You are on page 1of 13

RESEARCH/Original Article

Journal of Telemedicine and Telecare


0(0) 1–13
Cognitive telerehabilitation in mild ! The Author(s) 2017
Reprints and permissions:
cognitive impairment, Alzheimer’s sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1357633X17740390

disease and frontotemporal dementia: journals.sagepub.com/home/jtt

A systematic review

Maria Cotelli1, Rosa Manenti1, Michela Brambilla1, Elena Gobbi1,


Clarissa Ferrari2, Giuliano Binetti3,4 and Stefano F Cappa5,6

Abstract
Introduction: Given the limited effectiveness of pharmacological treatments, non-pharmacological interventions in neurode-
generative diseases have gained increasing attention in recent years and telerehabilitation has been proposed as a cognitive
rehabilitation strategy. The purpose of this systematic review is to examine the evidence for the efficacy of cognitive telereh-
abilitation interventions compared with face-to-face rehabilitation in patients with mild cognitive impairment, Alzheimer’s
disease and frontotemporal dementia.
Methods: In accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, a
systematic search of the Medline database was conducted. Out of 14 articles assessed for eligibility, five studies were identified,
three in participants with mild cognitive impairment or Alzheimer’s disease, two in patients with primary progressive aphasia.
Results: The Physiotherapy Evidence Database scale was used to assess the methodological quality of four out of five studies
included in this systematic review, with only one report receiving a high-quality rating. Effect-size analysis evidenced positive
effects of telerehabilitation interventions, comparable with those reported for face-to-face rehabilitation.
Discussion: The available evidence for the effectiveness of cognitive telerehabilitation is limited, and the quality of the evidence
needs to be improved. The systematic review provides preliminary evidence suggesting that cognitive telerehabilitation for
neurodegenerative disease may have comparable effects as conventional in-person cognitive rehabilitation.
Keywords
Cognitive, telerehabilitation, dementia, mild cognitive impairment, progressive aphasia

Date received: 25 September 2017; Date accepted: 12 October 2017

and social cognitive abilities reflect the location of the


Introduction
underlying pathology.
Increases in life expectancy during the last decades The FTD spectrum comprises the behavioural variant
have resulted in a large number of people living of FTD (bvFTD) and primary progressive aphasias
to old ages and an increased risk of developing neuro- (PPA). The bvFTD is characterized by behavioural
degenerative diseases. Alzheimer’s disease (AD) and
frontotemporal dementia (FTD), two of the most 1
Neuropsychology Unit, IRCCS Centro San Giovanni di Dio Fatebenefratelli,
prevalent neurodegenerative diseases, are extremely Italy
2
debilitating and increasingly common and affect millions Statistics Service, IRCCS Istituto Centro San Giovanni di Dio
of people worldwide.1,2 Fatebenefratelli, Italy
3
MAC Memory Center, IRCCS Istituto Centro San Giovanni di Dio
These disorders result in an impairment of the indi-
Fatebenefratelli, Italy
vidual’s abilities to perform daily tasks. As their disease 4
Molecular Markers Laboratory, IRCCS Istituto Centro San Giovanni di Dio
progresses, patients become dependent on medical ser- Fatebenefratelli, Italy
5
vices and family support. AD is a progressive disorder IRCCS Istituto Centro San Giovanni di Dio Fatebenefratelli, Brescia, Italy
6
that impacts memory, language and several other NEtS, Scuola Universitaria Superiore IUSS-Pavia, Italy
cognitive functions, while the diagnostic label of FTD
Corresponding author:
encompasses a number of heterogeneous clinical presen- Maria Cotelli, IRCCS Centro San Giovanni di Dio, Fatebenefratelli Via
tations, in which different patterns of neuropsychological Pilastroni, 4 25125 Brescia, Italy.
impairment in linguistic processing, executive function Email: mcotelli@fatebenefratelli.eu
2 Journal of Telemedicine and Telecare 0(0)

changes and alterations in social conduct due to orbito- the telerehabilitation of cognitive abilities among patients
frontal and insular atrophy,3–5 whereas the PPA are an with neurodegenerative diseases.
anatomically and pathologically heterogeneous group The aim of this systematic review was to analyse and
of neurodegenerative disorders6–8 characterized, in the synthesize the evidence of the efficacy of cognitive telereh-
early stages, by a selective deterioration of language.9 abilitation interventions in patients with MCI, AD and
The canonical PPA syndromes include different entities: FTD compared with conventional FTF rehabilitation.
(a) non fluent/agrammatic variant PPA (NfPPA/avPPA),
characterized by apraxia of speech and agrammatism, due
to left prefrontal cortex atrophy; (b) semantic variant PPA Methods
(svPPA), in which an impaired semantic knowledge and
Search strategies and selection of the studies
focal left anterior temporal lobe atrophy have been well
documented; and (c) logopenic/phonological variant PPA We performed this systematic review according to the
(lvPPA/pvPPA), characterized by slow spontaneous pro- Preferred Reporting Items for Systematic Reviews and
duction with anomia and phonological errors, defective Meta-Analyses (PRISMA) guidelines and flow diagram.35
repetition and atrophy centred in left posterior-superior The protocol for this review was registered in
temporal and inferior parietal regions.7 the International Prospective Register of Systematic
In addition, the focus of present research in cognitive Reviews (PROSPERO) under registration number
rehabilitation is on the predementia stage (mild cognitive CRD42017069884 and is available in full on the pro-
impairment (MCI) and related concepts), with the aim of gramme website (https://www.crd.york.ac.uk/CRDWeb/
identifying people who are years away from the develop- HomePage.asp). The electronic database Medline
ment of dementia and who may be well-suited for preven- (Pubmed) was searched for records which either contained
tion studies.10–13 the MESH terms ‘Telerehabilitation’ OR ‘Telemedicine’
Given the limited effectiveness of pharmacological treat- OR ‘online training’ AND the terms ‘Mild Cognitive
ments, non-pharmacological interventions to prevent and Impairment’ OR ‘Alzheimer disease’ OR ‘Frontotemporal
treat cognitive deficits and the associated difficulties with dementia’ OR ‘behavioural variant of Frontotemporal
activities of daily living, neurodegenerative disease patients dementia’ OR ‘frontal variant of Frontotemporal demen-
have gained attention in recent years, and among these tia’ OR ‘temporal variant of Frontotemporal dementia’ OR
cognitive training offers a potential approach for dementia ‘primary progressive aphasia’ OR ‘logopenic variant’
prevention and the improvement of cognitive function.14–23 OR ‘semantic variant’ OR ‘semantic dementia’ OR ‘agram-
A critical aspect of cognitive training programmes is that matic variant’ OR ‘non-fluent PPA’. Only English-lan-
the most promising interventions have involved intensive guage articles were selected (see Figure 1). Abstracts were
in-person sessions that are unlikely to be cost-effective or reviewed and all relevant original research articles were
feasible for large-scale implementation.24 reviewed in detail, including a review of the references in
An increasing need for alternative kinds of dementia each publication to identify additional sources. References
service delivery is necessary because of the growing and ‘cited by’ information of identified articles were further
demand and cost of healthcare and is made possible by scanned, but no additional studies were found.
the rapid developments of technology. In particular,
equitable access to services, improvement of quality of
care, ongoing intervention and the promotion of self-man-
Study selection criteria
agement are several benefits to be delivered from telereh- We included studies based on the following criteria: (a)
abilitation. The American Telemedicine Association original research; (b) conducted on participants with MCI
defines telerehabilitation as the delivery of rehabilitation or AD or FTD; (c) comprising telerehabilitation cognitive
services via information and communication technologies interventions; (d) providing at least one outcome on cogni-
(ICT).25 ICT provides services remotely in patients’ homes tive abilities; and (e) published before August 2017. Further,
or other environments, also allowing patients living in articles published in languages other than English, animal
rural settings or those with difficulties in mobility or studies, reports of secondary data such as meta-analyses or
travel expenses access to care.26–31 reviews or letters were excluded.
A variety of technologies are now available to support The selection of studies was first based on a screening
the continuum of care for people with cognitive impair- of the title and abstract, followed by a reading of the full
ment. A systematic review on the use of telerehabilitation text of the remaining reports (see flowchart in Figure 1).
interventions on individuals with physical impairments
showed that the application of this methodology leads
to clinical improvements that are generally equal to
Data collection, extraction and quality assessment
those induced by conventional face-to-face (FTF) Two investigators (MB and RM) examined the eligibility
rehabilitation programs.32 This methodology has been of the studies using a standardized data extraction form,
shown to have a broad potential also for speech-language and disagreements were resolved through consensus
treatment in post-stroke patients.26,29,33,34 To the best of or referral to a third reviewer (SFC). We extracted base-
our knowledge, there is no systematic review available on line information from the individual studies, including
Cotelli et al. 3

Figure 1. Summary of the literature search – PRISMA flow diagram.

publication year, study design, participants characteristics item relates to the external validity (specifically, the par-
and disease type. Moreover, outcome measures scores ticipant selection criteria). The remaining 10 items assess
were extracted at baseline and at post-treatment assess- the internal validity of each trial, and whether the trial
ment visit. Intervention method, such as treatment fre- contains sufficient statistical information to make it inter-
quency, session duration and number of treatment pretable. Thus, the internal validity of each study is
sessions, was extracted in the individual studies. ranked based on a total score out of 10 (i.e. excluding
The methodological quality was assessed using the the first item). On the PEDro scale, the cut-off for high
Physiotherapy Evidence Database (PEDro) scale. The quality of methodology is a score 56/10.36 All scores
PEDro scale consists of 11 criteria. Each satisfied item assigned to each study were agreed upon by consensus
contributes one point to the total PEDro score. The first and are presented in Table 1.
4 Journal of Telemedicine and Telecare 0(0)

Statistical analysis

Eligibility criterion is related to external validity, so it does not reflect the dimensions of quality assessed by the PEDro scale and it is not used to calculate the total score (which is why the 11-item scale gives a
score Level of
(1–10) quality

High
Low

Low

Low
Data were extracted for the description of methodology
and main outcomes. Due to the heterogeneity of outcome
PEDro
Total
measures, it was not possible to perform a meta-analysis.

3
However, whenever possible, an effect size (ES) was cal-
CRITERION 10:

culated for all results mentioned in each selected article.


Point estimate
and variability

Cohen’s d ES for mean differences were computed


reported

using the classic definition for independent samples.37,38


In the case of a paired sample, the Cohen’s d was adjusted
1

0
for the correlation coefficient according to the formula
Between-group
CRITERION 9:

reported in Dunlap et al.38 We are aware of the methodo-


difference

logical doubts regarding the appropriateness of the adjust-


reported

ment of Cohen’s d for paired sample,38 especially when the


correlation is not known or impossible to estimate from
0

1
the data reported in the primary studies (see, e.g. Morris
CRITERION 4: CRITERION 5: CRITERION 6: CRITERION 7: CRITERION 8:

treat analysis
Intention to-

and DeShon39). We thus performed comparisons of the


ES adjusted and not adjusted for a hypothesized correl-
PEDro: Physiotherapy Evidence Database. Each criterion was scored as either 1 or 0 according to whether the criteria was met or not, respectively.

ation coefficient. Considering the relatively small sample


1

size of the studies in analysis, a plausible pre-/post-treat-


ment correlation coefficient has been chosen (equal to 0.5)
for all treatments analysed by paired samples Cohen’s d.
dropouts
<15%

An evaluation in terms of magnitude (absolute value of


the ES according to Cohen’s classification: low ES 4 0.2;
1

medium 0.2 < ES 4 0.5; quite large 0.5 < ES 4 0.8; very
large ES > 0.8) and 95% confidence intervals (CIs) were
Assessors

reported.
blinding

0
Table 1. Evaluation of the quality criteria fulfilment and risk of bias assessment with PEDro scale.

Results
Therapists

Baseline characteristics of included studies


blinding

Out of 14 articles assessed for eligibility, five works pub-


0

lished between 2005 and 2016 fit the inclusion criteria for


the systematic review (Figure 1, Table 2). Three of the
selected studies included participants with MCI or AD,
Subjects
blinding

whereas two of them included patients with PPA. None


of the selected studies investigated the effects of telereh-
0

0
CRITERION 3:

abilitation in patients with bvFTD.


A total of 89 patients with MCI or dementia and 23
similar at
baseline

healthy older adults were enrolled in the selected studies.


CRITERION 1: CRITERION 2: Groups

The number of patients enrolled in the telerehabilitation


1

intervention groups varied from three to 18, with a total of


44 patients receiving cognitive training via telerehabilita-
Concealed
allocation

tion and 45 patients receiving a FTF intervention.


0

Description of the selected studies


Regarding the efficacy of the cognitive intervention pro-
criteriaa allocation
Eligibility Random

gram that involved telerehabilitation intervention, Poon


et al.40 first explored in a randomized controlled trial
1

(RCT) the feasibility and clinical outcome of a cognitive


intervention program for patients with MCI and mild
maximum score of 10).

dementia using telemedicine versus a conventional FTF


1

method. A total of 22 subjects with mild dementia or


Vermeij et al.,

Meyer et al.,

MCIs were recruited from a community centre, and 12


Poon et al.,

Jelcic et al.,
200540

201441

201642

201643

sessions of cognitive intervention were conducted via


Study

videoconferencing or by FTF method. The authors


found a comparable cognitive improvement in both
a
Table 2. Review of studies that assessed the effects of telerehabilitation on the cognitive functions in patients with MCI and dementia (see ‘‘Methods’’ section and Figure 1 for inclusion
criteria).

Patients’ characteristics
Cotelli et al.

Intervention
Number and type Cognitive global description (total
of patients (proto- Age assessment number of sessions, Follow-
Study col design) Mean (SD) Mean (SD) session duration) up Outcome measures Results

Poon et al. 22 patients with NA MMSE: Videoconference No Global assessment Overall, the telerehabilitation group and
(2005)40 mild dementia or Telerehabilitation cognitive inter- C-MMSE the FTF group did not differ signifi-
MCI (randomized group: 18.73 vention (12 ses- HDS cantly in neuropsychological outcomes.
controlled trial; (2.15); sions over 6 Memory In particular:
11 telerehabilita- FTF group: 18.27 weeks) vs. FTF C-RBMT – Improvement in C-MMSE, C-RBMT
tion, 11 FTF) (2.41) cognitive inter- Ad hoc satisfaction and HDS in telerehabilitation and FTF
vention (12 ses- questionnaire groups;
sions over 6 – Improvement in HDS spatial con-
weeks) struction subtest only in FTF group;
– 90% of the patients were highly satis-
fied by the telerehabilitation treatment.
Jelcic et al. 27 patients with Telerehabilitation MMSE: Tele, lexical-seman- No Global assessment Overall the telerehabilitation group and
(2014)41 AD (pilot con- group: 86.0 (5.1); Telerehabilitation tic stimulation MMSE the FTF lexical-semantic stimulation
trolled trial; 7 FTF group: 82.7 group: 23.7 (2.8); (24 sessions over Language group did not differ significantly in
telerehabilitation, (6.0); FTF group: 24.9 12 weeks, 60 Verbal Naming Test neuropsychological outcomes. In par-
10 FTF; 10 Unstructured (2.5); minutes) vs. FTF Phonemic verbal ticular:
unstructured cognitive stimula- Unstructured lexical-semantic fluency – Improvement in MMSE in tele, lexical-
cognitive tion group: 82.3 cognitive stimula- stimulation (24 Semantic verbal semantic stimulation and FTF lexical-
stimulation) (5.9) tion group: 24.8 sessions over 12 fluency semantic stimulation groups;
(2.7) weeks, 60 min- Memory – Improvement in phonemic and
utes) vs. unstruc- Story recall semantic verbal fluency only in tele,
tured cognitive (immediate recall) lexical-semantic stimulation group;
stimulation (24 Story recall (delayed – Improvement in delayed verbal
sessions over 12 recall) memory only in FTF lexical-semantic
weeks, 60 AVLT (immediate stimulation group;
minutes) recall) – Improvement in story immediate
AVLT (delayed recall) recall only in FTF lexical-semantic
Digit span forward stimulation group;
Digit span backward – Improvement in Digit span forward
Rey–Osterrieth only in FTF lexical-semantic stimulation
complex figure, group and significant worsening in Digit
recall span forward only in unstructured cog-
Praxis nitive stimulation group;
Rey–Osterrieth – Improvement in Digit cancellation only
complex figure, copy in tele, lexical-semantic stimulation
Attention group;
(continued)
5
6

Table 2. Continued

Patients’ characteristics
Intervention
Number and type Cognitive global description (total
of patients (proto- Age assessment number of sessions, Follow-
Study col design) Mean (SD) Mean (SD) session duration) up Outcome measures Results

Digit cancellation – 86% of the patients were highly satisfied


TMT part A and part by the telerehabilitation treatment.
B
Ad hoc satisfaction
questionnaire
Savage et al. 5 patients with Telerehabilitation ACE-R: Telerehabilitation No Language Improvement in naming trained pictures
(2014)44 svPPA (case group: 61.8 (5.6) Telerehabilitation Individualized Picture Naming task in all patients;
series pre–post; group: 50.0 (11.9) word training Video description – Improvement in Video description
all patients program (40 ses- task task in four out of five patients;
included in the sions over 8 Household request – Improvement in Household request
telerehabilitation weeks, 30 task task in two out of five patients;
group) minutes) Word–picture – Improvement in Word–picture
matching task matching task in two out of five
patients;
– Improvements were not found for
untrained items.
Vermeij et al. 18 patients with Patients with aMCI: MMSE: Telerehabilitation 3 months Memory – Improvement in Digit span forward in
(2016)42 aMCI and 23 68.4 (6.3); Patients with working memory AVLT (immediate healthy older adults;
healthy older Healthy older aMCI: 27.1 (2.2); training (25 ses- recall) – Improvement in Digit span backward in
adults (pilot adults: 70.1 (5.4) Healthy older sions over 5 AVLT (delayed recall) patients with aMCI;
study; both adults: 29.3 (1.0) weeks, 45 AVLT (recognition)m – Improvement in Spatial span backward
patients with minutes) Digit span forward in both groups;
aMCI and healthy Digit span backward – Improvement in RFFT in both groups;
older adults Spatial span forward Gains maintained at follow-up in:
included in the Spatial span back- – Digit span forward in healthy older
telerehabilitation ward adults;
group) Executive functions – Digit span backward in patients with
Stroop colour–word aMCI;
task – Spatial span backward in both groups;
RFFT RFFT only in healthy older adults.
Cognitive com-
plaints
CFQ
Meyer et al. 17 patients with avPPA: MMSE: Tele individualized No Language Overall, the three telerehabilitation par-
(2016)43 PPA (case series Telerehabilitation: avPPA: phonologic or – Oral confrontation ticipants and the FTF group did not
pre-post): 5 48; FTF group: Telerehabilitation: orthographic naming of treated differ significantly in language outcomes
patients with 67.8 (9.2); 24; FTF group: treatment vs. FTF and untreated with regard to trained items. In par-
avPPA (1 svPPA: 27.8 (1.5); phonologic or. items; ticular:
Journal of Telemedicine and Telecare 0(0)

(continued)
Table 2. Continued

Patients’ characteristics
Intervention
Cotelli et al.

Number and type Cognitive global description (total


of patients (proto- Age assessment number of sessions, Follow-
Study col design) Mean (SD) Mean (SD) session duration) up Outcome measures Results

telerehabilitation; Telerehabilitation: svPPA: orthographic – Written confron- avPPA


4 FTF); 4 patients 68; FTF group: Telerehabilitation: treatment tation naming of – Improvement in oral confrontation
with svPPA (1 63.7 (6.4) 12; FTF group: (68 sessions over treated and naming after tele phonologic and
telerehabilitation; lvPPA: 26.5 (0.7); 24 weeks) untreated items; orthographic treatment;
3 FTF); 8 patients Telerehabilitation: lvPPA: Naming during – Improvement of Naming during scene
with lvPPA (1 tel- 69; FTF group: Telerehabilitation: scene description description only after tele phonologic
erehabilitation; 7 71.4 (7.7) 18; FTF group: of treated and treatment.
FTF) 23.4 (4.0) untreated items. Comparison of telerehabilitation vs.
FTF treatment:
– Telerehabilitation had treatment
effects that were significantly larger
than FTF treatment.
svPPA
– Improvement in oral confrontation
naming only after tele phonologic
treatment.
Comparison of telerehabilitation vs.
FTF treatment:
Telerehabilitation and FTF group did
not differ significantly in language out-
come.
lvPPA
– Improvement in oral confrontation
naming after tele phonologic and
orthographic treatment;
– Improvement of written confronta-
tion naming only after tele ortho-
graphic treatment.
Comparison of telerehabilitation vs.
FTF treatment:
Telerehabilitation and FTF group did
not differ significantly in language
outcome.
ACE-R: Addenbrooke’s Cognitive Examination; aMCI: amnestic MCI; avPPA: agrammatic variant PPA; AD: Alzheimer’s disease; C-MMSE: Cantonese Mini-Mental State Examination; C-RBMT: Cantonese Rivermead
Behavioural Memory Test; CFQ: Cognitive Failure Questionnaire; FTF: face-to-face; HDS: Hierarchic Dementia Scale; lvPPA: logopenic variant PPA; MMSE: Mini-Mental State Examination; MCI: mild cognitive
impairment; NA: not available; PPA: primary progressive aphasia; AVLT: Auditory Verbal Learning Test; RFFT: Ruff Figural Fluency test; svPPA: semantic variant PPA; TMT: Trail-Making Test; PPA: primary progressive
aphasia; SD: standard deviation.
7
8 Journal of Telemedicine and Telecare 0(0)

groups. Only spatial construction abilities, tested using a regard to trained items, with effects similar to participants
subtest of the Hierarchic Dementia Scale, showed an assigned to FTF treatment, indicating that telerehabilita-
improvement in the FTF group, which was absent in the tion-based anomia training programs may be a promising
telerehabilitation group. In addition, the authors high- method to provide rehabilitation intervention strategies in
lighted a compliance rate above the 90% in the telereh- patients with PPA.
abilitation group.40
Jelcic et al.41 described the effects of telerehabilitation
in a sample of patients with AD. The authors compared
Methodological quality assessment
the effects of lexical-semantic stimulation (LSS) through Only one out of the five studies was a RCT,40 four were
telecommunication technology (LSS-tele) with in-person quasi-experimental studies: three pretest–posttest
LSS (LSS-direct) and with unstructured cognitive stimu- between-group controlled trials41–43 and one pretest–
lation (UCS) treatment. Twenty-seven patients with AD posttest trial without a control group.44 The latter was
were divided into three groups: seven patients received not included in the analysis of the overall methodological
LSS-tele treatment, 10 received standard LSS-direct inter- quality.
vention and 10 participants underwent UCS. Intervention PEDro scale score results on the four selected studies
treatments consisted of two weekly sessions throughout a showed that only one study is characterized by high meth-
three-month period. The authors reported that the Mini- odological quality.41 The mean PEDro score was 4.3,
Mental State Examination (MMSE) score improved sig- standard deviation (SD) 1.3. (lowest 3, highest 6). First,
nificantly after both LSS-tele and LSS-direct treatments. eligibility criteria were specified in all the four studies.
LSS-tele improved language and attentional abilities, Results show that none of the four studies satisfied con-
while LSS-direct enhanced verbal short-term and long- cealed allocation (criterion 2), blinding of all subjects (cri-
term memory. Moreover, 86% of the patients were terion 4) and blinding of all therapists who administered
highly satisfied by the telerehabilitation treatment. the therapy (criterion 5). The result on criterion 4 and
Cognitive improvement was not observed in any neuro- criterion 5 is not surprising because blinding of partici-
psychological test score after UCS. The authors concluded pants and therapists is impractical in cognitive interven-
that telecommunication technology applied to cognitive tion trials.45 Moreover, only one study (25%) reported the
rehabilitation in AD patients is feasible and may improve random allocation of patients to groups (criterion 1), and
cognitive performance.41 used blinded assessors who measured at least one key out-
A pilot study has investigated the effect of a working come (criterion 6). In two studies (50%) the groups were
memory (WM) telerehabilitation training in healthy older similar at baseline regarding the most important prognos-
adults and in patients with amnestic MCI.42 All partici- tic indicators (criterion 3), whereas in three studies out of
pants completed adaptive WM training with telerehabil- four (75%) all subjects for whom outcome measures were
itation for 25 sessions over a five-week period. The study available received the treatment or control condition as
showed an improvement in verbal and non-verbal short- allocated (criterion 8). The results of between-group stat-
term memory, and these gains were maintained at three istical comparisons are reported (criterion 9) and meas-
months follow-up. ures of variability for at least one key outcome are
Finally, two works explored the effects of language provided (criterion 10). All the included studies reported
training conducted by telerehabilitation in patients with measures of key outcome obtained from more than 85%
PPA.43,44 Savage et al.44 included five patients with of the subjects initially allocated to groups (criterion 7).
svPPA in a two-month language telerehabilitation train- Regarding the quality assessment of the studies, limited
ing program. Overall, participants showed clear gains in disagreement (less than 6%) was reached between the two
language production (naming the trained pictures and reviewers, who independently assessed them, concluding
video description). In addition, milder patients showed that the quality assessment tool did not allow for a great
an improvement in comprehension tasks and severe amount of misjudgement. Table 1 highlights the scores for
patients showed improvements in matching trained each criterion using the PEDro scale.
words to pictures. As expected, improvements were not
found for untrained items.
A recent case-series pre–post study has compared the
Effect sizes (ES) analysis
efficacy of a conventional language rehabilitation with a The ES for the telerehabilitation intervention (post–pre)
telerehabilitation approach to anomia treatment in was evaluated for outcomes of four studies,40–42,44 con-
patients with PPA.43 Following a baseline evaluation of sidering both Cohen’s d formulas for correlated and inde-
language and cognition, a phonological and orthographic pendent samples. One study43 was not included in this
treatment was administered to 17 PPA participants over analysis because it included single patients in the telereh-
the course of six months. Fourteen participants were abilitation intervention. As reported in Table 3, although
assigned to FTF treatment, while three patients received the ES computed for paired samples (with r ¼ 0.5) were
the treatment via telerehabilitation. The three patients higher than those for independent samples (with r ¼ 0), the
that received telerehabilitation treatment obtained a sig- ES in which CIs did not include zero were the same in
nificant improvement in language performances with both paired and independent computations. Thus, we
Cotelli et al. 9

Table 3. Effect sizes for telerehabilitation intervention effects.

Telerehabilitation intervention Effect size

Baseline Post-treatment Paired 95% CI Paired 95% CI


Cohen’s Cohen’s
d with d with
Study Outcome Mean SD Mean SD r ¼ 0.5 Lower Upper r¼0 Lower Upper

Poon et al., Global cognition


200540 C-MMSE 18.73 2.15 21.91 2.95 1.26 0.35 2.18 1.23 0.32 2.14
Global cognition
HDS 154.82 24.99 169.27 26.06 0.59 –0.26 1.45 0.57 –0.29 1.42
Memory
C-RBMT 5.64 2.29 8.81 3.12 1.19 0.28 2.09 1.16 0.26 2.06
Jelcic et al., Global cognition
201441 MMSE 23.7 3 25.7 2 0.82 –0.27 1.91 0.78 –0.30 1.87
Language
Verbal Naming Test 32.4 5 35.3 2 0.72 –0.36 1.80 0.76 –0.32 1.85
Phonemic verbal fluency 14.3 7 18.1 6 0.63 –0.45 1.70 0.58 –0.49 1.65
Semantic verbal fluency 17.9 5 20.4 3 0.62 –0.45 1.69 0.61 –0.47 1.68
Memory
Story immediate recall 7.4 2 5.7 3 –0.69 –1.77 0.38 –0.67 –1.74 0.41
Story delayed recall 5.4 3 6.3 3 0.32 –0.73 1.38 0.30 –0.75 1.35
AVLT immediate recall 25.3 9 20.9 4 –0.61 –1.68 0.46 –0.63 –1.71 0.44
AVLT delayed recall 3.3 3 3.4 4 0.03 –1.02 1.08 0.03 –1.02 1.08
Digit span forward 5 1 5 1 0.00 –1.05 1.05 0.00 –1.05 1.05
Digit span backward 3 1 3 1 0.00 –1.05 1.05 0.00 –1.05 1.05
Rey’s Complex figure recall 4.6 3 5 5 0.10 –0.95 1.15 0.10 –0.95 1.15
Praxis
Rey’s Complex figure copy 20 11 18.6 12 –0.13 –1.18 0.92 –0.12 –1.17 0.93
Attention
Digit cancellation 35.6 10 38 12 0.23 –0.82 1.28 0.22 –0.83 1.27
Savage et al., Language
201444 treated words (%)
Picture Naming Task 29.84 19.37 79.31 20.14 2.80 1.05 4.54 2.50 0.85 4.16
Video Description Task 32.85 19.93 61.83 11.85 1.87 0.38 3.35 1.77 0.31 3.23
HouseHold Request Task 39.66 19.95 64.26 26.9 1.14 –0.20 2.47 1.04 –0.28 2.36
Word–Picture Matching Task 54.42 24.63 78.37 13.15 1.25 –0.10 2.61 1.21 –0.14 2.56
Untreated words (%)
Picture Naming Task 35.57 14.15 40.63 22.01 0.29 –0.95 1.54 0.27 –0.97 1.52
Video Description Task 36.69 27.19 34.71 26.53 –0.08 –1.32 1.16 –0.07 –1.31 1.17
HouseHold Request Task 36.28 34.05 51.85 34.6 0.51 –0.75 1.77 0.45 –0.80 1.71
Word–Picture Matching Task 58.78 23.56 69 23.79 0.48 –0.77 1.74 0.43 –0.82 1.69
Vermeij et al., Memory 0.00
201642 AVLT immediate recall 33.6 7.5 32.1 7.2 –0.21 –0.86 0.45 –0.20 –0.86 0.45
AVLT delayed recall 3.9 3.4 3.6 3 –0.10 –0.75 0.56 –0.09 –0.75 0.56
AVLT recognition 25.2 3.7 24.8 3.8 –0.11 –0.76 0.54 –0.11 –0.76 0.55
Digit span forward 8.5 1.5 9.2 1.8 0.43 –0.23 1.09 0.42 –0.24 1.08
Digit span backward 6.2 1.6 7.3 1.9 0.64 –0.03 1.31 0.63 –0.04 1.30
Digit span – total score 14.7 2.8 16.5 3.2 0.61 –0.06 1.28 0.60 –0.07 1.27
Spatial span forward 7.2 2.1 7.8 1.7 0.32 –0.34 0.98 0.31 –0.34 0.97
Spatial span backward 6.8 2.1 7.4 2.3 0.28 –0.38 0.94 0.27 –0.38 0.93
Spatial span – total score 13.9 3.9 15.3 3.6 0.38 –0.28 1.04 0.37 –0.29 1.03
(continued)
10 Journal of Telemedicine and Telecare 0(0)

Table 3. Continued

Telerehabilitation intervention Effect size

Baseline Post-treatment Paired 95% CI Paired 95% CI


Cohen’s Cohen’s
d with d with
Study Outcome Mean SD Mean SD r ¼ 0.5 Lower Upper r¼0 Lower Upper

Executive functions
Stroop Colour–Word Task 66.4 30 60.9 24.6 –0.20 –0.86 0.45 –0.20 –0.86 0.45
RFFT – total patterns 59.3 19.2 64.4 18.8 0.28 –0.38 0.93 0.27 –0.39 0.92
RFFT – total errors 7.4 5.6 6.9 6.5 –0.08 –0.74 0.57 –0.08 –0.74 0.57
RFFT – error ratio 0.13 0.09 0.1 0.08 –0.36 –1.02 0.30 –0.35 –1.01 0.31
Cognitive complaints 0.00 –0.65 0.65
CFQ – total score 90.5 10.6 89.7 10.6 –0.08 –0.73 0.58 –0.08 –0.73 0.58
Effect sizes are calculated on the difference between post-treatment and pre-treatment scores. Bold scores highlight significant effect sizes and significant
confidence intervals.
C-MMSE: Cantonese Mini-Mental State Examination; C-RBMT: Cantonese Rivermead Behavioural Memory Test; CFQ: Cognitive Failure Questionnaire; CI:
confidence interval; HDS: Hierarchic Dementia Scale; MMSE: Mini-Mental State Examination; AVLT: Auditory Verbal Learning Test; RFFT: Ruff Figural Fluency
test; SD: standard deviation.

Figure 2. Forest plot of Cohen’s d related to the effect of telerehabilitation intervention on memory domain.

conservatively considered the ES for independent samples. (ES ¼ 0.76) in Jelcic et al.41 and for the HouseHold
Specifically, Cantonese Mini-Mental State Examination Request Task (ES ¼ 1.04) and Word–Picture Matching
(ES ¼ 1.23, 95% CI [0.32, 2.14]), Cantonese Rivermead Task (ES ¼ 1.21) in Savage et al.44 In particular, telereh-
Behavioural Memory Test (ES ¼ 1.16, 95% CI [0.26, abilitation intervention in patients with MCI or AD
2.06]), tests measured in the Poon et al.40 study, as well appears to induce improvements in the memory domain
as Picture Naming Task (ES ¼ 2.50, 95% CI [0.85, 4.16]) (see Figure 2).
and Video Description Task (1.77, 95% CI [0.31, 3.23]), With regards to the comparison between telerehabilita-
tests measured in Savage et al.,44 showed very large sig- tion and FTF treatment (Table 4), we found very large
nificant ES. Moreover, ES which were quite large (i.e. and significant ES for story immediate recall (1.28, 95%
ES > 0.75) but not significantly different from zero were CI [2.43, 0.13]) and Auditory Verbal Learning Test
found for MMSE (ES ¼ 0.78) and the Verbal Naming Test immediate recall (ES ¼ 1.04, 95% CI [2.16, 0.08])
Table 4. Effect sizes for the comparison between telerehabilitation intervention effects and the effects of control interventions.

Comparison with Comparison with


face-to-face unstructured
Cotelli et al.

intervention cognitive stimulation

Unstructured
Telerehabilitation Face-to-face cognitive
intervention intervention stimulation Effect size Effect size

Baseline Post-treatment Baseline Post-treatment Baseline Post-treatment 95% CI 95% CI

Study Outcome Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Cohen’s d Lower Upper Cohen’s d Lower Upper

Poon et al., Global cognition


200540 C-MMSE 18.73 2.15 21.91 2.95 18.27 2.41 22.09 3.53 –0.22 –1.06 0.62
Global cognition
HDS 154.82 24.99 169.27 26.06 156.09 17.1 170.64 15.95 0.00 –0.84 0.83
Memory
C-RBMT 5.64 2.29 8.81 3.12 7.45 2.16 10.36 2.73 0.10 –0.74 0.93
Jelcic et al., Global cognition
201441 MMSE 23.7 3 25.7 2 24.9 2 26.9 2 24.8 3 24.1 4 0.00 –1.05 1.05 0.85 –0.24 1.95
Language
Verbal Naming Test 32.4 5 35.3 2 35.1 2 36.8 1 33.2 3 32.8 4 0.36 –0.69 1.42 0.83 –0.27 1.92
Phonemic verbal fluency 14.3 7 18.1 6 25.4 14 26.6 11 20.6 11 22.2 11 0.26 –0.80 1.31 0.24 –0.81 1.29
Semantic verbal fluency 17.9 5 20.4 3 26.9 7 27.5 8 26.1 8 26.5 9 0.31 –0.75 1.36 0.31 –0.74 1.36
Memory
Story immediate recall 7.4 2 5.7 3 4.1 2 5.4 2 3.6 1 3.9 2 –1.28 –2.43 –0.13 –0.89 –1.99 0.20
Story delayed recall 5.4 3 6.3 3 4.4 3 6.4 1 4.4 2 3.8 3 –0.39 –1.45 0.67 0.53 –0.54 1.60
AVLT immediate recall 25.3 9 20.9 4 25.5 4 29.7 10 23.8 6 23.5 8 –1.04 –2.16 0.08 –0.55 –1.61 0.52
AVLT delayed recall 3.3 3 3.4 4 4.5 2 5.9 3 3.7 3 2.4 2 –0.41 –1.47 0.65 0.44 –0.62 1.50
Digit span forward 5 1 5 1 5 1 5.4 0.5 5.2 1 4.8 1 –0.43 –1.49 0.63 0.40 –0.66 1.46
Digit span backward 3 1 3 1 3.2 1 3.6 1 2.9 1 2.9 1 –0.40 –1.46 0.66 0.00 –1.05 1.05
Rey’s Complex figure recall 4.6 3 5 5 8.5 8 9.1 9 7.6 5 6.5 6 –0.03 –1.08 1.02 0.30 –0.75 1.35
Praxis
Rey’s Complex figure copy 20 11 18.6 12 23.9 10 25.8 7 28.8 11 25.6 14 –0.32 –1.37 0.73 0.15 –0.90 1.20
Attention
Digit cancellation 35.6 10 38 12 41.1 11 40.5 11 36.4 8 36 13 0.27 –0.78 1.32 0.25 –0.80 1.30
Effect sizes are calculated on the difference between telerehabilitation and other intervention effects. Bold scores highlight significant effect sizes and significant confidence intervals.
C-MMSE: Cantonese Mini-Mental State Examination; C-RBMT: Cantonese Rivermead Behavioural Memory Test; CI: confidence interval; HDS: Hierarchic Dementia Scale; MMSE: Mini-Mental State Examination; AVLT:
Auditory Verbal Learning Test; SD: standard deviation.
11
12 Journal of Telemedicine and Telecare 0(0)

tests evaluated in Jelcic et al.,41 both showing a better cognitive telerehabilitation interventions with a conven-
effect of FTF treatment with respect to telerehabilitation. tional FTF cognitive rehabilitation in patients with
Finally, by comparing telerehabilitation with UCS,41 no MCI, AD and FTD. A crucial limitation of this review,
significant ESs were found, although large effects were which made it difficult to draw consistent conclusions, was
found for MMSE (ES ¼ 0.85) and Verbal Naming Test the large variation in methodology among studies. The
(ES ¼ 0.83). Telerehabilitation had a better effect than variability of exercise training features (frequency, inten-
UCS; whereas in story immediate recall (ES ¼ –0.89) the sity, time, type) and the heterogeneity of the cognitive tests
former treatment had lower effects than the latter. used to measure different cognitive domains are possible
reasons for large variations in the cognitive benefits
between studies. All the included studies had a very
Discussion
small sample size, with an average of less than 5.3 (SD
To our knowledge, this review is the first attempt to sys- 8.8) individuals. Many of the included studies might have
tematically evaluate current evidence regarding the efficacy lacked the power to detect differences between the inter-
of cognitive telerehabilitation interventions compared with vention and control group. Finally, some unpublished stu-
a conventional FTF cognitive rehabilitation in patients dies may not have been identified through the literature
with neurodegenerative diseases. Only five studies of cog- search, and a publication bias cannot be excluded.
nitive training via telerehabilitation were identified for
inclusion in the review. Only one RCT of cognitive telereh- Declaration of Conflicting Interests
abilitation was identified,40 showing encouraging results The authors declared no potential conflicts of interest with
and indicating that telerehabilitation is likely to provide respect to the research, authorship, and/or publication of this
cognitive improvement for patients with mild dementia or article.
MCI with comparable efficacy to FTF intervention.
Overall, we found that cognitive telerehabilitation has Funding
comparable effects to conventional rehabilitation in The authors disclosed receipt of the following financial support
improving cognitive abilities in neurodegenerative disease. for the research, authorship, and/or publication of this article:
ES analysis evidenced positive effects of telerehabilitation this work was supported by Italian Ministry of Health (Ricerca
interventions, generally comparable with FTF interven- Corrente).
tion effects.
We highlighted that available evidence regarding cog- References
nitive telerehabilitation remains limited, and the quality of 1. Fratiglioni L, De Ronchi D and Aguero-Torres H.
the clinical studies designs needs to be improved, particu- Worldwide prevalence and incidence of dementia. Drugs
larly in relation to the blinding of assessors and conceal- Aging 1999; 15: 365–375.
ment of allocation. Trial reports indicate that some gains 2. Jorm AF and Jolley D. The incidence of dementia: a meta-
resulting from intervention may not be captured ade- analysis. Neurology 1998; 51: 728–733.
quately by available standardized outcome measures. 3. Neary D and Snowden J. Fronto-temporal dementia: nos-
The available evidence is insufficient to draw conclusions ology, neuropsychology, and neuropathology. Brain Cogn
about the effects of the intervention on cognition, health- 1996; 31: 176–187.
related quality of life or participant satisfaction with the 4. Rosen HJ, Gorno-Tempini ML, Goldman WP, et al.
Patterns of brain atrophy in frontotemporal dementia and
intervention. A serious limitation is the lack of any informa-
semantic dementia. Neurology 2002; 58: 198–208.
tion about possible functional effects of the treatment (trans- 5. Rascovsky K, Hodges JR, Knopman D, et al. Sensitivity of
fer to activities of daily living) or none of the five selected revised diagnostic criteria for the behavioural variant of
studies evaluated the cost-effectiveness of telerehabilitation frontotemporal dementia. Brain 2011; 134: 2456–2477.
and no study reported the occurrence of adverse events. 6. Grossman M. Primary progressive aphasia: clinicopatholo-
Further, well-designed studies of cognitive telerehabilitation gical correlations. Nat Rev Neurol 2010; 6: 88–97.
are necessary. While the preliminary evidence suggests that 7. Gorno-Tempini ML, Hillis AE, Weintraub S, et al.
cognitive telerehabilitation may have similar effects to FTF Classification of primary progressive aphasia and its vari-
cognitive interventions in improving cognitive abilities in ants. Neurology 2011; 76: 1006–1014.
neurodegenerative diseases, standardization of methodo- 8. Grossman M. The non-fluent/agrammatic variant of pri-
logical aspects of the studies is required to result in more mary progressive aphasia. Lancet Neurol 2012; 11: 545–555.
homogenous data and to determine the optimal type and 9. Rohrer JD, Knight WD, Warren JE, et al. Word-finding
dose of cognitive telerehabilitation. difficulty: a clinical analysis of the progressive aphasias.
Brain 2008; 131: 8–38.
10. Janoutova J, Sery O, Hosak L, et al. Is mild cognitive
Strengths and limitations impairment a precursor of Alzheimer’s disease? Short
review. Cent Eur J Public Health 2015; 23: 365–367.
There are inherent strengths and limitations of this sys- 11. Kidd PM. Alzheimer’s disease, amnestic mild cognitive
tematic review. First, due to the heterogeneity of outcome impairment, and age-associated memory impairment: cur-
measures it was not possible to perform a meta-analysis. rent understanding and progress toward integrative preven-
Moreover, only one RCT compared the efficacy of tion. Altern Med Rev 2008; 13: 85–115.
Cotelli et al. 13

12. Clare L, Bayer A, Burns A, et al. Goal-oriented cognitive 29. Mashima PA and Doarn CR. Overview of telehealth activ-
rehabilitation in early-stage dementia: study protocol for a ities in speech-language pathology. Telemed J E Health
multi-centre single-blind randomised controlled trial 2008; 14: 1101–1117.
(GREAT). Trials 2013; 14: 152. 30. Peretti A, Amenta F, Tayebati SK, et al. Telerehabilitation:
13. Clare L, van Paasschen J, Evans SJ, et al. Goal-oriented review of the state-of-the-art and areas of application. JMIR
cognitive rehabilitation for an individual with Mild Rehabil Assist Technol 2017; 4: e7.
Cognitive Impairment: behavioural and neuroimaging out- 31. Zampolini M, Todeschini E, Bernabeu Guitart M, et al.
comes. Neurocase 2009; 15: 318–331. Tele-rehabilitation: present and future. Ann Ist Super
14. Bahar-Fuchs A, Clare L and Woods B. Cognitive training Sanita 2008; 44: 125–134.
and cognitive rehabilitation for mild to moderate 32. Kairy D, Lehoux P, Vincent C, et al. A systematic review of
Alzheimer’s disease and vascular dementia. Cochrane clinical outcomes, clinical process, healthcare utilization and
Database Syst Rev 2013; 6: Cd003260. costs associated with telerehabilitation. Disabil Rehabil 2009;
15. Cotelli M, Calabria M and Zanetti O. Cognitive rehabilitation 31: 427–447.
in Alzheimer’s disease. Aging Clin Exp Res 2006; 18: 141–143. 33. Rosen MJ. Telerehabilitation. Telemed J E Health 2004; 10:
16. Gates NJ and Sachdev P. Is cognitive training an effective 115–117.
treatment for preclinical and early Alzheimer’s disease? 34. Cherney LR and van Vuuren S. Telerehabilitation, virtual
J Alzheimers Dis 2014; 42(Suppl 4): S551–S559. therapists, and acquired neurologic speech and language dis-
17. Hong YJ, Jang EH, Hwang J, et al. The efficacy of cognitive orders. Semin Speech Lang 2012; 33: 243–257.
intervention programs for mild cognitive impairment: a sys- 35. Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting
tematic review. Curr Alzheimer Res 2015; 12: 527–542. items for systematic reviews and meta-analyses: the
18. Kortte KB and Rogalski EJ. Behavioural interventions for PRISMA statement. BMJ 2009; 339: b2535.
enhancing life participation in behavioural variant fronto- 36. Maher CG, Sherrington C, Herbert RD, et al. Reliability of
temporal dementia and primary progressive aphasia. Int Rev the PEDro scale for rating quality of randomized controlled
Psychiatry 2013; 25: 237–245. trials. Phys Ther 2003; 83: 713–721.
19. Cappa SF, Benke T, Clarke S, et al. EFNS guidelines on 37. Cohen J. Statistical power analysis for the behavioral sci-
cognitive rehabilitation: report of an EFNS task force. Eur J ences. Hilsdale, NJ: Lawrence Earlbaum Associates, 1988,
Neurol 2003; 10: 11–23. p.2.
20. Cappa SF, Benke T, Clarke S, et al. EFNS guidelines on 38. Dunlap WP, Cortina JM, Vaslow JB, et al. Meta-analysis of
cognitive rehabilitation: report of an EFNS task force. Eur J experiments with matched groups or repeated measures
Neurol 2005; 12: 665–680. designs. Psychol Methods 1996; 1(2): 170–177.
21. Bahar-Fuchs A, Clare L and Woods B. Cognitive training 39. Morris SB and DeShon RP. Combining effect size estimates
and cognitive rehabilitation for persons with mild to mod- in meta-analysis with repeated measures and independent-
erate dementia of the Alzheimer’s or vascular type: a review. groups designs. Psychol Methods 2002; 7: 105–125.
Alzheimers Res Ther 2013; 5: 35. 40. Poon P, Hui E, Dai D, et al. Cognitive intervention for
22. Clare L. Rehabilitation for people living with dementia: a community-dwelling older persons with memory problems:
practical framework of positive support. PLoS Med 2017; telemedicine versus face-to-face treatment. Int J Geriatr
14: e1002245. Psychiatry 2005; 20: 285–286.
23. Clare L, Linden DE, Woods RT, et al. Goal-oriented cog- 41. Jelcic N, Agostini M, Meneghello F, et al. Feasibility
nitive rehabilitation for people with early-stage Alzheimer and efficacy of cognitive telerehabilitation in early
disease: a single-blind randomized controlled trial of clinical Alzheimer’s disease: a pilot study. Clin Interv Aging 2014;
efficacy. Am J Geriatr Psychiatry 2010; 18: 928–939. 9: 1605–1611.
24. Corbett A, Owen A, Hampshire A, et al. The effect of an 42. Vermeij A, Claassen JA, Dautzenberg PL, et al. Transfer
online cognitive training package in healthy older adults: an and maintenance effects of online working-memory training
online randomized controlled trial. J Am Med Dir Assoc in normal ageing and mild cognitive impairment.
2015; 16: 990–997. Neuropsychol Rehabil 2016; 26: 783–809.
25. Brennan D, Tindall L, Theodoros D, et al. A blueprint for 43. Meyer AM, Getz HR, Brennan DM, et al. Telerehabilitation
telerehabilitation guidelines. Int J Telerehabil 2010; 2: 31–34. of anomia in primary progressive aphasia. Aphasiology 2016;
26. Brennan D, Georgeadis A and Baron C. Telerehabilitation 30: 483–507.
tools for the provision of remote speech-language treatment. 44. Savage SA, Piguet O and Hodges JR. Giving words new life:
Top Stroke Rehabil 2002; 8: 71–78. generalization of word retraining outcomes in semantic
27. Brennan DM, Mawson S and Brownsell S. dementia. J Alzheimers Dis 2014; 40: 309–317.
Telerehabilitation: enabling the remote delivery of health- 45. Lampit A, Hallock H and Valenzuela M. Computerized cog-
care, rehabilitation, and self management. Stud Health nitive training in cognitively healthy older adults: a system-
Technol Inform 2009; 145: 231–248. atic review and meta-analysis of effect modifiers. PLoS Med
28. Hailey D, Roine R, Ohinmaa A, et al. Evidence of benefit 2014; 11: e1001756.
from telerehabilitation in routine care: a systematic review.
J Telemed Telecare 2011; 17: 281–287.

You might also like