Effectiveness of an online interval group therapy for children with developmental language disorders: protocol for a randomised controlled intervention study

Abstract

Introduction

With a prevalence of around 7.6%, developmental language disorders (DLDs) without comorbidities are among the most common and most frequently treated childhood disorders. Standard DLD therapy in Germany consists of individual therapy sessions once per week within speech–language therapy practices. In reality, these sessions only take place every 10–14 days on average. Online therapy may be beneficial but is not yet standard practice in Germany. Although DLD group therapy has been proven to be effective, it is rarely undertaken in Germany. The aim of this study is to compare the effectiveness of online DLD therapy for small groups of children with standard one-to-one therapy.

Methods and analysis

The effectiveness of two treatment settings is evaluated in 212 children with moderate-to-severe DLD (ages 3 years to 6 years 11 months) in the multicentre, block randomised controlled trial (RCT) THErapy ONline. Five centres in Germany participate. Children are randomly assigned to the intervention group (online interval-intensive therapy, IG1, n=106) or the control group (extensive standard in-person therapy, IG2, n=106). A speech and language assessment is conducted at baseline (study entry, T0), 12 months (T1) and 18 months (T2) after therapy start. The co-primary outcome parameters are the speech and language test scores of phonological speech sound production, expressive vocabulary, grammar production and language comprehension at T1. The secondary outcome parameters comprise two composite speech and language test scores at T1 and T2, including phonological working memory scores and the individual scores of the aforementioned tests at T2, as well as process evaluation parameters (time expenditure, resource utilisation, such as salary costs of speech–language therapists, additional costs of the online therapy, adherence to appointments and therapy acceptance).

Ethics and dissemination

This study has been approved by the Institutional Ethics Review Board of Westphalia-Lippe (2022-282 f-S). Parents provide written informed consent. Findings will be disseminated through presentations, peer-reviewed journals and conferences.

Trial registration number

DRKS00030068

STRENGTH AND LIMITATIONS OF THIS STUDY.

• The study will contribute to the body of multicentre, randomised controlled trials on interventions for developmental language disorders (DLDs).

• This approach (online group language therapy) is expected to improve the language and communication abilities of children with DLD, to empower parents in high-quality communication with their child and to increase adherence to appointments, treatment options during pandemics and therapy access for families who have additional logistical difficulties with participation.

• A follow-up period of 6 months allows a robust assessment of the clinical effectiveness and medium-term effects.

• The effectiveness, as well as the costs and cost-effectiveness, will be analysed.

• Only children with moderate or severe DLD and only native or non-native German-speaking children with more than 10 months of intensive contact (several hours per day) with the German language will be included.

Introduction

Developmental language disorders (DLDs) belong to the most common and most frequently treated disorders in childhood. The International Classification of Diseases 11th revision of the WHO¹ characterises DLD (6A01.2) as follows: “The […] ability to understand, produce or use language is markedly below what would be expected given the […] age and level of intellectual functioning. The language deficits are not explained by another neurodevelopmental disorder or a sensory impairment or neurological condition…”.¹ The prevalence of such cases, where none of the above-named co-occurring impairments (comorbidities) such as hearing loss, autism spectrum disorders, neurological disorders or intellectual disability are to be expected, is approximately 7.6%.²

DLDs often seriously impair the socio-emotional, cognitive, academic and professional development of the children affected^{3 4} and can continue into adulthood.^{5 6} Problems in reading and spelling acquisition are reported in 40% to 55% of children with DLD, learning difficulties in approximately 40%^{3 7} and significantly lower cognitive, educational and occupational levels than for typically developing peers.⁶ DLDs are often associated with impaired self-esteem, behavioural disorders and attention deficit/hyperactivity disorders, as well as motor and coordination impairments.^{6 8} They usually do not recover without specialised interventions.^{5 9} In Germany, 14.3% of boys and 9% of girls insured by the country’s largest health insurance company are currently receiving speech–language therapy, most of them between the ages of 5 and 9 years.¹⁰

The German clinical practice guideline by Neumann et al¹¹ highlighted the clear need for research on evidence-based DLD interventions in Germany, demonstrating, as it did, that speech–language therapy takes place relatively late in childhood in Germany, and that only five German randomised controlled trials (RCTs) from three working groups were identified as eligible for inclusion in the guideline. However, children and parents have the right to modern evidence-based interventions. Prescribers and therapists need to know about such interventions in order to counsel parents appropriately and reach an informed agreement and consent on individualised, evidence-based intervention for their child.

DLD therapies are increasingly being evaluated not only in terms of their effectiveness but also in terms of the efficiency of their practical implementation. Efficiency of implementation depends strongly on the therapy setting (eg, in person vs online), intensity or ‘dosage’ and the qualitative dosage component or ‘dose form’ (ie, the techniques, tasks, activities and intervention contexts used in a therapy session).12,14

The decision as to whether an individual’s therapy should be carried out in person or online (one aspect of the ‘setting’) must take account of the available evidence. While conducting speech–language therapy in face-to-face sessions meets the current professional standards in Germany, the American Speech–Language–Hearing Association already stated in 2005 that telemedicine has “the potential to extend clinical services to rural, remote, and underserved populations and to culturally and linguistically diverse populations.”¹⁵ One systematic review showed that telemedicine improves access to healthcare,¹⁶ for example, by providing remote care in several areas. As such, online provision became particularly relevant during the COVID-19 pandemic. The decision as to whether speech–language therapy should be conducted online is a joint decision to be made by the therapist and patient. The prerequisite for this is that the prescribing physician has not ruled out online therapy. The therapist must in all cases point out the possibility of face-to-face therapy to the patient, and face-to-face therapy must take precedence if the therapy goals could not be expected to be achieved as successfully by online therapy.

There are, however, very few studies regarding the effectiveness of online DLD therapy that enable such a comparison between online and face-to-face settings. One German intervention study with children who stuttered revealed that the effectiveness of combined face-to-face and online therapy versus face-to-face therapy alone was comparable.¹⁷ Evidence for the effectiveness and feasibility of online DLD treatment has been provided recently in an RCT conducted by some of the authors of the current paper. This study, named THESES (Therapien von Sprachentwicklungsstörungen), featured a 9-month follow-up period, in which significantly better language outcomes were found for 51 children undergoing online one-to-one therapy once a week than the 51 children receiving standard therapy (p=0.002, Cohen’s d=0.646).¹⁸

The selection of individual or group therapy is an important component of the setting of a patient’s DLD therapy. A Cochrane systematic review with meta-analysis reported positive effects of speech–language therapy for children with expressive phonological and expressive vocabulary difficulties and mixed findings for expressive syntax, but found no difference between individual and group therapy.¹⁹ A later systematic review by Nelson et al²⁰ provided additional evidence for improvements in receptive language outcomes and showed that, in children aged 3–5 years—the same age range examined in our study—both group-based and clinician-directed interventions were effective in enhancing expressive and receptive language skills in children with DLD. More recently, a systematic review and meta-analysis by Kunnari et al,²¹ focusing specifically on group interventions targeting expressive phonology in children with developmental speech and language disorders, reported positive effects on speech production accuracy, with medium-to-large average effect sizes for phonological accuracy.

Despite the heterogeneity and small sample sizes of many of the studies included across all three reviews and ongoing uncertainty regarding the long-term effectiveness of speech and language therapy, comparative research on different treatment formats, such as individual versus group therapy, remains warranted. However, group therapy is used only in 0.5% of voice, speech, language and swallowing therapy cases in Germany.²² In the above-named RCT, face-to-face DLD therapy with 51 children in small groups of 2–3 tended to show better results than standard individual therapy, but the difference failed to reach significance.²³ Other studies from Germany and Luxembourg have shown that very frequent speech–language therapy sessions in small groups of nursery-age children with DLD led to significant and stable benefits in the development of phonological (speech sound) awareness, grammar and reading skills.24,26 These studies demonstrated that this therapy strategy was more effective than extensive individual or group therapy²⁵ or no therapy²⁶ and found a possible beneficial effect of peer learning on language development,²⁵ among other factors. Other studies have shown at least comparable effectiveness between group and individual therapies but greater cost-efficiency for group therapy.^{27 28}

Therapy frequency and dose must be considered when planning DLD therapy. Standard DLD therapy in Germany takes place once a week for 45 min. However, for various reasons, such as illnesses, holidays and organisational difficulties, treatment sessions only take place every 14 days on average.²⁹ Even under the strict therapy regimens implemented as part of an intervention study,³⁰ it was only possible to undertake on average 0.7 therapy sessions per week. The question of the effectiveness of ‘extensive’ standard therapy compared with interval therapy featuring intensive therapy blocks remains to be answered. Motsch and Schmidt reported that shorter intensive DLD group therapy was more beneficial than longer extensive individual therapy.²⁵ Another German study³¹ found that high-frequency block and interval therapy in combined individual and group settings was effective, at least in the short and medium term.

Other studies have also demonstrated the effectiveness of high-frequency input for language-associated receptive skills or phonological awareness.^{32 33} Gillam et al³⁴ compared four intensive treatment approaches for children with DLD in an RCT, including a non-specific comparison treatment group. Participants received therapy sessions five times per week over 6 weeks. Interestingly, the study found superior outcomes only for phonological awareness in the Fast ForWord–Language and Computer-Assisted Language Intervention groups compared with both conventional therapy and the comparison group, but no differences in overall language abilities.³⁴ More intensive therapy (semantics) showed medium-term effectiveness even several months after the end of therapy,³⁵ and therapy efficiency could also be increased in late talkers (2–3 years) through higher therapy frequency.³⁶ In contrast, no significant proof of efficacy could be provided for very extensive individual therapy.³⁷

No direct comparison of the effectiveness of individual, face-to-face, outpatient DLD therapy and online therapy (specifically in small groups) has taken place until now, at least not in German-speaking countries. A recent literature review³⁸ reported that while many therapy providers for children with communication disorders have moved rapidly from a face-to-face to a telepractice setting following the COVID-19 outbreak, the relevant literature indicates a clear need for an evidence base for such conversions, including a description and analysis of the physical and psychological capacity components of teletherapy participants.

Aims of the study

Online speech–language therapy in small groups may be more effective than individual face-to-face therapy,²³ at least regarding subtypes of DLD, and interval therapy with intensive treatment phases and breaks may be more effective than extensive ones. The project THErapy ONline (THEON) therefore compares the benefits of online interval therapy in small groups of 2–3 children with DLD with standard, individual face-to-face therapy.

The primary hypothesis (H1) is that video-assisted, online, interval group therapy will result in significantly greater improvements in language performance 1 year (time point T1) after study entry (T0) for children with DLD than standard, extensive face-to-face therapy.

The secondary hypotheses (H2) are:

• H2-1: Online, interval, group therapy leads to significantly higher improvements in language performance measured 18 months after the start of therapy (follow-up; T2) in children with DLD than standard, individual face-to-face therapy.

• H2-2: Online, interval, group DLD therapy is more cost-effective than standard therapy.

• H2-3: Online, interval, group DLD therapy leads to increased adherence and fewer dropouts and is well accepted by children, parents and therapists, making it suitable for use during pandemics.

• H2-4: Online, interval, group DLD therapy improves access to speech–language therapy for families with migrant backgrounds and/or low social status, single parents, large families and rural populations. This study is designed to generate robust evidence on the effectiveness and feasibility of online small-group language therapy for children with DLD, an intervention for which empirical data are currently lacking. In the long term, the findings are expected to contribute to improving the standard of care for children with DLD.

Methods and analysis

Study design

THEON is a prospective, multicentre, block randomised controlled study with two parallel intervention groups. 212 participants are randomly assigned to an online intervention group (IG1: interval group therapy) or a standard intervention group (IG2: individual face-to-face therapy). The randomisation is stratified according to age group (3 years, 0 months to 5 years, 5 months; and 5 years, 6 months to 6 years, 11 months) and sex at baseline. Block randomisation will be performed within the strata. The random lists for the strata are created by the Kompetenzzentrum für Klinische Studien Bremen (Competence Center for Clinical Studies, University of Bremen) (authors MS and WB) using SAS V.9.4 (TS1M6). The allocations are distributed to centres either in sealed envelopes or online. Neither the recruiters nor the participants have access to the randomisation sequence. Therapists do not administer assessments to children they have personally treated. Further blinding is not possible.

Study settings

The department of phoniatrics and paediatric audiology at the University Hospital Münster (DPPAUHM) coordinates this multicentre study (principal investigator: KN). It is conducted in cooperation with the Special Education and Therapeutic Centre Neuwied gGmbH (SETCN). Online group therapy in both facilities is carried out using the CompuGroup Medical video programme ELVI. 20 cooperating speech–language therapy practices are also recruited for the standard, individual DLD therapy.

Participants, recruitment and tests

The recruiting regions are Münster and the surrounding area, up to a radius of 200 km (North Rhine-Westphalia, Lower Saxony), and the region around Neuwied, up to a radius of 50 km (Rhineland-Palatinate). The involvement of these large inclusion areas (which also include rural districts), as well as the fact that almost one in three children (31%) within nurseries in the western federal states of Germany (including North Rhine-Westphalia) has a migration background,³⁹ ensures that this study is truly population-based, incorporating, as it does, children with DLD from rural areas, with a migration background and from families with low social status (see study hypothesis H2-4). The children participating are recruited from and within DPPAUHM, SETCN and from speech–language therapy practices and nurseries that work together with these institutions. Potential study participants are recruited within the participating institutions, via the websites of the study centres and via appeals on social media and in newspapers.

The project includes children with moderate-to-severe DLD, aged between 3.0 and <7.0 years at study entry.

The criteria for moderate-to-severe DLD in children are:

1. A T value of ≤35 in at least one subtest covering the linguistic domains of:

   1. Language comprehension.

   2. Expressive syntax and morphology.

   3. Expressive semantics and lexicon.

2. And/or (4) phonological speech sound production delayed by ≥6 months in at least 3 phonemes (speech sounds) and/or phonological processes (physiological temporary simplifications of speech sounds in adult speech that are overcome during speech and language development) or show pathological phonological processes (speech sound substitutions that do not occur in typical speech and language development).

3. mean Intelligibility in Context Scale≤3.6.

   For enrolment in the study, participants must meet the following criteria:

• Hearing: bilateral peripheral normal hearing, that is, any hearing loss determined either from the pure-tone audiogram according to Röser’s four-frequency table, developed in 1973 and commonly used in Germany, or by means of auditory evoked potentials (auditory brainstem responses or auditory steady-state responses), must be ≤20%. This criterion is aligned with the WHO classification of hearing loss (2021),⁴⁰ which, however, applies only to adults and defines normal hearing as a hearing threshold (minimum sound intensity that an ear can detect, averaged across 500, 1000, 2000 and 4000 Hz) of less than 20 decibels hearing level (dB HL) in the better-hearing ear.

• Cognitive developmental assessment, for example using the SON-R 2–8 (Snijders-Oomen Nonverbaler Intelligenztest (German version)⁴¹) or the PITVA (Potsdamer Intelligenztest für das Vorschulalter)⁴²: IQ≥85.

• Language background: monolingual German-speaking children, or children with German as a second language who have had at least 10 months of exposure to German for several hours per day, eg, in a nursery or with German-speaking siblings.

Participants should not have received any speech–language therapy in the previous 3 months. Speech–language therapy parallel to the online intervention group may not be undertaken. Other non-speech–language therapies such as physiotherapy, occupational therapy and general early intervention are permitted.

The four coprimary outcome parameters are the four speech and language test scores of phonological speech sound production, expressive vocabulary, grammar production and language comprehension (see below for details on tests implemented) at T1. The secondary outcome parameters comprise two composite speech and language test scores evaluated for T1 and T2: (1) a composite score resulting from a principal component analysis using the first principal component derived from the four above-mentioned individual test scores combined with the phonological working score memory with weightings obtained at T0 and (2) a composite score calculated as Percentage of Consonants Correct (PCC)/100 multiplied by the mean T value of the other speech and language tests scores mentioned above and the phonological working memory score, allowing percentage-based and T-scaled measures to be combined into a single metric. This approach has already been applied in our previous THESES study and allows a comparison of both studies.⁴³ Phonological working memory was included in the composite score because it is a core component of DLD and a strong predictor of later reading and spelling abilities. Further secondary outcome parameters comprise the individual scores of the aforementioned tests at T2, in addition to process evaluation parameters (time expenditure, resource utilisation, such as salary costs of speech–language therapists, additional costs of the online therapy, adherence to appointments and therapy acceptance).

The original study protocol had specified a composite score of all four language test measures at T1 as the primary outcome parameter. Although the use of a single composite outcome measure is methodologically appealing and statistical pooling of test scores is mathematically straightforward (as illustrated by the second composite score), we decided not to retain this original approach. The main reason was that differential weighting of individual linguistic domains is inherently problematic and dependent on varying theoretical and clinical perspectives. For example, some expert frameworks prioritise morphosyntactic performance, whereas others place greater emphasis on phonological outcomes. At the same time, phonological speech sound disorders tend to have the most favourable spontaneous prognosis, while exerting a comparatively strong impact on later literacy development. Therefore, we deliberately refrained from applying differential weightings to combine the language measures into a single primary composite score. Instead, the individual language test scores are now defined as coprimary outcome parameters and a composite score, based on weightings derived at T0, is retained as a secondary outcome parameter. The statistical analysis and sample size calculation are adapted accordingly, accounting for the multiple coprimary endpoints.

At the beginning of the study (T0), potential participants undergo phoniatric, audiological, speech and language and psychological examinations, including a medical history and a history of speech and language development, an audiometric assessment using either conditioned play audiometry or pure-tone audiometry and tympanometry, cognitive examination using a non-verbal intelligence test and a speech and language assessment using age-appropriate standardised and validated tests of expressive and receptive language (figure 1). In our study, the Sprachentwicklungstest für drei- bis fünfjährige Kinder (Language Development Test for Three- to Five-year-old Children) is used to test language comprehension in children aged 3.0 to 4.5 years, morphological rule formation for 3.0 to <5.0 year olds and phonological working memory for non-words and syntax for 3.0 to <4.0 year olds. The Sprachstandserhebungstest für Kinder im Alter zwischen 5 und 10 Jahren (Language Assessment Test for Children Aged between 5 and 10 Years) is used to examine the formation of plural rules in 5.5–8.5 year olds. The Test zur Überprüfung des Grammatikverständnisses (Test for Reception of Grammar, German version) is used to test morpho-syntactic comprehension in 4.5–8.5 year olds. Expressive vocabulary is tested with the Aktiver Wortschatztest für 3- bis 5-jährige Kinder—Revision (Active Vocabulary Test—Revision) for 3.0 to <5.5 year olds, and the ‘expressive’ subtest of a short form of the Wortschatz- und Wortfindungstests für 6–10 Jährige (Vocabulary and Word Finding Test for 6–10 year olds) for 5.5 years to 8.5 year olds. The Potsdam-Illinois Test für Psycholinguistische Fähigkeiten (Potsdam-Illinois Test for Psycholinguistic Abilities) subtest for repeating sentences is used to test expressive syntax in 4.0–8.5 year olds. The Mottier Test is used to assess phonological working memory, that is, phonological processing capacity and accuracy, which implies speech sound perception (phoneme differentiation and identification), storage and retrieval through to sound production, as a repetition of sequences of nonsense syllables, in ≥4.0 to 8.5 year olds. Speech sound production (phonology/phonetics) is tested using the Psycholinguistische Analyse kindlicher Aussprachestörungen II (Psycholinguistic Analysis of Children’s Speech Sound Disorders) screening test plus four additional words from the main test with the percentage of correctly-pronounced consonants and consonant conjunctions (PCC) as the target parameter. These speech and language tests are repeated 12 months (test time T1) and 18 months (test time T2) after the start of therapy (figure 1).

Figure 1. Flowchart of the THEON protocol with diagnostic assessments and treatment phases. AWST-R, Aktiver Wortschatztest für 3- bis 5-jährige Kinder—Revision; PLAKSS II, Psycholinguistische Analyse kindlicher Aussprachestörungen II; P-ITPA, Potsdam-Illinois Test für Psycholinguistische Fähigkeiten; SET 5–10, Sprachstandserhebungstest für Kinder im Alter zwischen 5 und 10 Jahren; SETK 3–5, Sprachentwicklungstest für drei- bis fünfjährige Kinder; TROG-D, Test zur Überprüfung des Grammatikverständnisses; WWT, Wortschatz- und Wortfindungstests.

Study procedures

Participating children are randomised into groups for standard, outpatient, individual face-to-face therapy or online, interval group therapy, delivered by qualified speech and language therapists. The children in the standard therapy group receive 30 weekly, individual, face-to-face speech–language therapy sessions of 45 min each. The children in the online interval group receive therapy in groups of 2–3 children, three times a week in 3 therapy blocks (4–3–3 weeks) with a 3-month break between each block, for a total of 30 therapy sessions (figure 1). Thus, participants in both groups undergo the same number of therapy sessions. If possible, participants should not undertake any other speech–language therapy between the end of their therapy for this study and timepoint T2 (18 months after therapy initiation). In cases where acute therapeutic needs persist after the initial 30 sessions, treatment of participants from either group is continued as individual face-to-face therapy. The number of these additional sessions is documented and included as a covariate in the statistical analysis. Children who fulfil the inclusion criteria and whose parents have consented to their child taking part in the study will subsequently start treatment.

Sample size

For the primary research question, T values for the test results in the four above-named linguistic domains and auditory phonological working memory (figure 1) are compared between the two intervention groups. We presume that the results of the language tests are uncorrelated and will be 0.3 SD better in IG1 after treatment than in the standard therapy group (IG2). This presumption is supported by the fact that the study situation described in the introduction indicates that interventions with a high therapy frequency in intensive blocks tend to be slightly more effective than treatments with low therapy frequency, as is small group therapy compared with individual therapy. Online therapy is also known to achieve results comparable to, or even better than, those of face-to-face therapy. In order to achieve a probability of 80% with t-tests at the Bonferroni-adjusted significance level of 1.25% (overall significance level 5%) and therefore demonstrate the superiority of the online group intervention in at least one of the language tests (power), a sample size of 96 per group is required. With a dropout rate of approximately 10%, this means that n=106 patients per group should be included. This results in a total sample size of n=212.

Intervention: background and implementation

The study design builds on experience from a preceding RCT (THESES, trial registration number DRKS00023244). That trial compared standard face-to-face individual therapy (45 min, once per week) with identical therapy delivered online, with each group involving 51 children. Online therapy demonstrated significantly greater efficacy than individual face-to-face therapy (d=0.646, p=0.002; aggregated overall language score).⁴³ This effect was largely attributed to the involvement of parents, who, through continuous presence and interaction with both child and therapist, developed a language-promoting communication style—similar to effects reported in other parent-training programmes.⁴⁴

THESES also included an RCT comparing standard individual therapy with face-to-face group therapy involving small groups of 2–3 children (n=51 per group). After 3 months, group therapy outperformed standard individual therapy (d=1.270, p<0.001), with the difference approaching significance 1 year after therapy onset.⁴³ The trial involved children aged 3 years to 6 year and 11 months, corresponding to the age range of the present study.

Findings from THESES informed the current design, combining online and small-group therapy formats. Experience from the earlier trial confirmed that children with heterogeneous disorder profiles can be grouped together effectively, as they tend to motivate one another to speak. The DLD profiles—phonological, semantic-lexical, and morpho-syntactic—are included as covariates in the statistical analyses.

Therapies in both THESES and the current study follow the Clinical Practice Guideline on Interventions for Developmental Language Delay and Disorders,¹¹ developed in Germany by 24 professional societies and associations, including patient organisations, based on a systematic review. The guideline integrates current evidence on DLD interventions. Based on baseline language assessments, therapy integrates implicit (input enhancement and optimisation, modelling, focused stimulation, conversational recasting, corrective feedback and enhancement) and explicit (elicitation of specific verbal responses and linguistic structures, creation of opportunities for language production, metalinguistic explanation and reflection, and visualisation) methods.⁴⁵ Implicit methods, which enrich the learning context without placing direct demands on the child, are used initially, particularly with younger children; explicit methods targeting specific linguistic structures are applied later and more frequently with older children.

In terms of dose form, qualitative dosage components—such as techniques, tasks, activities and intervention contexts—constitute a therapy unit,¹⁴ within which two to three modules are implemented (eg, interactive picture-book reading, tactile play with modelling clay or board games). Randomisation ensures comparable baseline conditions across groups. Outcome independence from specific therapists is supported by the inclusion of a large number of clinicians who provide treatment according to the above guideline in routine practice, thereby ensuring comparability across intervention groups.

In individual face-to-face therapy, children are engaged through activities such as board games, role play and drawing games. In online group therapy, picture stories and games are shared via screen and actively co-constructed by children and therapists. The adaptive structure accommodates each child’s language-learning goals, enabling them to demonstrate strengths while addressing specific weaknesses. The primary focus is on shared speech and language activities, allowing children with DLD to benefit from peer language models. Individual face-to-face therapy is delivered without parental presence, whereas in online group therapy, a parent remains available in the background for support if required. In both settings, parents are informed about the session content and homework exercises at the end of each session. If children in the online group experience persistent difficulties with online therapy, they are transferred to standard therapy. These changes are documented. Furthermore, therapies that have taken place and those that have been cancelled are documented. If therapy is discontinued and a participant withdraws from the study, the reasons are requested and documented.

Data analysis

The speech and language levels achieved at T1 are measured using the test of language comprehension, expressive syntax and morphology, expressive semantics and lexicon and speech sound production, as well as phonological working memory (figure 1). These data regarding speech and language levels achieved are analysed using descriptive statistics. For each of the speech and language tests, the language level between the two therapy groups will be compared using a linear model with the target variable being language test score (at T1) and the covariates being language test score at baseline (T0), age group, gender and centre. The primary short-term therapy comparison (speech and language test scores—excluding phonological working memory—at T1, primary outcome) will be carried out at a Bonferroni-adjusted significance level of 1.25% (overall significance level 5%). The primary analysis will be performed according to the intention-to-treat principle. Missing values are replaced by suitable imputation methods (eg, multiple imputation). All effects are presented using 95% CIs in order to enable a discussion of clinical relevance and statistical significance. The results will be verified using model-independent descriptive statistics and graphs. The secondary endpoints (process evaluation) are evaluated descriptively. The secondary follow-up therapy comparison (speech and language development test scores at T2, secondary outcomes and composite language development scores) are performed in the same manner as the primary analysis but without Bonferroni correction.

For the process evaluation (see details below), univariate or bivariate descriptive analyses are carried out (quantitative methods). Suitable test procedures, such as χ² tests, single-factor variance analyses, t-tests (independent samples for normally distributed parameters) or non-parametric tests (eg, Mann-Whitney U test), will be used to determine statistical significance.

In order to estimate the natural language development of children without the influence of DLD therapy, we will include a wait-list contrast group from the THESES study in our analysis. This group comprises children with DLD of the same age range and corresponding inclusion criteria as in the present study.⁴³

Process evaluation and economic analysis

The process evaluation will be carried out by CS, SN and AN. The first step involves exploration of attitudes, requirements and barriers from the perspective of parents, speech–language therapists and paediatricians using guided interviews. The results will serve as the basis for the preparation of quantitative questionnaires that aim to survey the adherence to the treatment and its acceptance by children, parents and therapists, as well as the ease of access to it. Based on the results of the RCT and the accompanying survey, recommendations for action to improve the treatment of children with DLD will be developed. For this purpose, the results will be summarised and recommendations will be derived, circulated and discussed in a stakeholder workshop. The recommendations will then be finalised within the project consortium. In the economic analysis, which is also carried out by CS, SN and AN, the resource consumption of the intervention will be assessed for both groups. Data collection will be carried out using questionnaires completed alongside the therapy sessions. The use of resources will be priced using standardised unit costs and the costs will be analysed by calculating and comparing mean values.

Trial monitoring

Three-monthly online meetings of the participating centres are held to assess project progress and discuss upcoming tasks.

Current status of the study

Recruitment began on 1 September 2022. The study will end on 30 September 2026. Treatment interventions are currently being conducted within both groups.

Patient and public involvement

All kindergartens in the study region are offered the opportunity to conduct language screening tests for children within the relevant age group with the option of subsequent study participation. Parents are therefore involved in the recruitment process and can decide whether their child should take part in the language screening tests and, if the inclusion criteria are met, in the study itself.

Parents are also included in the process evaluation, which examines the practicability, satisfaction, acceptability, time requirements and barriers associated with implementing speech–language therapy in both study arms. The parents of ten children are involved in developing a questionnaire for process evaluation through semistructured interviews regarding perceived shortcomings, barriers and obstacles related to the online intervention. All parents are then surveyed using this questionnaire in order to identify facilitating and hindering factors, as well as needs at structural, technical and individual levels for integrating online speech–language therapy into routine care. The aim is to ensure that the perspectives of those affected are incorporated into the resulting recommendations for action to guide the future design of similar interventions.

Parents will also participate in a stakeholder workshop to develop a catalogue of such recommendations and to contribute to a white paper summarising advantages and disadvantages, including potential conflicts of objectives, as a basis for health policy decision-making. The lay dissemination of study results on the project website will likewise reflect the parents’ perspective.

Ethics and dissemination

Ethical issues

The clinical protocol and written informed consent were approved by the Institutional Ethics Review Board of the Medical Association of Westphalia-Lippe and the University of Münster (approval number 2022-282 f-S). All procedures described in the protocol follow the Good Clinical Practice guidelines and the ethical principles described in the current revision of the Declaration of Helsinki.^{46 47} All local legal and regulatory requirements will be fulfilled. The main ethical issues are informed consent of the parents of participants, the use of the THEON programme and the protection of data privacy. All parents addressed will receive detailed information and explanations about the study protocol prior to enrolment. Written consent must be given by at least one parent or legal guardian of each participant (online supplemental material 1). Contact addresses are provided for further questions regarding participation in THEON or in cases of withdrawal. In case of important protocol modifications, the relevant parties are informed in additional online meetings.

Dissemination plan

The main results will be published in a final report, as required by the German Innovation Fund directive. The scientific results will be disseminated via articles submitted to peer-reviewed scientific journals following the International Committee of Medical Journal Editors authorship eligibility guidelines and via presentations at national and international scientific conferences. The individual de-identified participant data (including data dictionary), statistical code and any other materials will be accessible on the homepage of the Deutsches Register Klinischer Studien (German Registry of Clinical Trials) (https://drks.de/search/en/trial/DRKS00030068/entails).