Trial Outcomes & Findings for Digital Interventions in Neurorehabilitation: iTALKbetter (NCT NCT04566081)

NCT ID: NCT04566081

Last Updated: 2026-03-10

Results Overview

The WRT is custom test of word retrieval comprised of a subset of the trained and untrained items form the iTALKbetter therapy. Participants accuracy on the word retrieval test is measured at two timepoints: pre intervention (T3), and post intervention (T4). There are 110 trained and 110 untrained items. Accuracy is reported as a percentage score (min = 0%, max = 100%) for each of the four conditions. The outcome is the percentage change in performance on the WRT. To investigate the efficacy of iTalkBetter in improving the retrieval of trained single words on the WRT, a repeated measures ANOVA was used which compared change in accuracy over the pre-therapy block and the therapy block. There were two factors, each with two levels: time (pre-therapy (Baseline to T3) and therapy (T3 to T4)); and item (trained and untrained lexical items).

Recruitment status

COMPLETED

Study phase

NA

Target enrollment

31 participants

Primary outcome timeframe

Two time points: 1. Pre- intervention (T3 = 6 weeks post baseline) 2. Post-intervention (T4 = 12 weeks post baseline)

Results posted on

2026-03-10

Participant Flow

31 PWA were recruited for the study. Recruitment was from the Predicting Language Outcome and Recovery After Stroke (PLORAS) database and from a local outpatient clinic. Between Sep 7, 2020 and Mar 1, 2022.

Participant milestones

Participant milestones
Measure
iTalkBetter Therapy Group
The iTalkBetter clinical study was a small-scale, repeated measures, control trial in which participants completed 6 weeks of therapy with the iTalkBetter app, aiming to complete 60 hours of therapy
Overall Study
STARTED
31
Overall Study
COMPLETED
27
Overall Study
NOT COMPLETED
4

Reasons for withdrawal

Reasons for withdrawal
Measure
iTalkBetter Therapy Group
The iTalkBetter clinical study was a small-scale, repeated measures, control trial in which participants completed 6 weeks of therapy with the iTalkBetter app, aiming to complete 60 hours of therapy
Overall Study
Death
1
Overall Study
Withdrawal by Subject
3

Baseline Characteristics

Race and Ethnicity were not collected from any participant.

Baseline characteristics by cohort

Baseline characteristics by cohort
Measure
iTALKbetter
n=27 Participants
The iTalkBetter clinical study was a small-scale, repeated measures, control trial in which participants completed 6 weeks of therapy with the iTalkBetter app, aiming to complete 60 hours of therapy
Age, Continuous
62 Years
STANDARD_DEVIATION 12 • n=27 Participants
Sex: Female, Male
Female
7 Participants
n=27 Participants
Sex: Female, Male
Male
20 Participants
n=27 Participants
Word Retrieval Test (WRT)
32 Percentage
STANDARD_DEVIATION 15.77 • n=27 Participants

PRIMARY outcome

Timeframe: Two time points: 1. Pre- intervention (T3 = 6 weeks post baseline) 2. Post-intervention (T4 = 12 weeks post baseline)

Population: Results are for the 27 participants who completed the study.

The WRT is custom test of word retrieval comprised of a subset of the trained and untrained items form the iTALKbetter therapy. Participants accuracy on the word retrieval test is measured at two timepoints: pre intervention (T3), and post intervention (T4). There are 110 trained and 110 untrained items. Accuracy is reported as a percentage score (min = 0%, max = 100%) for each of the four conditions. The outcome is the percentage change in performance on the WRT. To investigate the efficacy of iTalkBetter in improving the retrieval of trained single words on the WRT, a repeated measures ANOVA was used which compared change in accuracy over the pre-therapy block and the therapy block. There were two factors, each with two levels: time (pre-therapy (Baseline to T3) and therapy (T3 to T4)); and item (trained and untrained lexical items).

Outcome measures

Outcome measures
Measure
iTALKbetter
n=27 Participants
The therapy is a single-word picture-naming task with error-reducing, phonological cues, and includes 220 lexical items from all major word classes. The therapy content also includes both concrete (highly imageable words which portray tangible concepts e.g., 'cat') and abstract words (low imageability words which refer to intangible concepts e.g., 'confidence'). A novel automatic speech recogniser was developed by a member of the team and incorporated into the therapy. This naming utterance verification system (NUVA) utilised a deep learning element that classified, in real time, a naming attempt as correct or incorrect. This allowed the app to determine the next cue level and also enabled the provision of immediate feedback to the PWA for each individual practice of a word (trial). See Supplementary materials for the item progression algorithm. iTalkBetter also utilises gamification to reduce the boredom and fatigue effects often associated with repetitive and intensive tasks, thereby decreasing participant drop-out rates. In iTalkBetter, bright and engaging colours are used in the therapy via the use of an 'outer space' theme, which was designed and developed by our software team, SoftV. Visual feedback, in the form of words ('Well done!') and actions (pictures are collected and stored in sight on screen), are also provided to motivate users when items are correctly named.
Accuracy Performance (%) on a Bespoke Word Retrieval Test: WRT. Two Conditions: 1. Trained Items 2. Untrained Items
Untrained items (T4)
43 percentage
Standard Error 1.95
Accuracy Performance (%) on a Bespoke Word Retrieval Test: WRT. Two Conditions: 1. Trained Items 2. Untrained Items
Trained items (T3)
38.3 percentage
Standard Error 0.85
Accuracy Performance (%) on a Bespoke Word Retrieval Test: WRT. Two Conditions: 1. Trained Items 2. Untrained Items
Untrained items (T3)
38.4 percentage
Standard Error 0.95
Accuracy Performance (%) on a Bespoke Word Retrieval Test: WRT. Two Conditions: 1. Trained Items 2. Untrained Items
Trained items (T4)
50.6 percentage
Standard Error 1.8

SECONDARY outcome

Timeframe: Baseline and Post-intervention (T4 = 12 weeks post Baseline)

Semi-structured interviews were carried out with partcipants to better understand why they took part in the study, and what they hoped to gain (baseline) and then reflections on what it was like taking part in the study (Last time point)

Outcome measures

Outcome data not reported

Adverse Events

iTALKbetter

Serious events: 0 serious events
Other events: 0 other events
Deaths: 1 deaths

Serious adverse events

Adverse event data not reported

Other adverse events

Adverse event data not reported

Additional Information

Professor Alex Leff

UCL Queen Square Institute of Neurology

Phone: 02076791129

Results disclosure agreements

  • Principal investigator is a sponsor employee
  • Publication restrictions are in place