Larger RCT (n=56, ages 3-7): VR-supported speech therapy for developmental language disorder was feasible and safe, but no more effective than standard therapy
How this was rated
Studio controllato randomizzato monocentrico registrato (NCT07438639), CONSORT-2025, con cecità dei valutatori e un'appropriata analisi tra i gruppi a effetti misti lineari - un chiaro passo avanti metodologico rispetto allo studio pilota Cappadona 2023. Moderata anziché alta perché la questione dell'efficacia è sostanzialmente nulla (nessuna interazione Gruppo x Tempo significativa su nessuno dei 14 esiti) e potrebbe essere sottodimensionata a n=56; 14 esiti sono stati testati con un controllo limitato della molteplicità; lo strumento di usabilità era un questionario non standardizzato, specifico per lo studio; il campione era monocentrico con uno squilibrio al basale nella denominazione degli oggetti; e il moderatore relativo ai bambini più piccoli è esplorativo. Le conclusioni su fattibilità, sicurezza e accettabilità sono ben supportate; qualsiasi lettura di un vantaggio in termini di efficacia non lo è.
Ratings use a simplified four-tier scheme (High, Moderate, Low, Very Low) informed by the GRADE working group. Learn more about how studies are rated.
This randomized controlled trial followed up an earlier pilot, enrolling 56 preschoolers (ages 3-7, mean 4.5) with developmental language disorder and randomizing them to VR-supported speech therapy or standard therapy for six months. The VR system (VRRS, Khymeia) was non-immersive, used through a touchscreen, not a headset. Both groups improved substantially. On the primary between-group test the VR group did not significantly outperform standard therapy on any language outcome, though it was feasible, safe, and well-accepted, with 100% retention.
The larger follow-up to the Cappadona 2023 pilot, from the same Messina group and using the same non-immersive VRRS touchscreen system, and the most rigorous study in this line: a registered, CONSORT-2025, assessor-blinded RCT (n=56) with a proper between-group analysis. Both VR-supported and standard speech therapy produced large within-group language gains over six months, but the mixed-effects Group x Time interaction was non-significant for all 14 outcomes, so VR did not significantly outperform standard therapy. The headline contribution is feasibility, safety, and acceptability (100% retention, no adverse events, good therapist-rated usability), not added efficacy. An exploratory analysis hinted that younger children may benefit more from the VR arm, but some between-group differences were confounded by baseline imbalance. Read it as evidence for a feasible, well-tolerated delivery format, not as evidence that VR adds clinical benefit over conventional therapy for DLD.
Key findings
- Larger follow-up to the Cappadona 2023 pilot: 56 children (ages 3-7, mean 4.5; roughly 70% male) with DSM-5 DLD, randomized to VR-supported (VRRS) vs standard speech therapy, two 60-min sessions per week for six months
- Both arms improved substantially within-group, with large Cohen's d on most language measures (frequently 0.7 to 1.1); the experimental group reached significance on slightly more domains
- PRIMARY between-group test (linear mixed-effects Group x Time interaction): NON-significant for all 14 language outcomes - VR did not significantly outperform standard therapy
- A few post-treatment between-group differences favored the VR group (comprehension of words, naming of body parts, total naming), but the object-naming difference was confounded by a baseline imbalance present at T0
- Exploratory moderator analysis suggested the VR arm may be more effective for younger children (age-by-group interaction), with no effect of developmental quotient
- Feasibility and acceptability: 100% retention in the experimental group, no adverse events, and good therapist-rated usability on a non-standardized questionnaire
Background
Developmental language disorder affects an estimated 7-10% of children and is one of the most common reasons for speech and language therapy referral. Conventional therapy works, but sustaining the attention and regular attendance of young children over long protocols is a persistent challenge, and interest has grown in technology-supported formats that may improve engagement. An earlier pilot from this group (Cappadona 2023, also in this Hub) reported broad within-group language gains from a VR-supported approach but did not include a formal between-group comparison. This trial set out to extend that pilot with a larger sample and a proper between-group analysis.
What the researchers did
De Domenico and colleagues at IRCCS Centro Neurolesi Bonino-Pulejo in Messina, Italy, ran a registered (NCT07438639), single-blind, CONSORT-2025 randomized controlled trial. They enrolled 56 children aged 3-7 (mean 4.5 years, roughly 70% male) with DSM-5 developmental language disorder and a developmental quotient of at least 85, randomizing them to an experimental group (n=28) or control group (n=28). Both arms received the same therapeutic content across the same core language domains, twice weekly for one hour over six months. The control group received traditional in-person therapy with printed materials and games; the experimental group received the same content delivered through the VRRS (Khymeia, Padova). As in the pilot, the VRRS was used in non-immersive mode: children interacted through a touchscreen, not a head-mounted display. The primary outcome was change on the Italian Test of Language Development (TVL) from baseline (T0) to post-treatment (T1). Assessors were blind to group allocation. Analyses included within-group Wilcoxon tests, between-group comparisons, and linear mixed-effects models with a Group x Time interaction as the primary test of differential effect.
What they found
Both groups improved substantially. Within-group effect sizes were large in both arms across most language measures (Cohen’s d frequently 0.7 to 1.1), and the experimental group reached significance on slightly more domains than the control group. The decisive test, however, was the between-group comparison. In the linear mixed-effects models, the Group x Time interaction was non-significant for all 14 language outcomes, meaning the magnitude of improvement did not differ significantly between VR-supported and standard therapy once within-subject variability was accounted for. A handful of post-treatment between-group differences favored the experimental group (comprehension of words, naming of body parts, and total naming), but the object-naming difference was confounded by a baseline imbalance already present at T0. An exploratory moderator analysis suggested the experimental approach may be more effective for younger children, with no effect of developmental quotient. On the secondary aims, the experimental group had 100% retention, no adverse events were observed, and therapists rated the system as usable and easy to integrate, using a non-standardized, study-specific questionnaire.
Why this matters
This is the most rigorous study in this line, and it is a useful corrective to its own pilot. The pilot’s broad within-group gains looked promising, but when this larger trial applied a formal between-group test, VR-supported therapy did not significantly outperform conventional therapy on any outcome. The honest reading is that, for preschool DLD, a non-immersive touchscreen VR system is a feasible, safe, and well-accepted way to deliver therapy content, not a demonstrated way to improve language outcomes beyond standard care. That distinction matters: feasibility and engagement are real and valuable, but they are not the same as added efficacy. It is also worth being clear about modality. This is a 2D touchscreen rehabilitation system, not immersive head-mounted VR, and the population is preschoolers building foundational language, so the findings do not transfer to immersive speaking-situation practice with older children or adults.
Limitations
The efficacy question may be underpowered at n=56, and a non-significant interaction is not proof of equivalence. Fourteen outcomes were tested with limited control for multiplicity. The usability measure was a study-specific questionnaire, not a validated instrument. The trial was single-site, with a baseline imbalance on object naming. Children and families cannot be blinded to a VR intervention. Outcomes were standardized language test scores; real-world or functional communication was not directly assessed. The finding that the approach may work better for younger children is exploratory and needs direct confirmation.
Implications for practice
For preschool children with DLD, this trial supports VR-supported speech therapy (in non-immersive, touchscreen form) as a feasible, safe, and well-accepted way to deliver conventional therapy content, not as a more effective one. On the rigorous between-group test it did not beat standard therapy, so it is best framed as an engaging delivery option rather than a route to better language outcomes. The hint that younger children benefit more is worth testing directly but is not yet actionable. Note the modality: this is a 2D touchscreen system, not immersive head-mounted VR, so it does not speak to immersive speaking-situation practice with older children or adults.
Implications for research
This study answered the Cappadona 2023 pilot's call for a larger trial with formal between-group testing, and the between-group efficacy signal did not hold. The logical next steps are an adequately powered, multi-site trial with a pre-registered single primary outcome and multiplicity control, a direct test of the younger-child moderator, and real-world or functional communication outcomes rather than standardized test scores alone.
Cite this study
If you reference this study in your work, the canonical citation formats are:
@article{dedomenico2026,
author = {De Domenico, C. and La Fauci, M. and Mancuso, N. and Caputo, M. and Di Cara, M. and Piccolo, A. and Fulgenzi, A. and Borzelli, D. and Impallomeni, C. and Tripodi, E. and Calabrò, R. S. and Quartarone, A. and Cucinotta, F.},
title = {Virtual Reality-Supported Speech Therapy in Children with Developmental Language Disorder: A Randomized Controlled Trial},
journal = {Medical Sciences},
year = {2026},
doi = {10.3390/medsci14020291},
url = {https://withvr.app/evidence/studies/de-domenico-2026}
}TY - JOUR
AU - De Domenico, C.
AU - La Fauci, M.
AU - Mancuso, N.
AU - Caputo, M.
AU - Di Cara, M.
AU - Piccolo, A.
AU - Fulgenzi, A.
AU - Borzelli, D.
AU - Impallomeni, C.
AU - Tripodi, E.
AU - Calabrò, R. S.
AU - Quartarone, A.
AU - Cucinotta, F.
TI - Virtual Reality-Supported Speech Therapy in Children with Developmental Language Disorder: A Randomized Controlled Trial
JO - Medical Sciences
PY - 2026
DO - 10.3390/medsci14020291
UR - https://withvr.app/evidence/studies/de-domenico-2026
ER - Know of research that should be in this hub? If a relevant peer-reviewed study is not listed here, send the reference to hello@withvr.app. The hub is kept up to date as the literature grows.
Funding & independence
Current Research Funds 2025, Ministry of Health, Italy. The study used the commercial VRRS system (Khymeia, Padova) but was not funded by Khymeia, and the authors declare no conflicts of interest. No withVR BV involvement in funding, study design, or authorship. Summary prepared independently by withVR using the published peer-reviewed paper.