Impact of visuospatial neglect post-stroke on daily activities, participation and informal caregiver burden: a systematic review

Objectives: Visuospatial neglect (VSN) is a common cognitive disorder after stroke. The primary aim of this systematic review was to provide an overview of the impact of VSN in 3 aspects: (1) activities of daily living (ADL), (2) participation, and (3) caregiver burden. The second aim was to investigate the differences in studies focusing on populations with mean age < 65 versus (cid:2) 65 years. Methods: PubMed, EMBASE, Web of Science, Cochrane Library, Emcare, PsychINFO, Academic Search Premier and CENTRAL were searched systematically. Quality was assessed with the Mixed Methods Appraisal Tool. Results: Of the 115 included studies, 104 provided outcomes on ADL, 15 on participation (4 studies with mean age (cid:2) 65), and 2 on caregiver burden (1 study with mean age (cid:2) 65). Quality assessment yielded scores ranging from 0 to 100%. VSN had a negative impact on ADL (i.e., independence during ADL and performance in self-care, household tasks, reading, writing, walking, wheelchair navigation) and participation (i.e., driving, community mobility, orientation, work). The impact of VSN on fulﬁlling social roles was unclear. VSN had a negative effect on caregiver burden. We found no clear age-related differences. Conclusions and implications: VSN has a negative impact not only on patients’ independence but particularly on the performance of ADL. Despite the far fewer studies of VSN as compared with ADL, VSN also seems to hamper participation and increase caregiver burden, but further research is needed. Because of the large impact, VSN should be systematically and carefully assessed during rehabilitation. A considerable number of different instruments were used to diagnose VSN. Diagnosing VSN at more than one level [function (i.e., pen-and-paper test), activities, and participation] is strongly recommended. Consensus is needed on how to assess VSN and its negative impact for research and rehabilitation practice. Systematic review Registration No.


Introduction
Visuospatial neglect (VSN) is a common cognitive disorder after stroke.Patients with VSN have problems reporting and responding or orienting to visual stimuli in the contralesional hemispace that cannot be attributed to sensory or motor impairments [1].Estimates of the incidence of VSN after stroke differ, ranging from 20% to 82%, largely dependent on the patient sample, time post-stroke onset, and number and types of tests used for assessment [2].VSN can easily become a chronic disorder; half of the stroke patients with VSN still have VSN at 1 year after stroke onset [3].
The impact of VSN can be far-reaching for patient's everyday life.Stroke patients with VSN show slower recovery and less recovery in activities of daily living (ADL) as compared with non-VSN patients [4][5][6][7][8] and remain more dependent on their environment during (basic) ADL [4].Considering this negative impact, VSN may also have severe consequences for informal caregivers (hereafter called caregivers) of VSN patients.
Previous reviews have mainly focused on the consequences of VSN for body function and body structure [9,10] or on the impact of VSN mainly measured by independence during ADL [5,8].The aim of this review was to provide a systematic overview of the impact of VSN on (1) ADL, (2) participation, and (3) informal caregiver burden.In contrast to previous studies, the present review focuses on the quality of ADL and participation.We include all types of study designs, not just randomized controlled trials (RCTs) or large cohort studies, to provide a wider overview of the current situation.As well, although stroke is common and VSN occurs with increasing frequency and increasing severity in the geriatric population [11], many studies address younger populations.Therefore, to gain more insight into potential differences by age, the second aim was to provide a descriptive overview of differences between studies (i.e., amount of data available, outcomes examined) including younger versus older patients (mean age < 65 vs. !65 years).

Protocol and registration
The protocol for this systematic review was pre-registered in the PROSPERPO Database (registration no.CRD42018087483).This review followed the standard guidelines of Preferred Reporting items for Systematic Reviews and Meta-Analyses (PRISMA) [12].

Search strategy and article selection
We conducted a systematic literature search on May 1, 2018 in the electronic databases PubMed, EMBASE, Web of Science, Cochrane Library, Emcare, PsychINFO, Academic Search Premier and CENTRAL by using the terms visuospatial neglect, stroke, activity, participation and caregiver burden (Appendix).We had no restrictions on publication date or research design.After removing duplicates, we excluded study reports that were: not in English; not related to humans; not original research (e.g., an editorial or review); intervention studies; focusing on other levels than activity or participation; not related to stroke; not related to VSN; not describing a relation between VSN and outcome measure (non-relation); not available in full text.
In case of ambiguity in article selection with respect to the exclusion criteria related to activity and participation, activity and participation level were defined according to the International Classification of functioning, Disability and Health Framework (ICF) of the World Health Organization [13].Activity was considered the execution of a task or action by an individual [14].Only ADL within this framework were included (e.g., self-care and household task).Reading was included only when it involved more than single words.Reaching was included only if it was goaldirected grabbing or holding; otherwise it was considered a ''function'' according to the ICF.Virtual reality (VR) was included when it was used to measure an activity in daily life (e.g., walking) in a protected environment.According to the ICF, participation is defined as involvement in a life situation (e.g., driving) [14].
The first author (MB) screened titles and abstracts; in case of uncertainty, the second author (TN) was consulted to reach consensus.Among the included abstracts, full-text articles (when available) were screened.Two reviewers (MB and TN) independently assessed the full-text articles.The reviewers' selection and arguments were compared, and any differences regarding the inclusion of an article were discussed to reach consensus.

Data extraction and analyses
After the final selection, data were extracted by the first author (MB).In case of uncertainty, the second author (TN) was consulted.The following study characteristics were extracted: authors, year, study design, number of patients, proportion of VSN, mean age, mean time post-stroke onset, setting, country, diagnostic instrument used to assess VSN, outcome measures in ADL, participation or caregiver burden, main findings in (sub)categories of ADL and participation (e.g., self-care and household tasks by ADL, or driving by participation), and total methodological quality score.The primary outcome was the impact of VSN on ADL, participation and caregiver burden.The secondary outcome was the overview related to age (mean population age < 65 vs. !65 years).The widespread differences in study characteristics precluded the possibility to directly compare results; therefore, a meta-analysis was not possible.A narrative synthesis of the findings is presented.

Quality assessment
The Mixed Methods Appraisal Tool (MMAT v2011) was used to assess the methodological quality of studies [15].The MMAT was designed for appraising complex systematic literature reviews that include domains of qualitative, quantitative and mixed-methods studies.Scores on the MMAT range from 0% (no criteria met) to 100% (all criteria met).Each study included in this review was defined by a research design of the MMAT and assessed within its methodological domain [15].The included studies were assessed by the first 2 authors separately (MB and TN), then discussed in detail.Any disagreement was discussed until consensus was reached.If no consensus was reached, advice was sought from the third author (MC).

Study selection
The search yielded 4,669 citations.After removing duplicates, we screened 1,931 articles on the basis of the title and abstract and excluded 1,657.The 274 remaining studies were analyzed in full text.Finally, reports for 115 studies met the inclusion criteria and were included in this review (Fig. 1).

Study characteristics
Study characteristics are presented in Tables 1-3 based on ADL, participation, and caregiver burden.The studies had considerable diversity regarding study design; number of included patients; proportion of VSN; mean age of participants; mean time post-stroke onset; setting; diagnostic instrument used to assess the presence of VSN; outcome measure in ADL, participation or caregiver burden; categories of ADL; and participation in various activities.
Most studies were performed in Europe (n = 51); 10 were from the United States/Canada, 9 from Asia, 5 from Oceania, and 1 study covered both Europe and North America.For 39 reports, the study location of inclusion and assessment were not specified.Most studies were conducted in a rehabilitation setting (n = 50) and 10 were community-based.In another 35 studies, patients were selected from a hospital setting.Not all reports differentiated between in/outpatients.In 4 studies, patients were recruited from a Center for Evaluation of Fitness to-Drive and Car Adaptations [16][17][18][19].The settings of 16 studies were unclear or not specified.
All included studies used measurement tools to first assess VSN within the study population.These tools were used independent of the outcome measures.Tables 1-3, column 8 (VSN tests), shows that many different tests were used to diagnose VSN.As a diagnostic measurement for VSN, a pen-and-paper test was used in 98 studies.A cancellation task (n = 92) was mostly used, either in isolation or in a test battery (e.g., the Behavioral Inattention Test [BIT]).The line bisection task was used 57 times, also in isolation or in a test battery.For assessing VSN during ADL, tests used were the Catherine Bergego Scale (CBS) (n = 11), Kessler Foundation Neglect Assessment Process (n = 4) and Dublin Extrapersonal Neglect Assessment (n = 1).Most studies used more than one (diagnostic) assessment to identify VSN, independent of the outcome measure.In 13 studies, a single test was used (test battery such as BIT excluded).The (subtest of the) BIT was used in 36 studies.
Mean age of the population ranged from 50 to 84 years.We distinguished between study populations with mean age !65 years (n = 51 studies) and < 65 years (n = 55 studies).Nine reports did not include the mean age or the age could not be calculated from the data available.Overall, 104 reports provided data on ADL, 15 on participation level, and 2 on caregiver burden.These outcomes could be studied in isolation or combined.

Quality assessment
An overview of the quality assessment (based on MMAT total scores) is presented in Tables 1-3 (last column).We found a large variation in research designs and methodological quality.Total MMAT scores for the 115 included studies ranged from 0% (n = 2) to 100% (n = 34).
Most studies had a ''Quantitative non-randomized design'' (n = 88).Within this category, 45 cohort studies, 22 case-control studies, and 21 cross-sectional studies were specified.Their quality ranged from 0 to 100%.Of the 88 studies, 63 scored !75%, with 26 scoring 100%.In this category, the quality of the studies scoring < 100% was limited in terms of recruitment of participants, appropriate measurements, comparable groups, and complete outcome data.
In the MMAT category ''Quantitative descriptive studies'', 15 studies were specified: 9 incidence or prevalence studies, 5 case series, and 1 case report.The quality ranged from 25% to 100%.Of the 15 studies, 12 scored !75%, with 6 scoring 100%.In this category, the quality of the studies scoring < 100% was limited in terms of sampling strategy, representative population, appropriate measurement and response rate.Eight studies met the criteria of the MMAT category ''Qualitative design'': 3 had a phenomenology design and 5 were case studies.In this category, 7 studies scored 75%.Only 1 study [20] obtained a score of 100%.The quality of the studies scoring < 100% was limited in terms of sources of qualitative data (n = 2) and how the findings were related to researchers' influence (n = 5).
Finally, 4 studies met the criteria of the MMAT category ''Mixed method design'': 2 had a sequential explanatory design and 2 a triangulation design.The quality ranged from 25% to 100%.

Results of individual studies
We divided ADL into independency in ADL, self-care and household tasks, reading/writing and walking/wheelchair navigation.Participation was divided into the subcategories driving, community mobility/orientation, social roles and work.
One study found no significant associations between successfully stepping over an obstacle and presence of VSN [110].VSN had a negative impact on exercise such as walking and jogging [111,112].VSN severity was related to bumping into objects with a wheelchair [113].
Four studies discussed the negative impact of VSN during driving.Poor driving performance was visible in lane changing, understanding, and traffic participation [17,18].Even among patients showing no VSN on conventional pen-and-paper tests, their driving performance was hampered [19].
Six studies provided data on community mobility and orientation.VSN had a negative impact on regaining functional mobility in the community [25,121].Patients with VSN had difficulties orienting themselves in the environment and identifying where they were located [20,69,70,74].
Four studies examined the impact of VSN on fulfilling social roles: 3 used the LIFE-H (questionnaire used to estimate the level of participation).VSN was correlated with categorical independent variables but was not the best predictor at 6 months post-stroke [122].When returning to the community after a stroke, positive changes in participation over time were possible, even with cognitive deficits [123,124]; this study showed that patients with and without VSN have the same level of participation and fulfilling social roles.The fourth study indicated that the change in relationship to other people was the main source of concern for most VSN patients [20].
Only 2 studies described the impact of VSN on returning to work.One study reported that 11% of the patients with neglect or aphasia had resumed working at 1 or 3 years after stroke [68].The other study reported that a patient could sing and play the piano again but was unable to write a new musical [125].

Impact of VSN on caregiver burden
Only 2 of the 115 studies discussed caregiver burden.[126,127] The study of Buxbaum et al. (2004) had a quantitative descriptive design and a quality score of 100%.The study of Chen et al. (2017) had a mixed-methods design and scored 75% for quality (Table 3).
VSN independently contributed to predicting family burden [126].Caregivers of VSN patients were more likely to describe economic stressors and undesirable changes in career and vacation planning than caregivers of non-VSN patients [126].VSN in stroke survivors was associated with greater burden and stress in caregivers and involved allocating more caregiving as compared with caregivers of non-VSN patients [127].

Differences between study populations with mean age younger/ older than 65 years
Here, we provide according to the second aim, a descriptive overview of differences between studies focusing on populations with mean age < 65 versus !65 years (Tables 1-3 Of the 2 studies that provided findings on the impact of VSN on caregiver burden (Table 3, column 5), for 1, the population mean age was !65 years.From these preliminary findings, there is no indication that these outcomes would differ between younger and older populations.

Discussion
The results of this review of 115 studies (quality ranging from 0% to 100%) indicate that VSN had a negative impact on patients' independence and in particular the performance of ADL (self-care, household tasks, reading, writing, walking, wheelchair navigation).Additionally, VSN had a negative impact on participation (driving, community mobility, orientation, work).The impact of VSN on fulfilling social roles remains unclear.VSN had a negative effect on caregiver burden.We found no clear age-related differences.Few studies focused on participation (especially among older patients) and caregiver burden.

Impact of VSN on ADL
Our findings on independence and performance of ADL are consistent with the review of Jehkonen et al., who concluded that VSN had a negative influence on independence of activities [8].In contrast, Stein et al. reported no important relationship between hemi-inattention status and functional outcome in patients with right hemisphere stroke because of significant methodological limitations of the relatively few studies published in this field in the last decade [5].
Measuring independence in ADL is useful (e.g., FIM of BI) to produce an overall image regarding functioning, functional recovery and progress.However, an independence measure does not allow for analyzing the problems people experience or to indicate the problems in performance.When VSN is examined with a performance scale (e.g., CBS) better insight in the actual performance of activities is gained as well as an overview of real experienced needs in treatment of VSN or in compensation strategies.With more insight in ADL performance, personal goals can be better described, and treatment can be more adequate, coordinated and adjusted accordingly.A performance measure will help give more information to the patient and caregiver about the treatment, the needs, modifications and preparation both during rehabilitation and when returning home.We suggest that when assessing VSN, a measurement tool that gives insight in the impact of VSN on performance of activities is needed as well.
Many studies reported a discrepancy between the various VSN diagnostic measurements, particularly when using tests on different levels of outcomes (levels of body function, ADL and participation).For example, pen-and-paper test results differed from results of a measurement for ADL [69,86,108], VR [119], and dynamic measurement [106].In general, many studies reported that patients who showed VSN during an ADL task did not show VSN on a regular pen-and-paper task.To prevent missing VSN and underrecognizing it, VSN should be assessed at more than one level (i.e., levels of isolated [cognitive] function as well in more dynamic situations).
Overall, a considerable number of different instruments were used to diagnose VSN, which makes comparison and interpretation of the studies difficult and does not benefit research and practice.Therefore, future VSN studies will require consensus on outcome measures for both research and general practice.This requirement applies to all levels of measurements to diagnose VSN in terms of function (i.e., pen-and-paper test) and at the level of activities and participation.

Impact of VSN on participation
To our knowledge, no other review has examined the impact of VSN on participation.Our findings show that VSN had a negative influence on different aspects of participation and also show considerable variety regarding which aspect of participation showed related problems.The impact of VSN in fulfilling social roles was unclear.Further research is needed to gain more insight into the association between VSN and participation.

Impact of VSN on caregiver burden
Although we expected to find more studies on the impact of VSN on caregiver burden, only 2 met the inclusion criteria, both associated VSN with a greater burden on caregivers of VSN patients than caregivers of non-VSN patients.The effect of VSN on caregiver burden is underexposed, and more research is needed to explore caregivers' problems and to better guide VSN patients and caregivers in preparation for a return home.

Difference between study populations with mean age younger/ older than 65 years
We found no clear age-related differences in the aspects investigated.The results concerning age differences by level of participation and caregiver burden should be interpreted with caution.Our results indicate that more research in this field is needed.
Although VR is a relatively new instrument to detect problems in VSN activities in a safe environment, we found 8 studies using this instrument.However, 3 were from the same authors [114,115,120], 2 of which included the same population [115,120].Since only one study had a population with mean age !65 years, [83] more studies are needed on the usability and diagnostic criteria in this older age group.
VSN is a common disorder after stroke in older adults (mean age !65 years), but we found few studies of this population according to level of participation.Further research is needed to gain more insight into the association between VSN and participation, especially among older patients.

Strengths and limitations
A limitation of this study was the restriction in search strategy.First, our search strategy concerned only studies written in English.
Second, because of the large amount of included studies, we did not use a snowball method or include grey literature (e.g., theses).
One strength of this systematic review is the inclusion of both quantitative and qualitative study designs.By using the MMAT for methodological quality assessment, we could assess various types of study designs with one measurement tool within their own category.In addition, using the MMAT could also be considered a limitation.The MMAT can be used for every study design but is therefore also a coarse tool to measure quality.The quality score could differ when using another instrument for quality assessment.
Another strength is that we examined the impact of VSN on three different aspects (ADL, participation and caregiver burden); moreover, the outcome ADL was split into independence and performance.Although many studies have examined independence measures, these instruments do not assess actual performance.Measurements of performance are important for actual insight into the problems people with VSN really experience.More insight can ensure better individual treatment (goals) and better guidance for patients and caregivers for returning home.In the present review, both the inclusion of all study designs and examination of the impact of VSN provided a more comprehensive overview of the impact of VSN on patients and caregivers.

Conclusion
VSN has a negative impact not only on patients' independence but particularly on the performance of ADL.Despite the far fewer studies of VSN as compared with ADL, VSN also seems to hamper participation and increases caregiver burden, but further research is recommended.Because of the large impact, VSN should be systematically and carefully assessed during rehabilitation.A considerable number of different instruments were used to diagnose VSN.Diagnosing VSN at more than one level (function [i.e., pen-and-paper test], activities and participation) is strongly recommended.Consensus is needed for research and practice on how to assess VSN and its negative impact.

Table 1
Overview of studies: activities of daily living.

Table 1 (
Continued ) ADL: activities of daily living; BCoS: Birmingham Cognitive Screen; BI: Barthel Index; BIT: Behavioral Inattention Test; BIT B: Behavioral Inattention Test Behavioral subtest; BIT C: Behavioral Inattention Test Conventional subtest; BL: bilateral; BNIS: Barrow Neurological Institute Screen; Baking tray task; CBIT-HK: Chinese Behavioral Inattention Test-Hong Kong; CBS: Catherine Bergego Scale; CNS: Canadian Neurological scale; CR: continuous recovery group; DENA: Dublin Extra personal Neglect Assessment; EADL: Extended Activities of Daily Living Scale; EMS: Elderly Mobility Scale; FAC: Functional Ambulation Categories; FAI: Frenchay Activities Index; FIM: Functional Independence Measure; FR: fluctuating recovery group; Instrumental Activity Measure; IADL-CV: Chinese version of the Lawton Instrumental Activities of Daily Living scale; IAM: Instrumental Activity Measure; IVA-CPT: Integrated Visual Auditory Continuous Performance Test; KF-NAP: Kessler Foundation Neglect Assessment Process; LH: left hemisphere; LHS: London Handicap Scale; LIMOS: Lucerne international classification of function: disability and health-based Multidisciplinary Observation Scale; L-OMIT: omitted stimuli in left hemispace group; LOTCA: Loewenstein Occupational Therapy Cognitive Assessment; LSQ: Life Space Questionnaire; MAC: Mobility Assessment Course: MLI: Mild Left Inattention; MMAT: Mixed Methods Appraisal tool; MRS: modified Rankin Scale; MVPT: Motor-Free Visual Perception Test; MWCT: Mesulam and Weintraub random symbol Cancellation Task; N: neglect; NAT: Naturalistic action test; pn: personal neglect; PR: poor recovery group; R-BIAS: began tasks to the fight without omissions group; r-BIT: reduced version of the BIT-test; RH: right hemisphere; RIF-CAS: Rehabilitation Institute of Chicago Functional Assessment Scale; RKE-R: Rabideau Kitchen Evaluation; RMA: Rivermead Motor Assessment; RMI: Rivermead Mobility Index; RPAB: Rivermead Perceptual Assessment Battery; RWN: Moss Real World Navigation; TCT: Trunk Control Test; Unknown: unilateral neglect; USER: Utrechtse Scale for Evaluation of clinical Rehabilitation; USN: Unilateral Spatial Neglect; VISSTA: VIsual Spatial Search Task; VNRI: Visual Neglect Recovery Index; VR: Virtual Reality; VRLAT: Virtual Reality Lateralized Attention Test; VSN: VisuoSpatial Neglect; VWNT: VR Wheel-chair Navigation Test.

Table 2
Overview of studies: participation.
ADL: activities of daily living; BI: Barthel Index; BIT: Behavioral Inattention Test; BIT C: Behavioral Inattention Test Conventional subtest; BL: bilateral; BNIS: Barrow Neurological Institute Screen; CARA: department of the Road Safety Institute; CBS: Catherine Bergego Scale; FIM: Functional Independence Measure; IAM: Instrumental Activity Measure; KF-NAP: Kessler Foundation Neglect Assessment Process; LH: left hemisphere; LIFE-H, life-habits assessment; LSA: Life space assessment; LSQ: Life Space Questionnaire; MMAT: Mixed Methods Appraisal tool; PP: Peripheral Perception test; r-BIT: reduced version of the BIT-test; RH: right hemisphere; ROCF: Rey-Osterrieth Complex Figure; SDSA: Stroke Driver Screening Assessment; TRIP: Test Ride for Investigating Practical fitness to drive; VSN: visuospatial neglect; UN: unilateral neglect.

Table 3
Overview of studies: caregiver burden.
, column 5).Of the 104 studies that reported on ADL (Table1, column 5), 48 studies had a population with mean age < 65 years versus 47 with mean age !65 years.Mean age was not reported for 9 studies.Only 1 of 8 studies on VR had a mean age !65 years.Because overall the main findings did not differ, no difference is expected regarding the impact of VSN on independency and performance during ADL between younger and older persons with VSN.Of the 15 studies reporting on the impact of VSN on participation (Table2, column 5), 9 studies had a population with mean age < 65 years versus 4 with mean age !65 years.In 2 studies, the mean age was not specified.The studies with mean age of the population !65 years provided data on work (n = 1), community mobility and orientation (n = 1), and fulfilling social roles (n = 2, but from the same authors/population).No studies about driving had a population with mean age !65 years.Studies on fulfilling roles with population mean age !65 years showed no relation between VSN and fulfilling social roles, in contrast to studies with population mean age < 65 years (n = 2).