Neuropsychological rehabilitation or cognitive rehabilitation aims to assist patients’ in recovering or compensating for impaired cognitive functions and in emphasizing meaningful functional activity in everyday life. The holistic approach of neuropsychological rehabilitation emphasizes the interlinked nature of cognition, emotion, and psychosocial functioning (Wilson, Reference Wilson2008). Comprehensive-holistic cognitive interventions are designed to decrease the impact of cognitive deficits and increase the patient’s self-awareness of the disability, while supporting their emotional, interpersonal, and psychological functioning through the therapeutic environment (Cicerone et al., Reference Cicerone, Goldin, Ganci, Rosenbaum, Wethe, Langenbahn, Malec, Bergquist, Kingsley, Nagele, Trexler, Fraas, Bogdanova and Harley2019; Koehler et al., Reference Koehler, Wilhelm and Shoulson2012). In addition, neuropsychological rehabilitation can focus on improving or maintaining the patient’s cognitive abilities in specific cognitive areas (e.g., attention, vision and neglect, language and communication skills, memory, executive functioning) (Cicerone et al., Reference Cicerone, Goldin, Ganci, Rosenbaum, Wethe, Langenbahn, Malec, Bergquist, Kingsley, Nagele, Trexler, Fraas, Bogdanova and Harley2019). Interventions may include compensatory strategy-based techniques and/or individually adapted cognitive training (Miotto et al., Reference Miotto, Serrao, Guerra, Lúcia and Scaff2008; Mowszowski et al., Reference Mowszowski, Batchelor and Naismith2010; Wilson et al., Reference Wilson, Winegardner, van Heugten and Ownsworth2017). Furthermore, cognitive interventions may include family-centered or social interventions (De Goumoëns et al., Reference De Goumoëns, Rio, Jaques and Ramelet2018; Shaw, Reference Shaw2016; Robinson et al., Reference Robinson, Kaizar, Catroppa, Godfrey and Yeates2014).
Cognitive interventions can be utilized with a wide array of patients with neuropsychological impairments, and research has demonstrated its effectiveness in improving cognitive and psychosocial functioning in populations, such as acquired brain injury (ABI), stroke, multiple sclerosis (MS; Cicerone et al., Reference Cicerone, Langenbahn, Braden, Malec, Kalmar, Fraas, Felicetti, Laatsch, Preston, Bergquist, Azulay, Cantor and Ashman2011; Cicerone et al., Reference Cicerone, Goldin, Ganci, Rosenbaum, Wethe, Langenbahn, Malec, Bergquist, Kingsley, Nagele, Trexler, Fraas, Bogdanova and Harley2019; Turner-Stokes et al., Reference Turner-Stokes, Pick, Nair, Disler and Wade2015; Resch et al., Reference Resch, Rosema, Hurks, de Kloet and van Heugten2018; Rosti‐Otajärvi & Hämäläinen, Reference Rosti‐Otajärvi and Hämäläinen2014; Mattioli et al., Reference Mattioli, Stampatori, Bellomi, Capra, Rocca and Filippi2010). Furthermore, systematic reviews have demonstrated some promising findings from technologically-oriented cognitive interventions, such as virtual reality, video games, and computer-based training with persons with various neurological disorders, including autism spectrum disorder, attention-deficit/hyperactivity disorder (ADHD), learning disabilities, Parkinson’s disease, dementia, and ABI; however, the results are mixed, particularly for the drill-based cognitive training approaches with issues of generalizability (e.g., there was no clear transfer effect to patient’s daily life, or it was not investigated) (Mesa-Gresa et al., Reference Mesa-Gresa, Gil-Gómez, Lozano-Quilis and Gil-Gómez2018; Peijnenborgh et al., Reference Peijnenborgh, Hurks, Aldenkamp, Vles and Hendriksen2015; García-Casal et al., Reference García-Casal, Loizeau, Csipke, Franco-Martín, Perea-Bartolomé and Orrell2017; Nousia et al., Reference Nousia, Martzoukou, Tsouris, Siokas, Aloizou, Liampas, Nasios and Dardiotis2020; Resch et al., Reference Resch, Rosema, Hurks, de Kloet and van Heugten2018; Rivero et al., Reference Rivero, Nunez, Pires and Bueno2015; Sala et al., Reference Sala, Aksayli, Tatlidil, Tatsumi, Gondo and Gobet2019; Bogdanova et al., Reference Bogdanova, Yee, Ho and Cicerone2016).
In recent years, telerehabilitation has provided an alternative for patients with neuropsychological impairments to receive rehabilitation remotely at home or in other environments outside the clinic or hospital. Telerehabilitation has the potential to provide flexibility of time and location in rehabilitation (Chen et al., Reference Chen, Jin, Zhang, Xu, Liu and Ren2015). Remote services can increase access for persons who live in areas where traditional rehabilitation services may be limited (Peretti et al., Reference Peretti, Amenta, Tayebati, Nittari and Mahdi2017), and increase accessibility of services for patients who have mobility impairments or difficulties with traveling (Brennan et al., Reference Brennan, Mawson, Brownsell, Gaggioli, Keshner, Weiss and Riva2009). Telerehabilitation also has potential to be less labor intensive and serve as a lower cost alternative to traditional interventions (Kairy et al., Reference Kairy, Lehoux, Vincent and Visintin2009; Kueider et al., Reference Kueider, Parisi, Gross and Rebok2012), although the results considering cost-effectiveness are still inconclusive (Laver et al., Reference Laver, Adey-Wakeling, Crotty, Lannin, George and Sherrington2020).
Clinician-patient communication in telerehabilitation may be delivered remotely via multiple technologies, such as telephone, videoconference, or e-mail (Peretti et al., Reference Peretti, Amenta, Tayebati, Nittari and Mahdi2017; Stephenson et al., Reference Stephenson, Howes, Murphy, Deutsch, Stokes, Pedlow and McDonough2022). Remote interventions may be delivered synchronously with real-time clinician-patient interaction, asynchronously by using remotely monitored platforms with delayed feedback from a clinician, or using a hybrid approach combining both (Khoshrounejad et al., Reference Khoshrounejad, Hamednia, Mehrjerd, Pichaghsaz, Jamalirad, Sargolzaei, Jamalirad and Aalaei2021; Stephenson et al., Reference Stephenson, Howes, Murphy, Deutsch, Stokes, Pedlow and McDonough2022). Interventions can also include virtual reality (Rogante et al., Reference Rogante, Grigioni, Cordella and Giacomozzi2010), digital patient portals (Voigt et al., Reference Voigt, Benedict, Susky, Scheplitz, Frankowitz, Kern, Müller, Schlieter and Ziemssen2020), or other technologies such as sensory monitors, personal digital assistants, and smart phone applications (McCue et al., Reference McCue, Fairman and Pramuka2010). Using a computerized training program can offer a personalized approach to training, such as real-time feedback and automatic adjustments to the user’s ability level (Kueider et al., Reference Kueider, Parisi, Gross and Rebok2012). In this study the term “teleneuropsychological rehabilitation” refers to neuropsychological or cognitive interventions utilizing remote delivery methods involving synchronous or asynchronous interaction between the professional and patient with the aim to assist the recovery or compensation for impaired neuropsychological functioning (McCue & Cullum, Reference McCue, Cullum, Noggle and Dean2013; McLellan, Reference McLellan, Swash and Oxbury1991; Wilson, Reference Wilson2008) which distinguishes it from completely independently utilized (i.e., without clinician involvement) cognitive training programs or assistive technology (e.g., personal digital assistants).
Although there is increasing evidence demonstrating the potential of telerehabilitation in patients with neuropsychological impairments (e.g., Camden et al., Reference Camden, Pratte, Fallon, Couture, Berbari and Tousignant2020; Laver et al., Reference Laver, Adey-Wakeling, Crotty, Lannin, George and Sherrington2020; Meltzer et al., Reference Meltzer, Baird, Steele and Harvey2018), the literature of the systematic review of the teleneuropsychological rehabilitation/remote cognitive intervention is sparse. Findings from one systematic review demonstrated that remote interventions could be as effective as conventional rehabilitation to improve cognitive or depressive symptoms after traumatic brain injury (TBI; Betts et al., Reference Betts, Feichter, Kleinig, O'Connell-Debais, Thai, Wong and Kumar2018). One meta-analysis focused on home-based technology interventions in rehabilitation of executive functioning and memory for children and adolescents with ABI (Linden et al., Reference Linden, Hawley, Blackwood, Evans, Anderson and O'Rourke2016). Results were encouraging, but the quality of evidence for all outcomes was noted to be low. Similarly, a meta-analysis of computer-based remote training programs with behavioral and cognitive focus demonstrated promising results for patients with pediatric ABI (Corti et al., Reference Corti, Oldrati, Oprandi, Ferrari, Poggi, Borgatti, Urgesi and Bardoni2019), although high heterogeneity limited the ability to draw definite conclusions from the results. Altogether, even though these cited reviews showed some promising results, the conclusions were limited by with issues of quality of the evidence and methodological heterogeneity. In addition, most of the studies included in the previous reviews were conducted before 2016 and along with RCT’s included studies with diverse methodological quality. Given the rapidly evolving field of telehealth, technological development as well as the increasing use of telemedicine during the COVID-19 pandemic (Mann et al., Reference Mann, Chen, Chunara, Testa and Nov2020), there is a need for evaluate the current research field and practices of the teleneuropsychological rehabilitation.
Objective
The aim of this systematic literature review was to identify and evaluate the effects of teleneuropsychological rehabilitation on cognitive, behavioral, and socio-emotional functioning in adults and children.
Methods
This systematic review was conducted using the PRISMA reporting protocol (Moher et al., Reference Moher, Liberati, Tetzlaff and Altman2009), as this allows for a standardized non-biased approach to the review. The study protocol has been retrospectively published on the International Platform of Registered Systematic Review and Meta-analysis Protocols (INPLASY) with the registration number INPLASY202330093.
Eligibility criteria
Eligible studies included neuropsychological remote interventions for children and/or adults (aged 0–65 years) with neuropsychological impairments or their caregivers. Articles where the mean age of the participants was more than 65 years were excluded due to greater variability in patients’ health conditions and its impact on intervention outcomes. Given the scarcity of research on teleneuropsychological rehabilitation, there were no restrictions on duration, intensity, co-interventions, comparators, or outcome measures. In addition, a wide spectrum of neuropsychological rehabilitation approaches was included, such as cognitive retraining (paper-pencil and/or computerized), compensatory strategy training, and neuropsychological counseling or support. Remote interventions involving mainly self-training was included if there was regular involvement (synchronous/asynchronous) of the intervention provider. The goal was to achieve a large sample of studies to assess the current state of the field. The final sample of this study was narrowed to randomized controlled trials (RCTs). Experimental study design was considered necessary to determine the validity of the results and to evaluate the evidence supporting effectiveness. More detailed eligibility criteria are presented in Figure 1.
The eligibility (inclusion and exclusion) criteria employed for the systematic review.
Information sources
Online literature searches were conducted using MEDLINE (PubMed), Web of Science, Scopus, and PsycINFO electronic databases and filtered to include studies published in 2016 or later to focus on recent telerehabilitation practices. All database searches used the same set of search terms combining words “tele," “remote," “virtual," “distance," “online," and “web-based” with words “rehabilitation," therapy, intervention, and training and words “neuropsychology," “neurology," “neuropsychiatry," “neurodevelopmental," “neurocognitive," and “cognitive.” These terms were searched from title, abstract, and keyword fields, if available. In addition, searches were supplemented from relevant Finnish electronic databases (Journal.fi; Helda, psykologia.fi., Julkari) with the same set of search terms. Review articles identified in the searches and the accompanying reference lists were screened using the same criteria. Basic search expression, which was slightly modified according to each database, is presented in Appendix.
Study selection, data collection and extraction
An initial screening of titles and abstracts was carried out to quickly exclude clearly non-eligible articles. If it was unclear whether an article met inclusion or exclusion criterion after review of the title and abstract, the article was passed for full text assessment. Following initial screening, two authors (EN, IS) independently assessed the eligibility of full text articles using the previously described eligibility criteria. The same authors further discussed any article in which they provided discrepant ratings on whether it satisfied eligibility criteria. If there were remaining uncertainties regarding eligibility, the final decision was made based on a group discussion with two additional authors (MP, EP). A data extraction form was used for collecting the following information for each eligible study: (1) baseline information (e.g., demographics, study design), (2) intervention characteristics (e.g., content, duration), (3) outcome measures and their results, (4) effect sizes (ESs) if reported, and (5) main findings, as reported in the study. The data extraction form was developed based on the data extraction template by Cochrane Developmental, Psychosocial and Learning Problems group (available on https://dplp.cochrane.org/data-extraction-form).
Risk of bias assessment
Risk of bias (RoB) of the selected studies was assessed using The Cochrane’s RoB tool and by applying the guidelines of The Cochrane Handbook (Higgins et al., Reference Higgins, Altman, Gotzsche, Juni, Moher, Oxman, Savovic, Schulz, Weeks and Sterne2011; Higgins et al., Reference Higgins, Thomas, Chandler, Cumpston, Li, Page and Welch2022). One author (EN) completed the ratings independently and categorized the RoB as being low, high, or unclear. When there was uncertainty in the coding of the RoB, three authors (EN, IS, MP) discussed it and decided by consensus.
Data analysis and synthesis
Due to the heterogeneity of the included studies, a meta-analysis could not be undertaken. Thus, a narrative synthesis of quantitative results was used. In addition, for each eligible study, the reported ES was used, wherever effect was statistically significant (p-values < 0.05). If ES was not reported, Hedges’ g was computed using R version 3.6.2 (R Core Team, 2019) according to means and standard deviations reported in the study, whenever possible, and the ES’s were categorized as small (g = 0.2), moderate (g = 0.5), or large (g = 0.8). However, these categories are somewhat arbitrary and should be interpreted only as rough estimates of the ES. In the case of insufficient information regarding the ES, the authors were contacted. However, no responses were obtained.
Results
Results of the search and study selection
All databases were searched in November 2021. The processes for searching the literature and choosing final articles for this study are shown in Figure 2. There was 85% initial agreement between 2 researchers concerning the eligibility of the 292 studies. Altogether, there were 29 studies considered as eligible. Among these studies, there were 14 RCTs, 4 controlled studies with a pre-post-study design, and 11 intervention studies without a control group. Of the remaining articles, only RCTs were included in the qualitative synthesis based on the predefined criteria.
The flow chart of the process of choosing the articles on teleneuropsychological rehabilitation.
Intervention findings
Detailed descriptions of the interventions are presented in Table 1. The most used medium for remote communication was telephone (n = 7, 50%), followed by videoconference (n = 5, 36%). In addition, two studies had e-mail as an alternative to telephone and in two studies, the medium remained unclear. Furthermore, six studies (43%) combined remote sessions with face-to-face meetings. For children/adolescents, the most typical diagnosis was TBI (n = 6, 43%). Regarding children, other patient groups were pediatric onset MS or ADHD (n = 1), childhood cancer survivors (acute lymphoblastic leukemia/brain tumor; n = 1) and unilateral cerebral palsy (n = 1). The most typical diagnosis for adults was MS (n = 2) or stroke (n = 2). Additionally, there were one study including patients with epilepsy. The total length of the interventions varied from 5 weeks to 6 months and intensities varied from weekly/biweekly sessions to 6 sessions per week. Next, synthesis of the findings of the studies is summarized in Tables 1 and 2 (total n = 1063), with studies with children/adolescents synthesized separately from those with adults. From here on, the term therapist refers to the person who provided the intervention independent of his/her exact profession, unless specification is needed.
Descriptive summary of the randomized controlled studies on teleneuropsychological rehabilitation included in the analyses
Abbreviations for diagnoses: ABI = Acquired brain injury, ADHD = Attention-Deficit/Hyperactivity Disorder, ALL = Acute lymphoblastic leukemia, BT = Brain tumor, POMS = Pediatric onset multiple sclerosis, RRMS = Relapsing-remitting multiple sclerosis, SPMS = Secondary progressive multiple sclerosis, TBI = Traumatic brain injury, UCP = Unilateral cerebral palsy. Abbreviations for interventions: CAPS = Communication and problem-solving, ELT = Experimental linguistic treatment, f2f = face-to-face, HBCACR = Home-based computer assisted cognitive rehabilitation, HOBSCOTCH = The HOme-Based Self-management and COgnitive Training CHanges lives, I-InTERACT = Internet-Based Interacting Together Everyday: Recovery After Childhood TBI, IRC = Internet resources comparison, ST = Specific training, nST = Non-specific training, TOPS-Family = Teen Online Problem-Solving with Family, TOPS-TO = Teen Online Problem-Solving with Teen Only, VRRS = Virtual reality rehabilitation system.
Summary of the findings of the randomized controlled studies included in the analyses
Note: Participants in all studies had cognitive impairments or cognitive symptoms. Sample sizes indicate the number of randomized participants. *Statistical effects are reported only for statistically significant results. (Study 1) Effect size (ES) is Hedges’ g computed by the authors of the present study based on reported means and standard deviations. (Study 2) ES is Hedges’ g computed by the authors of the present study based on mean changes in intervention and control groups. (Study 3) ES is Hedges’ g computed by the authors of the present study based on reported means and standard deviations. (Study 4) ES reflect within-group changes derived from the mixed model results and are similar to Cohen’s d and Hedge’s g. (Study 5) ES regarding the linear mixed-effects model between-group differences. The group difference is mostly result of differences between post-hospitalization and post-intervention timepoints. Regarding EQ-5D, statistical effect was found only concerning the visual analogue scale (VAS). (Study 6) Hedge’s g computed from change scores between groups. (Study 7) Mann-Whitney ES r, calculated same way than in Vilou et al. (Reference Vilou, Bakirtzis, Artemiadis, Ioannidis, Papadimitriou, Konstantinopoulou, Aretouli, Messinis, Nasios, Dardiotis, Kosmidis and Grigoriadis2020). Intention-to-treat-analysis ES in brackets. (Study 10) ES is Hedges’ g computed by the authors of the present study based on reported means and standard deviations. (Study 13) ES is calculated using r = Z/sqrt(N), where r is the ES, Z is the score of Mann-Whitney U test, and N is the study sample. (Study 14) ES is Hedges’ g computed by the authors of the present study based on reported unadjusted post-intervention means and standard deviations. This may have little bias on the computed ES, but should be quite good approximate, because the group difference was very small in the baseline [TOPS-Family mean = 57.88 (SD = 12.89); TOPS-TO mean = 57.61 (SD = 12.81)]. Abbreviations for diagnoses: ABI = Acquired brain injury, ADHD = Attention-Deficit/Hyperactivity Disorder, ALL = Acute lymphoblastic leukemia, BT = Brain tumor, POMS = Pediatric onset multiple sclerosis, RRMS = Relapsing-remitting multiple sclerosis, SPMS = Secondary progressive multiple sclerosis, TBI = Traumatic brain injury, UCP = Unilateral cerebral palsy. Abbreviations for statistical effects: BL-PI = Within-group effect between baseline and post-intervention, BL-6MF = Within-group effect between baseline and 6 months follow-up, PI = Between-group effect in baseline-adjusted post-intervention scores or in change scores between baseline and post-intervention. Abbreviations for interventions: CAPS = Communication and problem-solving, ELT = Experimental linguistic treatment, f2f = face-to-face, HBCACR = Home-based computer assisted cognitive rehabilitation, HOBSCOTCH = The HOme-Based Self-management and COgnitive Training CHanges lives, Computerized cognitive training = CCT, I-InTERACT = Internet-Based Interacting Together Everyday: Recovery After Childhood TBI, IRC = Internet resources comparison, ST = Specific training, nST = Non-specific training, TOPS-Family = Teen Online Problem-Solving with Family, TOPS-TO = Teen Online Problem-Solving with Teen Only, VRRS = Virtual reality rehabilitation system. Abbreviations for outcome measures: ADRS = Aphasic Depression Rating Scale, AWMA = Automated Working Memory Assessment, BDI-fast screen = Beck Depression Inventory Fast Screen for Medical Patients, BRIEF-BRI = BRIEF Behavior Regulation Index, BRIEF-GEC = Behavior Rating Inventory of Executive Functioning – Global Executive Composite, BVMT-R = Brief Visuospatial Memory Test-Revised, CBCL = Child behavior checklist, CPT-II = Continuous Performance Test II, ENPA = Esame Neuropsicologico Per l'Afacia, EQ-5D – VAS = Euro-Qol-5D Visual analog scale, GVLT = Greek Verbal Learning Test, MFIS = Modified Fatigue Impact Scale, QOLIE-31 = Quality of Life in Epilepsy, RBANS = Repeatable Battery for the Assessment of Neuropsychological Status, SDMT = Symbol Digits Modalities Test, SPART = Spatial Recall Test, SPART-D = SPART Delayed, TMT = Trail Making Test, TT = Token test, WIAT-II = Wechsler Individual Achievement Tests II, WISC-IV = Wechsler Intelligence Scale for Children IV.
Telerehabilitation for children and adolescents
There were nine studies evaluating the effectiveness of telerehabilitation for children and adolescents with neuropsychological impairments, comprising a total of 844 participants. Regarding the content, three types of interventions were identified: (1) family-centered interventions with strategy training focused on parenting skills, cognitive, behavioral, or emotional functioning or a combinations of these (studies 1, 4, 11, and 14); (2) computerized cognitive training (studies 3, 7, and 10); and (3) cognitive training combined with motor training (i.e., gross motor, cognitive and upper limb tasks; studies 8, and 9). In all studies of the first intervention type (1), the telerehabilitation included synchronous videoconferencing with self-guided material reviewed between the videoconferencing sessions, with the exception of one study with unclear medium of delivery of the online sessions (study 4). For the second type (2), the cognitive training was home-based self-training with regular phone calls with the therapist. For the third type (3), the mode of the delivery was the same as in the second type.
Evidence for effectiveness was found mainly for the interventions offering computerized cognitive training. These interventions targeted specific cognitive domains (working memory or attention) with improvements in single (study 7) or several (studies 3 and 10) cognitive domains with ESs varying from small to large when compared to active or passive control groups. While there were positive near transfer effects, the sample sizes were relatively small and evidence for transfer of specific training to non-trained domains (e.g., functional cognition) was somewhat mixed. One attention training intervention showed improvements in some non-trained cognitive domains when compared to non-specific training (study 10), although there was no follow-up included. Furthermore, in this study improvements were found only in small group of children with pediatric onset MS, but not in children with ADHD. Regarding the interventions with working memory training, one study had positive long-term changes in some non-trained functions when compared to a passive control group (study 3), and another study showed long-term improvements on a single non-trained task when compared to an active control group (study 7). None of these studies included psychosocial or emotional outcome measures.
In relation to other types of telerehabilitation, the evidence supporting the effectiveness of family-centered interventions was somewhat mixed. No significant differences were found when cognitive outcomes across treatment groups and active or passive control groups were compared. For one family-centered intervention, there were significant improvements with moderate ES only for the psychological outcome (withdrawn/depressive behavior) compared to active control group (study 1); however, some significant changes were found for family-centered interventions compared to the active control group when subgroup analyses were applied. There were improvements in executive functioning and psychological outcomes in children and adolescents of lower educated caregivers (studies 1 and 14). In addition, children of caregivers with greater symptoms of depression benefitted from the intervention (study 1). Furthermore, improvements were found in social competence depending on age of adolescent and severity of TBI (study 11). Finally, there was no evidence of an effect of home-based cognitive training combined with motor training on cognitive or psychosocial functioning (studies 8 and 9).
Telerehabilitation for adults
There were five studies evaluating the effectiveness of telerehabilitation for adults with neuropsychological impairments, comprising a total of 219 participants. The content of the rehabilitation was focused on home-based computerized cognitive training with regular contact via videoconferencing or telephone with the therapist (studies 5, 6, 12, and 13). Considering the method of delivery of these interventions, the intervention was mainly comprised of self-administered cognitive training with at least weekly contact with the therapist by telephone (study 6) or via videoconferencing (study 5); the contact modality was unclear in one study (study 13) and in another study, contact via videoconference was mainly for monitoring (study 12). In addition to these interventions, one intervention included training of cognitive and self-management strategies with or without computerized cognitive training by telephone with the therapist (study 2).
All studies with adults offered support for the effectiveness of telerehabilitation, at least to some extent. Interventions including computerized cognitive training significantly improved specific or various domains of functioning with ESs varying from small to large, although for one study there was insufficient information to calculate ESs (study 12). In addition, sample sizes of these studies were relatively small and long-term effects remained unclear, as no follow-up was included. Regarding the transfer effect, three studies showed positive post-treatment changes in some psychosocial domains (i.e., quality of life, depression), in addition to the positive near transfer effects on cognitive domains, when compared to face-to-face rehabilitation (studies 5 and 12) or non-specific training (study 6). In one of these studies, the target of the intervention was mainly verbal skills (study 5); however, the ES’s (measured with R-squared) of this study were unusually large for psychological research, which should be interpreted with caution. The interventions of two other studies (studies 6 and 12) were targeting various cognitive domains. Additionally, one study with an intervention targeting multi-domain cognition showed improvements in trained domains when compared to a passive control group without psychosocial or emotional outcome measures included (study 13). Finally, the only intervention with compensatory cognitive strategy training over telephone (study 2) demonstrated significantly improved post-treatment quality of life and cognitive functioning when compared to usual care, suggesting some immediate far transfer effects.
Risk of bias
The detailed figure of risk of bias (RoB) assessment is available in the Supplementary Material 1. The most common (100%) methodological flaw was the blinding of participants and personnel. Regarding the random sequence generation, 71% of studies reported it adequately. In one study, the randomization was done based on participants’ distance from the hospital, which was assessed as a high RoB. The allocation concealment was reported adequately in 50% of studies (e.g., sealed envelopes), whereas 50% of studies did not report it. In 57% of studies, the blinding of outcome assessors was successful, but in some studies (29%) this was not reported, or the assessors were aware of the randomization (14%). The attrition bias (incomplete outcome data) was assessed as low for 71% of studies (e.g., adequate intent-to-treat analyses, attrition rates equally distributed across groups). This information was unclear for 21% of studies. In addition, the attrition rates of one study were unbalanced across groups reflecting high RoB. In regards to selective outcome reporting, most of the studies (93%) did not report a protocol for the study, or it was too imprecise for determining the RoB.
Regarding other potential bias, there were no other sources of bias found in 29% of the studies; however, this remained unclear for 43% of studies that did not report the compliance of the planned intervention (studies 2, 5, 6, 10, 11, and 12). Furthermore, the compliance of intervention was weak for three studies with participants reaching the planned training dosage poorly (studies 8, 9, and 13), which should be considered as possible RoB (Munder & Barth, Reference Munder and Barth2018). In addition, in one study, intervention and control groups differed in ethnicity (study 4).
Discussion
This systematic review included 14 RCT studies with varying interventions, comparators, as well as participants’ characteristics and outcome measures, which reflects the cross-disciplinary nature of neuropsychological rehabilitation (Wilson, Reference Wilson2008). Heterogeneity of studies has been common in previous systematic reviews on telerehabilitation as well (e.g., Amatya et al., Reference Amatya, Galea, Kesselring and Khan2015; Betts et al., Reference Betts, Feichter, Kleinig, O'Connell-Debais, Thai, Wong and Kumar2018; Camden et al., Reference Camden, Pratte, Fallon, Couture, Berbari and Tousignant2020). Most of the identified articles focused on children with TBI (six studies), followed by adults with MS (two studies) or stroke (two studies). Most typical intervention type was computerized cognitive training with regular contact with a therapist via telephone or videoconferencing (seven studies), or family-centered strategy training via videoconferencing (four studies). In line with this study, previous reviews focusing on neuropsychological or cognitive rehabilitation have found a growing trend to utilize computerized programs in interventions involving cognitive training (Rosti‐Otajärvi & Hämäläinen, Reference Rosti‐Otajärvi and Hämäläinen2014; Cicerone et al., Reference Cicerone, Goldin, Ganci, Rosenbaum, Wethe, Langenbahn, Malec, Bergquist, Kingsley, Nagele, Trexler, Fraas, Bogdanova and Harley2019; Resch et al., Reference Resch, Rosema, Hurks, de Kloet and van Heugten2018). However, a shift toward integrating emotional and cognitive interventions in face-to-face neuropsychological rehabilitation has also been detected (Cicerone et al., Reference Cicerone, Goldin, Ganci, Rosenbaum, Wethe, Langenbahn, Malec, Bergquist, Kingsley, Nagele, Trexler, Fraas, Bogdanova and Harley2019; Wilson, Reference Wilson2008).
Effectiveness of telerehabilitation
There was some support for the effectiveness of telerehabilitation in home-based computerized cognitive training with regular contact with the therapist via telephone or videoconferencing for children (studies 3, 7, and 10) and adults (studies 5, 6, 12, and 13). All of these interventions targeting specific or various cognitive domains improved single or several domains of cognitive functioning when compared to active or passive control groups, with ESs varying from small to large. In this review, improvements were found particularly in immediate trained cognitive functions suggesting near transfer effects with only few studies (3 and 7) showing long-term improvements in non-trained cognitive domains. Only few studies incorporated emotional and/or psychosocial outcome measures (5, 6, and 12) or included follow-up (3 and 7). The results showing effects mainly in trained skills with a lack of generalizability are in line with previous studies on the effectiveness of computerized cognitive interventions (Melby-Lervåg et al., Reference Melby-Lervåg, Redick and Hulme2016; Diamond & Ling, Reference Diamond and Ling2016; Lynch, Reference Lynch2002; Sigmundsdottir et al., Reference Sigmundsdottir, Longley and Tate2016; Mingming et al., Reference Mingming, Bolun, Zhijian, Yingli and Lanshu2022; Zucchella et al., Reference Zucchella, Capone, Codella, Vecchione, Buccino, Sandrini, Pierelli and Bartolo2014). However, previous studies have also demonstrated promising findings on far transfer effects when computerized training is implemented with face-to-face interaction with the professional (Cicerone et al., Reference Cicerone, Goldin, Ganci, Rosenbaum, Wethe, Langenbahn, Malec, Bergquist, Kingsley, Nagele, Trexler, Fraas, Bogdanova and Harley2019; De Luca et al., Reference De Luca, Calabrò, Gervasi, De Salvo, Bonanno, Corallo, De Cola and Bramanti2014; Fernández et al., Reference Fernández, Bringas, Salazar, Rodríguez, García and Torres2012). For adults (studies 5, 6, and 12), there were some immediate improvements in emotional and psychosocial functioning (e.g., quality of life). Even though the sample sizes were relatively small, these indications of some far transfer effects are promising, since previous systematic reviews on computerized-cognitive rehabilitation for similar populations have demonstrated mixed evidence of training effects on emotional or psychosocial functioning (Lampit et al., Reference Lampit, Heine, Finke, Barnett, Valenzuela, Wolf, Leung and Hill2019; Zhou et al., Reference Zhou, Feng, Li, Xu, Wu and Li2021).
Altogether, most of the studies on remote computerized interventions for children or adults focused mainly on impairment level of functioning with scarce evidence of the long-term meaningful, everyday functions of cognition, mood, behavior, or participation (e.g., social, educational, vocational). Finally, only two studies (study 5 and 12) compared the intervention to face-to-face rehabilitation, therefore more research is also needed to evaluate whether telerehabilitation can be used as an alternative to traditional rehabilitation.
Regarding the effectiveness of other types of teleneuropsychological rehabilitation, the results were somewhat mixed. No support was found for the effectiveness of interventions combining cognitive training with motor training aimed for children (studies 8 and 9). These interventions had challenges with treatment compliance and remote technology. In addition, these studies were conducted by the same research group, while the study samples were different. The sole intervention with compensatory cognitive and self-management strategy training with psychoeducation via telephone for adults (study 2) significantly improved post-treatment quality of life and cognitive functioning of participants with epilepsy, suggesting some far transfer effects. Although the evidence for the effectiveness of neuropsychological rehabilitation in epilepsy is limited, the findings of this review are consistent with previous reviews on face-to-face interventions suggesting the usefulness of more comprehensive approach for patients with epilepsy (Farina et al., Reference Farina, Raglio and Giovagnoli2015) as well as strategy-based training for patients with seizure-related deficits (Langenbahn et al., Reference Langenbahn, Ashman, Cantor and Trott2013).
Instead, regarding interventions for children, the results of family-centered interventions (studies 1, 4, 11, and 14) showed only tentative support for the intervention improving single psychosocial outcomes. This finding is consistent with prior studies with interventions focused on family members of individuals with TBI, which demonstrated some level of positive outcomes for the telehealth interventions but also a lack of high-quality evidence for the effectiveness of such programs (Brown et al., Reference Brown, Whittingham, Boyd and Sofronoff2013; Rietdijk et al., Reference Rietdijk, Togher and Power2012). The investigation of effectiveness of the family-centered interventions was based mainly on subjective parent-report outcome measures. This may be problematic due to factors, such as parental mental distress or low socio-economic status of the family, which may impact parental report (Najman et al., Reference Najman, Williams, Nikles, Spence, Bor, O'Callaghan, Le Brocque, Andersen and Shuttlewood2001; Narad et al., Reference Narad, Raj, Yeates, Taylor, Kirkwood, Stancin and Wade2019); however, the studies included in this review investigated some of these variables, such as caregivers’ education (studies 1, 4, 11, and 14) or parental depression (study 1) and found mixed results: these variables were significant for two studies (1, 14).
Implications for clinical practice
Regarding studies with computerized cognitive training, interventions included varying amount of interaction with the therapist. The content of remote interaction was typically concerning aspects of cognitive training (e.g., motivation) or it was unclear due to vague description in the articles. This type of involvement of the provider of the intervention is quite common in studies concerning home-based computerized cognitive training (Sigmundsdottir et al., Reference Sigmundsdottir, Longley and Tate2016). Future studies on remote interventions should describe and evaluate different aspects (e.g., content) of the interaction between the professional and patient more precisely. Individual needs of the patients should be the basis of neuropsychological rehabilitation (Wilson, Reference Wilson2008), and it remained somewhat unclear if these needs were met in the studies described above; however, interventions included specific rehabilitation program with individualized training (e.g., difficulty level adjustments, immediate feedback, systemized delivery), which are found to be advantages of these types of programs for patients with neurological impairments and features in promoting neuroplasticity (Cicerone et al., Reference Cicerone, Goldin, Ganci, Rosenbaum, Wethe, Langenbahn, Malec, Bergquist, Kingsley, Nagele, Trexler, Fraas, Bogdanova and Harley2019; Cramer et al., Reference Cramer, Sur, Dobkin, O'Brien, Sanger, Trojanowski, Rumsey, Hicks, Cameron, Chen, Chen, Cohen, deCharms, Duffy, Eden, Fetz, Filart, Freund, Grant and Vinogradov2011). Along with focusing on cognitive training, interventions aiming to meaningful functional improvements of everyday life may be appropriate approach for many patients.
To keep up with technological development this review focused on recent tele-practices on neuropsychological rehabilitation. The most common methods of delivering telerehabilitation was via telephone followed by videoconferencing. Videoconferencing (Appleby et al., Reference Appleby, Gill, Hayes, Walker, Walsh and Kumar2019; Camden et al., Reference Camden, Pratte, Fallon, Couture, Berbari and Tousignant2020) or phone calls (Ownsworth et al., Reference Ownsworth, Arnautovska, Beadle, Shum and Moyle2018; Betts et al., Reference Betts, Feichter, Kleinig, O'Connell-Debais, Thai, Wong and Kumar2018) has been common technology in previous systematic reviews on telerehabilitation for neurological populations, as well. Even though most of the studies showed effectiveness used mainly telephone contact, it is challenging to draw conclusions regarding whether the technology used influenced the outcomes due to the heterogeneity of the studies included in this review. In addition, there could be more relevant factors associated to positive changes. It has addressed that individual preferences and therapy approaches could be key features guiding the choice of technology (Camden et al., Reference Camden, Pratte, Fallon, Couture, Berbari and Tousignant2020; Laver et al., Reference Laver, Adey-Wakeling, Crotty, Lannin, George and Sherrington2020).
Nevertheless, considering the clinical practice, telephone could be used as a delivery method along with videoconferencing. Further, this is notable information considering the telerehabilitation in low-income communities as telephone can offer potential alternative to videoconferencing, which can be inaccessible to many of these communities due to its requirements of fast internet connection (Annaswamy et al., Reference Annaswamy, Verduzco-Gutierrez and Frieden2020). In addition, previous systematic review focused on TBI (Betts et al., Reference Betts, Feichter, Kleinig, O'Connell-Debais, Thai, Wong and Kumar2018), and now this review found that teleneuropsychological rehabilitation may be suitable delivery method for adults with epilepsy, MS, and stroke as well.
Risk of bias of the included studies
The risk of bias of attrition and randomization process was low for most of the studies, although many studies did not report precise information on allocation concealment. The most common risk of bias was the blinding of the personnel and participants, which is very usual in psychological interventions (Juul et al., Reference Juul, Gluud, Simonsen, Frandsen, Kirsch and Jakobsen2021). The providers of the intervention must know the type of intervention they deliver, and the participants have opinions about the treatment, as well (Munder & Barth, Reference Munder and Barth2018). Hence, it is almost impossible to fulfill every criterion with neuropsychological rehabilitation interventions; however, there were many studies without information on research protocols (selective outcome bias), or the completion of the intervention as planned. The latter was the case especially for interventions including computerized cognitive training for adults, which would have been important information for potential source of bias and acceptability of the treatment (Higgins et al., Reference Higgins, Thomas, Chandler, Cumpston, Li, Page and Welch2022; Munder & Barth, Reference Munder and Barth2018), especially since these interventions included primarily self-training.
Limitations
A comprehensive search was performed, and two independent authors evaluated the criteria of eligibility to reduce potential bias in the review process. Nonetheless, a publication bias cannot be excluded with the possibility of some studies being missed. Additionally, relevant studies with unclear or non-significant results may have remained unpublished. Due to heterogeneity of studies included, strong recommendations for specific subgroups could not be made. There were many studies with relatively small sample sizes, particularly for the cognitive training interventions, which can diminish the generalizability of the results. In addition, this review focused on children and adults (0–65 years), hence the findings of this review may not be generalizable to older adults with neuropsychological impairments. Although RCTs provide a strong evidence-base to the results, the inclusion of other designs could have given more information on telerehabilitation. For example, it may be more difficult to conduct RCT for more holistic programs (Khan et al., Reference Khan, Turner-Stokes, Ng, Kilpatrick and Amatya2007).
Conclusions
This systematic review found promising preliminary support for the effectiveness of computerized cognitive training with regular interaction with a therapist over telephone or via videoconference; however, it was challenging to draw conclusions on the effectiveness of teleneuropsychological rehabilitation, since interventions and comparators, as well as participants’ characteristics and outcome measures, varied significantly. Additionally, only a few of the studies included in this review integrated cognitive, emotional, and psychosocial aspects, and many studies had small sample sizes, as well. Despite these limitations, preliminary support for the effectiveness of computerized cognitive training on impairment level of cognitive functioning was evident for seven out of seven RCTs targeting specific or various cognitive domains with comparators of face-to-face rehabilitation (two studies), non-specific cognitive training (three studies), as well as passive control groups (two studies). However, only few of these studies incorporated follow-up or emotional or psychosocial measures, and their impact on long-term meaningful, everyday functioning remained unclear. For family-centered interventions, the results were somewhat mixed with improvements only noted in single psychosocial outcomes. No support was found for the effectiveness of interventions combining cognitive and motor training.
The field of teleneuropsychological rehabilitation is still evolving alongside the technological development. Thus, including RCT studies published in 2016 or later, this systematic review offers important information on the current state of experimental research as well as the current practices in teleneuropsychological rehabilitation for various patient groups; however, more research is needed, to draw conclusions regarding whether it can be used as an alternative to face-to-face rehabilitation. Furthermore, there is a lack of research investigating the long-term clinical benefits of teleneuropsychological rehabilitation on real-life functioning, which is crucial in the assessment of effectiveness of rehabilitation. Finally, there is also a need for research with more comprehensive rehabilitation approaches to reliably assess the effectiveness of wide spectrum of teleneuropsychological interventions.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S1355617723000565
Acknowledgements
We would like to thank the Research and Editing Consulting Program of the International Neuropsychological Society for editorial assistance.
Funding statement
This work was supported by the Social Insurance Institution of Finland (grant number 65/26/2021).
Competing interests
None.
Appendix: Search terms
(“Telerehab*” OR “Tele-rehab*” OR “Teletherap*” OR “tele-therap*” OR “Teleintervent*” OR “Tele-intervent*” OR “teletraining” OR “Tele-training” OR “Remote rehab*” OR “Remote therap*” OR “Remote intervent*” OR “Remote training” OR “Virtual rehab*” OR “Virtual therap*” OR “Virtual intervent*” OR “Virtual training” OR “Distance rehab*” OR “Distance therap*” OR “Distance intervent*” OR “Distance training” OR “Online rehab*” OR “Online therap*” OR “Online intervent*” OR “Online training” OR “Web-based rehab*” OR “Web-based therap*” OR “Web-based intervent*” OR “Web-based training”) AND (“Neuropsycholog*” OR “Neurolog*” OR “Neuropsychiatr*” OR “Neurodevelop*” OR “Neurocognit*” OR “Cognit*”)
These terms were searched from title, abstract, and keyword fields, if available.
Open access
