Condition category
Nervous System Diseases
Date applied
27/05/2018
Date assigned
12/06/2018
Last edited
07/06/2018
Prospective/Retrospective
Retrospectively registered
Overall trial status
Completed
Recruitment status
No longer recruiting

Plain English Summary

Background and study aims
About a third of people who have a stroke, have aphasia. Aphasia is a language disorder that can affect speaking, understanding what people say, reading and writing. Most people with aphasia have difficulty finding the words they want to say. In this study we looked at whether a specific speech and language therapy could improve word-finding for people with aphasia. The therapy is called Elaborated Semantic Feature Analysis (ESFA). We compared a group of people who had ESFA to a group of people on a waiting list. We also looked at whether there were any differences in outcomes if people had ESFA one-to-one with a speech and language therapist versus if they had ESFA in a combination of one-to-one sessions and group therapy.

Who can participate?
People with aphasia due to stroke at least 4 months after they had their stroke. Participants had to be over 18 years old and have no significant cognitive problems.

What does the study involve?
By the end of the study, all participants had 3 hours of ESFA therapy per week for 12 weeks (36 hours in total). Those who had one-to-one ESFA had three 1-hour sessions per week with a speech and language therapist. Those that had combination ESFA had two 45-min one-to-one sessions with the speech and language therapist and one 90-min group session with other people with aphasia and the speech and language therapist.

What are the possible benefits and risks of participating?
By taking part in this study, people receive speech and language therapy which can help them improve their word finding skills. They may also find it easier to communicate with other people and they may experience improved feelings of quality of life. There are no known risks in taking part in this study.

Where is the study run from?
The lead centre of this study is Eginitio Hospital in Athens, Greece. There are also another six hospitals/rehabilitation centres in Athens and the University General Hospital of Patras taking part.

When is the study starting and how long is it expected to run for?
The study started in January 2012 and finished in November 2015.

How long will the trial be recruiting participants for?
The study recruited participants between February 2013 and March 2015. The study was funded by the European Union (European Social Fund – ESF) and Greek national funds through the Operational Program "Education and Lifelong Learning" of the National Strategic Reference Framework (NSRF).

Who is the main contact?
Professor Spyridoula Varlokosta, svarlokosta@phil.uoa.gr

Trial website

http://thales-aphasia.phil.uoa.gr/

Contact information

Type

Public

Primary contact

Prof Spyridoula Varlokosta

ORCID ID

Contact details

Department of Linguistics
Faculty of Philology
National and Kapodistrian University of Athens
Panepistimioupoli
15784 Zografou
Athens
Greece
Athens
15784
Greece
+30 210 7277642
svarlokosta@phil.uoa.gr

Type

Scientific

Additional contact

Prof Katerina Hilari

ORCID ID

http://orcid.org/0000-0003-2091-4849

Contact details

Division of Language and Communication Science
City
University of London
Northampton Square
London
EC1V 0HB
United Kingdom
+442070404660
k.hilari@city.ac.uk

Additional identifiers

EudraCT number

ClinicalTrials.gov number

Protocol/serial number

70/3/11672

Study information

Scientific title

Efficacy of elaborated semantic features analysis in aphasia: a quasi-randomised controlled trial

Acronym

Thales Aphasia: SLT

Study hypothesis

1. Evaluate the efficacy of Elaborated Semantic Features Analysis (ESFA) for people with aphasia on different domains of the WHO ICF (International Classification of Functioning, Disability and Health) framework, including quality of life, as compared to a delayed-treatment control group. It was hypothesised that the therapy group will have improved language skills, while the delayed-treatment control group will not. It was also hypothesised that ESFA, although specifically targeting the underlying language impairment, could lead to secondary gains in other levels of the WHO ICF model, such as communication, and quality of life.
2. Compare and contrast the relative efficacy of ESFA therapy on different domains of the WHO ICF framework, and quality of life, as delivered in two different approaches - direct (individual) and indirect combination therapy (individual and group). It was hypothesised that direct therapy (individual therapy) will lead to greater benefits on participants’ language skills (naming), while indirect therapy (combination therapy) will lead to greater benefits on functional communication, i.e. the ability of people to get their message across, using whatever means they can. Combination therapy (individual and group therapy) may potentially have a greater effect on participants’ well-being and life quality due to the reported psychosocial benefits of groups therapy.

Ethics approval

Ethical approval was obtained in both Greece and the United Kingdom. In Greece, the project was evaluated by two research ethics committees: The University Hospital of Patras (approved 25/02/2013, 42/19.02.2013), for participants recruited in Achaia, and the University of Athens Eginitio Hospital (approved 02/06/2013, 325/16-01-13) for participants recruited in Attica. In the UK, the project was approved by the Division of Language and Communication Science’s Proportionate Review Committee of the School of Health Sciences, City, University of London (approved 21/08/2013, PhD/12-13/17).

Study design

Quasi-randomized single-blind controlled trial

Primary study design

Interventional

Secondary study design

Randomised controlled trial

Trial setting

Home

Trial type

Treatment

Patient information sheet

Condition

Aphasia due to stroke

Intervention

Participants were randomised based on the order of their enrolment in the study. It is quasi-randomised for two main reasons. Firstly participants were randomized after enrolment to the overall Thales project and before eligibility for speech and language therapy was checked. This resulted in participants being excluded for not meeting eligibility criteria after randomization and therefore uneven numbers in the groups. Secondly participants were randomized to one of three groups: direct therapy, combination therapy, delayed treatment. Therapies offered were either Elaborated Semantic Features Analysis (ESFA) or mapping therapy. This study reports on those who were allocated to ESFA therapy. The decision on whether a person with aphasia would receive ESFA or mapping therapy was based on their language deficits according to the Boston Diagnostic Aphasia Examination.

The intervention tested was ESFA, which has been previously described in detail and has good evidence of treatment fidelity (Kladouchou et al., 2017). In summary, participants received either 36 h of individual therapy (three 1-h sessions per week for 12 weeks) or 36 h of a combination of individual and group therapy (two 45-min individual therapy sessions and one 1.5 h group session per week for 12 weeks). The sessions took place mainly in the participants’ home and some in hospital settings. In therapy, the client chose a picture from a stimuli set and the therapist asked them to name it. Then, presenting a semantic feature chart (as in Boyle 2004), the therapist prompted the client to think of and say words related semantically with the target word (semantic features). The chart included six categories: superordinate category, use, action, physical properties, location and association. To elicit features, the therapist asked questions or provided the client with sentence completion cues while prompting them to write down the features generated. If needed, the therapist used an alphabet board to help clients write; and if they were unable to write, the therapist filled in the chart. Then the client was prompted to produce phrases with the target word and each of its features. In group therapy sessions, the same procedure was followed with participants asked in turn to complete the chart and produce phrases. In time, the therapist gave participants the opportunity to interact and provide appropriate cues to each other. The therapist controlled turn taking to ensure individuals got similar amounts of exposure to targets and cues, whilst being mindful of disturbing peer-to-peer interactions as little as necessary.

To select treatment stimuli, each participant completed an oral confrontation-naming task of the 260 Snodgrass and Vanderwart pictures –colour version (Rossion & Pourtois, 2004) three times before starting therapy. The pictures were presented in a random order to each participant, without any cueing or feedback. It took approximately 60 minutes to administer the full set of pictures, using a computerized task, and participants were given a maximum of 13 seconds to respond to each picture. The pictures that a participant failed to name on at least two trials were selected as potential treatment stimuli. Not all selected treatment items were used during the therapy procedure. Each participant was trained in a subset that was dependent on participant’s success on the probes that were taken during the therapy.

The delayed treatment control group had no speech and language therapy while the intervention groups were having therapy.

Intervention type

Behavioural

Phase

Drug names

Primary outcome measure

Oral confrontation-naming task of the 260 Snodgrass and Vanderwart pictures – colour version.
Participants on the therapy approaches (direct group, combination group) were assessed twice before therapy (double baseline, weeks 1 and 6), post-therapy (week 19), and 3-months later (follow-up). Participants on the delayed treatment control group were assessed three times before therapy (weeks 1, 6 and19) and then were randomly allocated to one of the two approaches for ESFA treatment and were reassessed after the 12-week treatment (post-therapy) and 3 months later (follow-up).

To test (a) the efficacy of ESFA therapy versus no therapy mixed within-between ANOVAs were used with group as the between variable (2 groups: ESFA versus control) and time as the within variable (3 levels: weeks 1, 6, 19). To test (b) the relative efficacy of direct versus combination ESFA, mixed within-between ANOVAs were used with group as the between variable (2 groups: direct versus combination ESFA) and time as the within variable (4 levels: two baselines, post-therapy and follow-up).

Secondary outcome measures

1. Greek Boston Naming Test (Simos, Kasselimis & Mouzaki, 2011)
2. Functional communication skills assessed using American Speech and Hearing Association Functional Assessment of Communication Skills for Adults (ASHA FACS) (Frattali et al., Holland, Thompson, Wohl, & Ferketic, 1995), which was completed by the partner / main carer of the participant with aphasia
3. Discourse scores from the Cookie Theft Picture Description of the BDAE (BDAE; Goodglass & Kaplan, 1983)
4. Emotional distress assessed using the Greek version of the General Health Questionnaire-12, (GHQ-12, Garyfallos et al., 2001)
5. Quality of life assessed using the Greek version of the Stroke and Aphasia Quality of Life Scale-39g scale (SAQOL-39g, Kartsona & Hilari, 2007; Efstratiadou et al., 2012)
6. Health status assessed using the Greek version of EQ-5D (Kontodimopoulos, 2008).

Secondary outcomes were assessed at the same time points as the primary outcome. The same analyses were carried out for the secondary outcome measures as for the primary outcome.

Overall trial start date

01/01/2012

Overall trial end date

30/11/2015

Reason abandoned (if study stopped)

Eligibility

Participant inclusion criteria

1. Aphasia following stroke, as reported by their referring clinician
2. >4 months post stroke and medically stable
3. Greek native speakers
4. Aged over 18 years
5. No significant cognitive decline (scored ≥32 out of 38 on Brief Cognitive Screening Test [Economou & Routsis, 2015], a test specifically developed for people with aphasia)

Participant type

Patient

Age group

Adult

Gender

Both

Target number of participants

A total sample of 78 participants gave 80% power; and a sample size of 92 gave 85% power. We aimed to recruit 96 participants to allow for attrition, which is common in stroke studies.

Participant exclusion criteria

1. Not living at home prior to stroke
2. Known history of mental health problems and/or cognitive decline prior to stroke
3. History of other neurological or psychiatric problems
4. Received other speech language therapy during this research

Recruitment start date

22/02/2013

Recruitment end date

16/03/2015

Locations

Countries of recruitment

Greece

Trial participating centre

Eginitio Hospital
Athens
GR-11528
Greece

Trial participating centre

General State Hospital- George Gennimatas
Athens
GR-11517
Greece

Trial participating centre

Evaggelismos Hospital
Athens
GR-10676
Greece

Trial participating centre

University General Hospital - Attikon
Athens
GR-12462
Greece

Trial participating centre

Filoktitis Rehabilitation Centre
Athens
GR-19400
Greece

Trial participating centre

Iatriki Askisis Rehabilitation Centre
Athens
GR-15342
Greece

Trial participating centre

IKA Neas Ionias
Athens
GR-14234
Greece

Trial participating centre

University General Hospital of Patras
Rio, Patras
GR-26504
Greece

Sponsor information

Organisation

University of Athens

Sponsor details

Panepistimioupoli
15784 Zografou
Athens
15784
Greece

Sponsor type

University/education

Website

Funders

Funder type

Not defined

Funder name

European Union (European Social Fund – ESF)

Alternative name(s)

Funding Body Type

Funding Body Subtype

Location

Funder name

Greek national funds through the Operational Program "Education and Lifelong Learning" of the National Strategic Reference Framework (NSRF)

Alternative name(s)

Funding Body Type

Funding Body Subtype

Location

Results and Publications

Publication and dissemination plan

The main findings will be published as a peer-reviewed journal article.

Results published in thesis: http://openaccess.city.ac.uk/19773/

Intention to publish date

31/12/2018

Participant level data

Stored in repository

Basic results (scientific)

Methods: The study is a randomised trial using a waiting list control. Of the 72 participants of Thales, 58 met the eligibility criteria for speech and language therapy and 39 were allocated to ESFA (19 allocated to mapping therapy). Participants were randomised via recruitment order to one of three groups- two groups of therapy (direct or combination) and the waiting list control group. Of the 38 that had ESFA, 12 were randomised to the waiting list control group and 26 to one of the two ESFA therapy approaches. Participants on the therapy approaches were assessed two times before therapy (double baseline, week 1- 6), post-therapy (week 19), and 3-months later (follow-up). Participants on the waiting list control were assessed three times before therapy (week 1-6-19) and then were randomly allocated to one of the two approaches for ESFA treatment and were reassessed after the 12-week treatment (post-therapy) and 3 months later (follow-up). Both therapy groups had equal intensity and dosage- three hours of ESFA per week for 12 weeks (36 hours): those that received direct ESFA had three 1-hour sessions per week; those that received combination ESFA had one 90-minute session of group ESFA and two 45-minute sessions of individual ESFA per week. The primary outcome measure was confrontation naming of the 260 colourised pictures initially developed by Snodgrass and Vanderwart (1980) (Rossion & Pourtois, 2004). Secondary outcome measures included a range of assessments tapping on all WHO ICF levels: Boston Naming Test (BNT), Discourse Measurement with Cookie Theft picture, Functional Assessment of Communication Skills for adults (ASHA – FACS), Stroke and Aphasia Quality of Life scale (SAQOL-39g), General Health Questionnaire (GHQ-12) and EQ-5D.

Therapy materials appropriate to each person were chosen at baseline before initiation of therapy. At baseline, each participant had to name the 260 pictures. The pictures were randomly presented to each participant for naming across three trials without any cuing or feedback. Based on the results of these trials, the pictures that participants failed to name on at least two trials were selected as potential treatment materials. This process of stimulus selection resulted in a set of treatment and probe items that were individual to each participant.

To test (a) the efficacy of ESFA therapy (n=26) versus no therapy (n=12) mixed within-between ANOVAs were used with group as the between variable (2 groups: ESFA versus control) and time as the within variable (3 levels: weeks 1, 6, 19). To test (b) the relative efficacy of direct (n=22) versus combination (n=14) ESFA, mixed within-between ANOVAs were used with group as the between variable (2 groups: direct versus combination ESFA) and time as the within variable (4 levels: two baselines, post-therapy and follow-up).

Results: After applying a Bonferroni correction for multiple comparisons, for (a) therapy versus control, there was a significant main effect of time on the primary outcome measure Greenhouse-Geisser F (1.1, 39.38) = 26.04, p< .001 with a large effect size (η2p = .42), and a significant interaction effect Greenhouse-Geisser F (1.1, 39.38) = 9.56, p= .003 with a large effect size (η2p = .21); whereby the therapy group improved significantly more from pre-therapy (week 6) [mean (SD) = 61.96 (49.40)] to post-therapy (week 19) [mean (SD) = 104.38 (73.91)] than the control group [week 6 mean (SD) = 74.33 (62.94), week 19 mean (SD) = 81.83 (69.90)]. There was a significant main effect of time for the BNT (p = .002) with a large effect size (η2p = .19), with the significant difference between the firsts two baselines and BL3/post therapy. There was an interaction effect, which did not remain significant after adjusting for multiple comparisons, for the SAQOL-39g psychosocial domain (p = .013) (η2p = .12) and the overall SAQOL-39g score (p = .015) (η2p =.11), with the therapy group improving with therapy, and the control group not improving.

For (b) direct versus combination ESFA, there was a significant main effect of time on the primary outcome measure for both approaches, Greenhouse-Geisser F (1.89, 64.53) = 32.95, p < 0.001 with large effect size (η2p = .49). Pairwise comparisons showed there was a significant difference between the two baselines (mean difference = 10.23, p= .003), a significant difference between both the baselines and post-therapy (mean differences= 49.70 and 39.45, ps< .001) and a significant difference between both the baselines and follow- up (mean differences = 43.45 and 33.22, ps< .001). The post therapy gains were maintained, i.e. there was no significant drop from post-therapy to follow up. There was also a significant main effect of time with large effect size for the BNT (p< .001) (η2p = .29), with significant differences in pairwise comparisons between both baselines and post therapy and both baselines and follow-up; and the ASHA-FACS (p = .001) (η2p =.18), with significant differences between both baselines and the follow -up assessment. The interaction and group effects were not significant.

Conclusion: This study is the first to explore the efficacy of ESFA in a randomised group design. Results supported the efficacy of ESFA therapy versus no therapy. ESFA therapy led to gains in naming, communication and quality of life for people with aphasia. Gains were similar in the two therapy approaches and were maintained over a three-month follow-up. Pending further research to confirm the reliability of the results and allow meaningful effects to be detected on a range of outcome measures, ESFA may be a useful therapy to adopt in practice.

Publication list

Publication citations

Additional files

Editorial Notes