery of their physical functions.^52,53 If patients develop hemiplegia symptoms and are not treated in time, they lose their ability to take care of themselves, which has a very adverse impact on quality of life.⁵⁴ However, the positive effects of wearable devices in this study may be attributed to the following factors: for one thing, various wearable sensors integrated with functions such as physiological monitoring, pressure sensing, and inertial sensing^55,56 offer functional and specific hand training for patients; for another, these devices enable real-time, dynamic feedback and limb motor status in stroke patients, as well as assistance with their limb movements.⁵⁷ This is consistent with the findings of Chae et al., who reported that a wearable device-based smartwatch rehabilitation system could effectively improve motor function scores.³⁴ Previous studies have indicated that the estimated minimal clinically important difference (MCID) for upper limb motor recovery in stroke patients at different stages is 9–13 points on the Fugl-Meyer Assessment for the Upper Extremity (FMA-UE).^58,59 However, the results of this meta-analysis showed that the improvement in the intervention group compared with the control group exceeded this MCID threshold. Therefore, this finding indicates that the improvement in upper limb motor function observed in the present study was not only statistically significant but also clinically meaningful.

Wearable devices can facilitate the improvement of balance ability by assisting with daily activity training (e.g., climbing stairs and walking on uneven surfaces). This training modality effectively enhances active balance control, weight shift, and muscle activation capacity through postural control, all of which are crucial for improving balance function in stroke patients.^60,61 Wearable robots can also break through the limitations of conventional rehabilitation training scenarios and enhance patients’ engagement and motivation through novel and diverse training settings, thereby facilitating the long-term improvement of balance ability and gait function.⁶² This is consistent with the findings of Brianne Darcy and Wang et al., who reported that wearable devices can help stroke patients achieve long-term improvements in balance ability.^63,64 Previous research has indicated that a 5-point increase in the Berg Balance Scale (BBS) constitutes a useful indicator for reducing walking assistance in early subacute stroke patients.⁶⁵ The improvement in balance ability in the present study exceeded the MCID threshold, indicating clinical significance. However, the result was high heterogeneity. Therefore, the findings should be interpreted with caution. Previous studies have reported that the MCID for activities of daily living in patients with stroke ranges from 1.9 to 5 points. The improvement in activities of daily living observed in the present study exceeded the MCID threshold, indicating that the effect of wearable devices on patient autonomy is not only statistically significant but also clinically meaningful.^66,67

The present study showed no significant improvement in grip strength. Subgroup analysis revealed that wearable devices had a significant positive effect on improving grip strength in the subgroup with intervention duration < 1 month, but no significant effect in the subgroup with intervention duration > 1 month. The potential reasons are as follows: firstly, short-term wearable intervention has an early training effect, which promotes a rapid increase in grip strength during the early rehabilitation phase; secondly, long-term intervention is associated with decreased treatment adherence and dropout data, leading to insufficient statistical power.

This study failed to find a significant effect on lower limb motor function and quality of life in stroke patients, and the potential reasons need to further interpretation. The potential causes are as follows: on the one hand, most wearable devices focus on postural control and gait rhythm in lower limb interventions, whereas the improvement of lower limb motor function requires high-intensity, multi-joint collaborative training.⁶⁸ On the other hand, only three original studies were included in this study on lower limb motor function, and the intervention was concentrated within 3–8 weeks, failing to further explode the long-term effect. Interpretation of the negative findings on quality of life outcomes are as follows: On one hand, only three original studies were included, which may have led to insufficient effect size. On the other hand, meaningful improvements in quality of life can only be achieved through sustained interventions lasting 3–6 months or longer.⁶⁹

This study had several limitations. Firstly, there was heterogeneity in some outcome indicators. As only 3 original studies were included, subgroup analysis was not conducted and only qualitative methods were used for analysis. Secondly, no formal power analysis was performed to determine the adequate sample size for detecting clinically meaningful effects. The sample sizes of some studies were relatively small and the follow-up periods were short, which might have certain impacts on the accuracy of the results. Thirdly, the number of included original studies was fewer than five for some outcome indicators in this study. In accordance with the Cochrane Handbook for Systematic Reviews of Interventions, publication bias tests were not performed for these outcomes. Therefore, larger and more in-depth studies are needed in the future to explore the impact of wearable devices on stroke patient.

Conclusion

In this meta-analysis, wearable devices are found to have positive impact on the upper limb motor function, balance ability and activity of daily living of stroke patients, but have no significant intervention effect on lower limb motor function. However, we need to carefully consider the results related to balance ability, activity of daily living and quality of life.

Supplemental material

Supplemental material - Efficacy of wearable devices for upper and lower limb rehabilitation in stroke patients: A systematic review and meta-analysis of randomized controlled trials

Supplemental material for Efficacy of wearable devices for upper and lower limb rehabilitation in stroke patients: A systematic review and meta-analysis of randomized controlled trials by Chang Liu, Baojian Wei, Xiaolei Wang, Yuzhen Xu and Ning Jiang in DIGITAL HEALTH.

Supplemental material

Supplemental material - Efficacy of wearable devices for upper and lower limb rehabilitation in stroke patients: A systematic review and meta-analysis of randomized controlled trials

Footnotes

Acknowledgements

We would like to express our sincere gratitude to Professor NJ and Professor YX for their contributions to the conceptualization and design of this study. Additionally, we thank Professor BW and Professor Wang for careful revisions to the article content.

ORCID iD

Chang Liu

Author contributions

CL conducted the analyses and wrote paper; BW and XW wrote paper; NJ and YX conceived the content, design of the study, contributed and approved the final version of the manuscript. All authors reviewed the manuscript.

Funding

The research was supported by the following fund: Shandong Provincial Natural Science Foundation, General Program (No. ZR2025MS1169) and the Shandong Provincial Humanities and Social Sciences Research Project (2025).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

No datasets were generated or analysed during the current study.

Registration and protocol

The protocol for this review was registered in the PROSPERO (CRD42024558176).

Guarantor

NJ.

Supplemental material

Supplemental material for this article is available online.

References

Dong-dong

ZHU

Zhen-lei

JIA

Xu-ting

, et al. Effects of health management based on narrative medicine model on self-managed cognitive function and limb function of perimenopausal stroke patients. Maternal and Child Health Care of China 2024; 39(10): 1883–1886.

Parker

Powell

Mawson

. Effectiveness of Upper Limb Wearable Technology for Improving Activity and Participation in Adult Stroke Survivors: Systematic Review. Journal of Medical Internet Research 2020; 22(1): e15981. https://doi.org/10.2196/15981

Kuriakose

Xiao

. Pathophysiology and Treatment of Stroke: Present Status and Future Perspectives. International Journal of Molecular Sciences 2020; 21(20): 7609. https://doi.org/10.3390/ijms21207609

Feigin

Brainin

Norrving

, et al.

World Stroke Organization (WSO): Global Stroke Fact Sheet 2022

International Journal of Stroke 2022; 17(1): 18–29. https://doi.org/10.1177/17474930211065917

Feigin

Forouzanfar

Krishnamurthi

, et al. Global and regional burden of stroke during 1990–2010: findings from the Global Burden of Disease Study 2010. The Lancet 2014; 383(9913): 245–255. https://doi.org/10.1016/s0140-6736(13)61953-4

Prince

Guo

, et al. The burden of disease in older people and implications for health policy and practice. The Lancet 2015; 385(9967): 549–562. https://doi.org/10.1016/S0140-6736(14)61347-7

Wolf

Spiers

Doherty

, et al. The effect of self-management education following mild stroke: an exploratory randomized controlled trial. Topics in Stroke Rehabilitation 2017; 24(5): 345–352. https://doi.org/10.1080/10749357.2017.1289687

Fryer

Luker

McDonnell

, et al. Self management programmes for quality of life in people with stroke. Cochrane Database of Systematic Reviews 2016; 2019(3): CD010442.

Bogle

BCV

. Adherence to medication and self-managment in stroke patients. British Journal of Nursing 2014; 23(3): 158.

10.

Meng

Zhang

Chen

, et al. Exploration of Human Activity Recognition Using a Single Sensor for Stroke Survivors and Able-Bodied People. Sensors (Basel) 2021; 21(3): 799. https://doi.org/10.3390/s21030799

11.

Bicki

Seth

McCulloch

, et al. Use of activity trackers to improve blood pressure in young people at risk for cardiovascular disease: a pilot randomized controlled trial. Pediatr Nephrol 2024; 39(8): 2467.

12.

Zhang

Xie

, et al. Wearable Health Devices in Health Care: Narrative Systematic Review. JMIR mHealth and uHealth 2020; 8(11): e18907. https://doi.org/10.2196/18907

13.

Guo

Wang

, et al. Clinical Study of a Wearable Remote Rehabilitation Training System for Patients With Stroke: Randomized Controlled Pilot Trial. Jmir Mhealth and Uhealth 2023; 11(1): e40416. https://doi.org/10.2196/40416

14.

Park

Jayaraman

. Enhancing the Quality of Life Through Wearable Technology. IEEE ENGINEERING IN MEDICINE AND BIOLOGY MAGAZINE 2003; 22(3): 41–48. https://doi.org/10.1109/memb.2003.1213625

15.

Sanders

Chan

Augsburger

, et al. Feasibility of Wearable Sensing for In-Home Finger Rehabilitation Early After Stroke. Ieee Transactions on Neural Systems and Rehabilitation Engineering 2020; 28(6): 1363–1372. https://doi.org/10.1109/TNSRE.2020.2988177

16.

Park

Lim

. Effects of a Rehabilitation Program Using a Wearable Device on the Upper Limb Function, Performance of Activities of Daily Living, and Rehabilitation Participation in Patients with Acute Stroke. International journal of environmental research and public health 2021; 18(11): 5524. https://doi.org/10.3390/ijerph18115524

17.

Nicholson

Sniehotta

Van Wijck

, et al. A Systematic Review of Perceived Barriers and Motivators to Physical Activity after Stroke. International Journal of Stroke 2012; 8(5): 357–364. https://doi.org/10.1111/j.1747-4949.2012.00880.x

18.

Lutes

CT-LL

. Why Do Pedometers Work? A Reflection upon the Factors Related to Successfully Increasing Physical Activity. Sports Med. 2009;39(12):981–993.

19.

Cleland

Alex

Madhavan

. Concurrent validity of walking speed measured by a wearable sensor and a stopwatch during the 10-meter walk test in individuals with stroke. Gait & Posture 2023; 107: 61–66. https://doi.org/10.1016/j.gaitpost.2023.09.012

20.

Hall

Hyde

Bassett

, et al. Systematic review of the prospective association of daily step counts with risk of mortality, cardiovascular disease, and dysglycemia. International Journal of Behavioral Nutrition and Physical Activity 2020; 17(1): 78. https://doi.org/10.1186/s12966-020-00978-9

21.

Venetsanou

Emmanouilidou

Soutos

, et al.

Towards a Functional Approach to the Assessment of Daily Life Physical Activity in Children: Are the PAQ-C and Fitbit Flex-2 Technically Adequate?

International Journal of Environmental Research and Public Health 2020; 17(22): 8503. https://doi.org/10.3390/ijerph17228503

22.

Holubová

Malá

Hoidekrová

, et al. The Accuracy of Commercially Available Fitness Trackers in Patients after Stroke. Sensors (Basel) 2022; 22(19): 7392. https://doi.org/10.3390/s22197392

23.

Powell

Parker

Martyn St-James

, et al. The Effectiveness of Lower-Limb Wearable Technology for Improving Activity and Participation in Adult Stroke Survivors: A Systematic Review. Journal of Medical Internet Research 2016; 18(10): e259.

24.

Page

McKenzie

Bossuyt

, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ, 2021.

25.

Thompson

Dawson

Legasto-Mulvale

, et al. Mobile Technology–Based Interventions for Stroke Self-Management Support: Scoping Review. JMIR mHealth and uHealth 2023; 11: e46558.

26.

Song

Qin

Suen

LKP

, et al. Effects of wearable device training on upper limb motor function in patients with stroke: a systematic review and meta-analysis. Journal of International Medical Research 2024; 52(10): 3000605241285858. https://doi.org/10.1177/03000605241285858

27.

Min

Yan

. Validity and Reliability of the Wolf Motor Function Test Scale for Evaluating Upper Limb Function in Acute Stroke Patients. Chinese Journal of Rehabilitation Medicine 2009; 24(12): 992–998.

28.

Collen Dtw

Robb

Bradsha

. The Rivermead Mobility Index: a further development of the Rivermead Motor Assessment. Int Disabil Stud 1991; 13(2): 50.

29.

Higgins

JPT

Altman

Gotzsche

, et al. The Cochrane Collaboration’s tool for assessing risk of bias in randomised trials. BMJ 2011; 343: d5928-d. https://doi.org/10.1136/bmj.d5928

30.

MC Master University and Evidence Prime 2021 . GRADEpro Guideline Development Tool. (Computersoftware). https://www.gradepro.org/

31.

DerSimonian

RLN

Laird

. Meta-analysis in clinical trials. Control Clin Trials 1986; 7(3): 177–188. https://doi.org/10.1016/0197-2456(86)90046-2

32.

Higgins

JTJ

Chandler

(eds). Cochrane Handbook for Systematic Reviews of Interventions version 6.3 (updated February 2022). Cochrane, 2022. Available from. https://www.training.cochrane.org/handbook

33.

Lin

, et al. Feasibility and efficacy of wearable devices for upper limb rehabilitation in patients with chronic stroke: a randomized controlled pilot study. European journal of physical and rehabilitation medicine 2018; 54(3): 388–396. https://doi.org/10.23736/S1973-9087.17.04691-3

34.

Chae

Kim

Lee

K-S

, et al. Development and Clinical Evaluation of a Web-Based Upper Limb Home Rehabilitation System Using a Smartwatch and Machine Learning Model for Chronic Stroke Survivors: Prospective Comparative Study. Jmir Mhealth and Uhealth 2020; 8(7): e17216. https://doi.org/10.2196/17216

35.

Scronce

Ramakrishnan

Vatinno

, et al. Effect of Self-Directed Home Therapy Adherence Combined with TheraBracelet on Poststroke Hand Recovery: a Pilot Study. Stroke research and treatment 2023; 2023: 3682898. https://doi.org/10.1155/2023/3682898

36.

Lawrie

Dong

Steins

, et al. Evaluation of a smartwatch-based intervention providing feedback of daily activity within a research-naive stroke ward: a pilot randomised controlled trial. Pilot and feasibility studies 2018; 4: 157. https://doi.org/10.1186/s40814-018-0345-x

37.

Da-Silva

Moore

Rodgers

, et al. Wristband Accelerometers to motiVate arm Exercises after Stroke (WAVES): a pilot randomized controlled trial. Clinical Rehabilitation 2019; 33(8): 1391–1403. https://doi.org/10.1177/0269215519834720

38.

Wang Sha-Sha

Yong-Li

Xian-Ju

, et al. Application of Internet Plus - Based Wearable Devices in Health Care after Stroke Patients Undergoing Stenting. World Latest Medicine Information (Electronic Version) 2021; 21(59): 049.

39.

Zhuang F-Juan

Yan

. Analysis on the Application Effect of Wearable Devices in Stroke Patients Discharged from Hospital. China Digital Medicine 2019; 14(7): 033.

40.

Li Meng-qian

LD-H

. Application of Network Education Combined with Sports Bracelet in ADL and Functional Training of Stroke Patients with Hemiplegia. Today Nurse 2020; 27(23): 042.

41.

Ding

Dai

, et al. Clinical Application Research on Stroke Situational Intelligent Rehabilitation Training System Based on Wearable Devices: A Randomized Controlled Trial. Healthcare 2025; 13(7): 708. https://doi.org/10.3390/healthcare13070708

42.

Yang

Y-S

Tseng

C-H

Fang

W-C

, et al. Effectiveness of a New 3D-Printed Dynamic Hand–Wrist Splint on Hand Motor Function and Spasticity in Chronic Stroke Patients. Journal of Clinical Medicine 2021; 10(19): 4549. https://doi.org/10.3390/jcm10194549

43.

Villafañe

Taveggia

Galeri

, et al. Efficacy of Short-Term Robot-Assisted Rehabilitation in Patients With Hand Paralysis After Stroke: A Randomized Clinical Trial. Hand 2017; 13(1): 95–102. https://doi.org/10.1177/1558944717692096

44.

Alon

Levitt

McCarthy

. Functional Electrical Stimulation (FES) May Modify the Poor Prognosis of Stroke Survivors with Severe Motor Loss of the Upper Extremity. American Journal of Physical Medicine & Rehabilitation 2008; 87(8): 627–636. https://doi.org/10.1097/PHM.0b013e31817fabc1

45.

Chang

Kim

T-W

Kim

, et al. Interim results of exoskeletal wearable robot for gait recovery in subacute stroke patients. Scientific Reports 2025; 15(1): 11671. https://doi.org/10.1038/s41598-025-96084-6

46.

Wang

Liu

Wang

, et al. Internet + wearable device training effects on limb function recovery and serum neurocytokine content in stroke patients. NeuroRehabilitation 2024; 55(1): 17–28. https://doi.org/10.3233/NRE-230347

47.

Wei

WXJ

Fong

KNK

Chung

RCK

, et al. “Remind-to-Move” for Promoting Upper Extremity Recovery Using Wearable Devices in Subacute Stroke: A Multi-Center Randomized Controlled Study. IEEE Transactions on Neural Systems and Rehabilitation Engineering 2019; 27(1): 51–59. https://doi.org/10.1109/TNSRE.2018.2882235

48.

Kim

Baek

Suram

, et al. Effect of Wearable Robot-Assisted Gait Training on Balance and Walking Ability in Subacute Stroke Patients. American Journal of Physical Medicine & Rehabilitation 2025; 104(10): 859–868. https://doi.org/10.1097/PHM.0000000000002735

49.

Barry

Ross

Woehrle

. Therapy Incorporating a Dynamic Wrist-Hand Orthosis Versus Manual Assistance in Chronic Stroke. Journal of Neurologic Physical Therapy 2012; 36(1): 17–24. https://doi.org/10.1097/NPT.0b013e318246203e

50.

Watanabe

Tanaka

Inuta

, et al. Locomotion Improvement Using a Hybrid Assistive Limb in Recovery Phase Stroke Patients: A Randomized Controlled Pilot Study. Archives of Physical Medicine and Rehabilitation 2014; 95(11): 2006–2012. https://doi.org/10.1016/j.apmr.2014.07.002

51.

Steven

Wolf

CJW

Philip Miller

, et al. Effect of constraint-induced movement therapy on upper extremity function 3 to 9 months after stroke: the EXCITE randomized clinical trial. JAMA 2006; 296(17): 2095.

52.

Han

Zhang

Kang

, et al. Clinical Evidence of Exercise Benefits for Stroke. Exercise for Cardiovascular Disease Prevention and Treatment. Advances in Experimental Medicine and Biology 2017; 1000: 131–151.

53.

Stewart

Cramer

. Patient-Reported Measures Provide Unique Insights Into Motor Function After Stroke. Stroke 2013; 44(4): 1111–1116. https://doi.org/10.1161/STROKEAHA.111.674671

54.

Wan

Cheng

Luo

, et al. Neuroprotection, learning and memory improvement of a standardized extract from Renshen Shouwu against neuronal injury and vascular dementia in rats with brain ischemia. Journal of Ethnopharmacology 2015; 165: 118–126. https://doi.org/10.1016/j.jep.2015.02.027

55.

Simpson

Menon

Hodgson

, et al. Clinicians’ perceptions of a potential wearable device for capturing upper limb activity post-stroke: a qualitative focus group study. J Neuroeng Rehabil 2021; 18(1): 135. https://doi.org/10.1186/s12984-021-00927-y

56.

Asada

Shaltis

Reisner

, et al. Mobile monitoring with wearable photoplethysmographic biosensors. IEEE Eng Med Biol Mag 2003; 22(3): 28–40. https://doi.org/10.1109/memb.2003.1213624

57.

Radder

Prange-Lasonder

Kottink

, et al. A wearable soft-robotic glove enables hand support in ADL and rehabilitation: A feasibility study on the assistive functionality. J Rehabil Assist Technol Eng 2016; 3: 2055668316670553. https://doi.org/10.1177/2055668316670553

58.

Arya

Verma

Garg

. Estimating the minimal clinically important difference of an upper extremity recovery measure in subacute stroke patients. Top Stroke Rehabil 2011; 18(Suppl 1): 599–610. https://doi.org/10.1310/tsr18s01-599

59.

Huynh

DiCarlo

Vora

, et al. Sensitivity to Change and Responsiveness of the Upper Extremity Fugl-Meyer Assessment in Individuals With Moderate to Severe Acute Stroke. Neurorehabil Neural Repair 2023; 37(8): 545–553. https://doi.org/10.1177/15459683231186985

60.

Chisholm

Alamro

Williams

, et al. Overground vs. treadmill-based robotic gait training to improve seated balance in people with motor-complete spinal cord injury: a case report. J Neuroeng Rehabil 2017; 14(1): 27. https://doi.org/10.1186/s12984-017-0236-z

61.

Park

Cho

Hong

, et al. Effects and Safety of Wearable Exoskeleton for Robot-Assisted Gait Training: A Retrospective Preliminary Study. J Pers Med 2023; 13(4): 676. https://doi.org/10.3390/jpm13040676

62.

Calafiore

Negrini

Tottoli

, et al. Efficacy of robotic exoskeleton for gait rehabilitation in patients with subacute stroke: a systematic review. Eur J Phys Rehabil Med 2022; 58(1): 1–8. https://doi.org/10.23736/S1973-9087.21.06846-5

63.

Darcy

Rashford

Huizenga

, et al. Balance Improvement and Fall Risk Reduction in Stroke Survivors After Treatment With a Wearable Home-Use Gait Device: Single-Arm Longitudinal Study With 1-Year Follow-Up. JMIR Form Res 2025; 9: e67297. https://doi.org/10.2196/67297

64.

Wang

Luo

, et al. Can wearable real-time biofeedback gait training devices improve gait speed, balance, functional mobility and activities of daily living (ADL) in individuals post-stroke? A systematic review and meta-analysis of randomized controlled trials. J Neuroeng Rehabil 2025; 23(1): 47. https://doi.org/10.1186/s12984-025-01863-x

65.

Tamura

Miyata

Kobayashi

, et al. The minimal clinically important difference in Berg Balance Scale scores among patients with early subacute stroke: a multicenter, retrospective, observational study. Top Stroke Rehabil 2022; 29(6): 423–429. https://doi.org/10.1080/10749357.2021.1943800

66.

Huang

Shi

Sun

, et al. Minimum Clinically Important Difference of the Modified Barthel Index in the Assessment of Ischemic Stroke Patients. Chinese Journal of Health Statistics 2022; 39(02): 012.

67.

Hsieh

Wang

, et al. Establishing the minimal clinically important difference of the Barthel Index in stroke patients. Neurorehabil Neural Repair 2007; 21(3): 233–238. https://doi.org/10.1177/1545968306294729

68.

Chang

Kim

, et al. Efficacy of Wearable Exoskeleton for Gait Recovery in Patients With Stroke: A Multicenter Randomized Controlled Trial. Stroke 2026; 57(3): 577–586.

69.

Lynch

Jones

Simpson

, et al. Activity monitors for increasing physical activity in adult stroke survivors. Cochrane Database Syst Rev 2018; 7(7): Cd012543. https://doi.org/10.1161/STROKEAHA.118.023088

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.06 MB

0.00 MB

1.23 MB