Reading comprehension intervention in populations with developmental language disorder: A scoping review · Lam JHY, Wang J, Pratt AS et al.

Results

Results included in the final synthesis represent research across seven countries between 1988 and 2022 with half of them conducted in the US (12/24; 48 %). All interventions reported were conducted in English, except one intervention was conducted in Finnish and one intervention was conducted in English with Spanish support. Over half of the studies (13/24, 54 %) were conducted in the last ten years (2013–2023). Studies adopted various terms in characterizing the DLD groups (details described in the section below). DLD was used in the current study for clarity of reporting. Participants with DLD included in these intervention studies ranged from 1 to 394 and the age range was from 3.83 to 14.25 years old.

Participants characteristics

Regarding the age of the participants in the intervention, three of the included studies (13 %) targeted children below age of five (i.e., pre-reading stage). The majority of the included studies included participants from age five to ten (17/24; 71 %), which was the learning to read stage. Four studies (17 %) included adolescents as participants (i.e., reading to learn stage). Regarding the language(s) spoken by the participants, the majority of the studies reported participants were monolingual English speakers or first-language English speakers (21/24; 88 %). Two studies (8 %) included bilingual Spanish-English participants and one study (4 %) included first-language Finnish participants.

Regarding the description of DLD, 14 included studies (58 %) adopted the terms of language impairment or language disorder/deficits/difficulties (LI), followed by specific language impairment (SLI; n = 2; 8 %), weak language skills (n = 5; 21 %), and language-learning disabilities (LLD; n = 2; 8 %). Regarding the identification of the DLD group, 12 of the included studies (50 %) identified participants as DLD with speech-language pathologists’ diagnosis, followed by use of standardized tests (n = 9; 38 %), and use of screening tools (n = 4; 17 %). Note that for some studies, the participants were characterized as having LI or weak language skills, in which some of them were qualified as DLD (Sanabria et al., 2022; Wright et al., 2015). For Staudt (2009), results were only reported from participant Will, as he was the sole participated diagnosed with a language disorder.

Intervention characteristics

Intervention context

The majority of the intervention studies (18/24; 75 %) were conducted at schools, followed by clinics (1/24; 4 %) and participants’ homes (1/24; 4 %). Four studies did not report the setting of the intervention programs. Eight studies (33 %) were conducted by researchers, followed by trained students (5/24; 21 %), teachers (4/24; 17 %), teaching assistants (4/24; 17 %), trained students (4/24; 17 %), computer programs (2/24; 8 %), and SLPs (1/24; 4 %).

Intervention design

Five of the included studies (21 %) adopted a randomized control trial (RCT) design. Twelve included studies (50 %) used group comparison design, and the remaining seven included studies (29 %) adopted a single-case design. For the studies that adopted RCT and group comparison designs (n = 17), eight studies compared a reading intervention group with a waiting control, consultation, or no treatment group, while the other nine studies compared two or three active reading comprehension intervention groups. Thus, there were a total of 37 active reading comprehension intervention programs conducted with participants with DLD. A majority of the interventions (31/37; 84 %) were face-to-face as the service delivery method, whereas the remaining six interventions (16 %) adopted electronic devices to conduct the intervention. Twelve of the interventions (32 %) adopted individual therapy, 11 adopted group therapy (30 %), and nine of them adopted a mix of individual and group therapy (28 %).

Intervention length

Regarding the active reading comprehension intervention programs reported (n = 37), the duration of the interventions ranged from three weeks to a school year (m = 12.24 weeks; SD = 10.52 weeks) and the dosage (total hours) of the interventions ranged from 2.5 h to 83.33 h (m = 22.31 h; SD = 23.11 h). Two to four times of interventions a week and daily intervention were the most commonly adopted dose frequencies (14/37; 38 %), followed by weekly intervention (9/37; 24 %). The intensity of the intervention ranged from 40 to 250 min per week (m = 75.95 min; SD = 62.76 min). See Table 4 for details about intervention length.

Intervention target based on the AVR

The skills targeted in the intervention programs reported can be categorized into four major skills, namly self-regulation (n = 11), decoding (n = 13), vocabulary knowledge (n = 4), and extended discourse (n = 9), with reference to the AVR framework. Only vocabulary knowledge was a primary targeted skill among all skills in the bridging process under the AVR, so vocabulary knowledge was used as a category. Extended discourse was used as a category, replacing language comprehension in the AVR, to specify that both receptive and expressive language were targeted in the interventions. A majority of the intervention programs (22/37; 59 %) targeted multiple skills, while 15 of them (41 %) targeted only one skill throughout the interventions.

Self-regulation.

Table 1 summarizes the interventions targeting self-regulation. One group-designed study focused on executive function skills; six group-designed studies and two single-case studies focused on different strategy use. These interventions recruited older children with DLD. Ritter, Colson et al. (2013) compared the effects of the 48-hour language and reading intervention with a four-hour Interactive Metronome addition, which focused on temporal processing and could improve reading comprehension. Participants were 49 nine-year-old children with DLD, and results showed the Interactive Metronome addition had higher performance on reading rate, reading fluency, and reading comprehension.

There were six group-designed studies and two single-case studies focused on the use of different strategies, which described nine different intervention programs and one replication. Intervention programs taught multiple reading strategies to children with DLD. Wright et al. (2015) implemented a 4-week reading intervention targeting six reading strategies with reciprocal teaching to 10, 13-year-old adolescents with DLD, and results showed a large effect size improvement in reading comprehension. This intervention was then replicated in the same study with 18, 13-year-old adolescents with DLD, and the results showed a large effect size improvement in word reading and reading comprehension. Takala (2006) provided a 5-week classroom-level reading strategies intervention, which included four cognitive reading strategies with reciprocal teaching, to 50 fourth- and sixth-grade Finnish students with DLD, and results showed gains in reading comprehension but did not reach statistical significance. McCartney et al. (2015) implemented a universal reading comprehension intervention for six- to eight-year-old typically developing children and children with DLD, targeting multiple text comprehension strategies in the routine classroom reading curriculum. Results showed a significant improvement in reading comprehension in both groups. Lukin and Estraviz (2010) taught multiple text comprehension strategies during the Reading Recovery program to six children with DLD for 21 weeks and found children reading book levels reached age-matched levels after the intervention.

Imagery training was the focus of two studies with three different intervention programs. Sanabria et al. (2022) adopted the EMBRACE program (Enhanced Moved by Reading to Accelerate Comprehension in English) for 56 second-to-fifth grade Spanish-English bilingual children with DLD, which involved the creation of multimodal simulations, including picture manipulation and imaginary picture manipulation. While the overall intervention effect was not significant, the intervention had a significant effect on comprehension of narrative and easier texts. Joffe et al. (2007) implemented a three-week imagery training program to nine 10-year-old children with DLD, and significant improvement was reported in literal and inferential reading comprehension questions.

There was one study related to the use of think-aloud strategy. Green and Roth (2013) conducted a single case study on intervention focusing on teaching “puzzle question” strategy under think-aloud condition to a fourth-grade student with DLD and found improvement in a standardized reading comprehension test.

A randomized controlled trial focused on a classroom-level intervention on information processing strategies, modifying teachers’ language input and direct vocabulary instruction (Starling et al., 2012). A total of 21 Grade 8 adolescents with DLD participated in the intervention group. Results showed that despite significant differences in reading comprehension between pre-post tests, the intervention group had no significant difference compared to the control group.

Overall, these studies provide early efficacy data on interventions targeting reading comprehension that incorporate self-regulation studies. Generally, pre- to posttest on reading comprehension measures are reported. But there is a need for an RCT to compare the efficacy of different strategies and determine the optimal dosage.

Word recognition.

Table 2 summarizes the interventions targeting word recognition. There were seven group-designed studies and one single-case study focused on decoding skills, phonemic awareness, phonological awareness, letter-sound knowledge, and word reading. The majority of these interventions included younger children. There were two intervention programs primarily focusing on decoding skills. Crowe (2005) compared a 10-hour intervention focusing on decoding and vocabulary with an intervention using communicative reading strategies with eight-to-eleven-year-old students with DLD. Findings indicated that the intervention using communicative reading strategies had significantly higher gains on standardized and non-standardized reading measures. In an earlier study, Crowe (2003) found that participants who received intervention using communicative reading strategies had significantly better reading comprehension performance than the participants who received intervention focusing on decoding and vocabulary and the control group.

There were two studies focused on comparing different intervention programs targeted at phonemic awareness. Loeb et al. (2009) compared three different intervention programs, including Fast ForWord computerized intervention, computer-assisted language intervention, and individualized language intervention, with 103 six-to-nine-year-old children with DLD. All the interventions focused on phonemic awareness and language structures with differences in time allocation. There were no significant changes in reading comprehension skills in the maintenance test. Pokorni et al. (2004) compared three different intervention programs focusing on phonemic awareness, including Fast ForWord intervention, Earobics Step 2, and Lindamood Phonemic Sequencing Program, on 54 nine-year-old children with DLD. The Fast ForWord intervention and Earobic Step 2 were computerized programs, whereas the Lindamood Phonemic Sequencing Program was a face-to-face group intervention. Findings indicated no significant difference in reading comprehension between pre- and post-tests.

There were two studies focused on phonological awareness. Ritter, Park et al. (2013) reported a 12-week instruction of phonological awareness and sound-symbol correspondence to 34 students in Grades 1 to 3 with DLD, and the participants had significantly higher gains in word decoding and read comprehension than the control group (n = 30). Bowyer-Crane et al. (2008) compared an intervention focusing on phonological awareness, letter-sound knowledge, and reading with an oral intervention focusing on expressive language, listening skills, vocabulary, and inferencing on 152 6-year-old children with DLD for 20 weeks. There were no significant differences in reading comprehension ability in post-test scores between the two groups.

There was one study that reported intervention programs focused on letter-sound knowledge. Kouri (2016) compared the use of graphophonemic feedback strategies, meaning-based strategies, and combination of phonemic and meaning feedback strategies on reading comprehension in 15 seven-to-nine-year-old children with DLD. It showed that there were no significant differences in reading comprehension between feedback strategy types.

Buck et al. (1988) implemented a home tutoring program focusing on word recognition, print knowledge, and expressive language for eight seven-to-ten-year-old children with DLD for 15 weeks. The results showed significant improvement in reading comprehension.

Overall, there were mixed results on the efficiency of the pre- to post-tests of reading comprehension interventions that incorporate word recognition skills. Decontextualized and computerized programs showed limited changes in reading comprehension ability compared to meaning-based integrated interventions. Further work is needed to investigate the efficacy of phonological awareness and word recognition interventions with larger sample sizes and more rigorous design.

Vocabulary knowledge.

Table 3 summarizes the interventions targeting vocabulary knowledge. There were four single-case studies focused on building vocabulary knowledge as the primary skill. Two focused on keyword knowledge, whereas the other two focused on the morphology of vocabulary. Rousseau et al. (1993) provided a 10-week intervention focusing on discussion of the meaning of 10 key words and listening preview (pointing to the words while listening) of the narrative passages to five 12-year-old Spanish-English bilingual adolescents with DLD and found that the combination of both conditions yielded higher performance than the baseline period. Staudt (2009) reported an intervention on expanding vocabulary knowledge through building connections between vocabulary and understanding multiple meanings of vocabulary during poem reading to a fourth-grade child with DLD and found improvement in word reading, reading speed, and reading comprehension after a school year.

Two interventions on morphology targeted at building up understanding of vocabulary. Katz and Carlisle (2009) provided a combination of derivational morphological analysis and context analysis in a 12-week intervention to three fourth-grade students with DLD and found small to large effect size improvement in reading comprehension. A 6-year-old child with a history of DLD received a 7-week both derivational and inflectional morphological awareness intervention on 12 targeted affixes and showed gains in reading comprehension but did not reach clinically relevant levels (McLeod & Apel, 2015).

Overall, there were improvements and early efficiency on the pre- to post-tests of reading comprehension interventions focusing on vocabulary knowledge. There is a need for group design with a larger sample size to confirm the efficacy of these interventions.

Extended discourse.

Table 4 summarizes the interventions targeting extended discourse. In addition to the five reading intervention programs mentioned in the self-regulation and decoding sections, which served as active comparison groups, there were three other studies reporting three different reading comprehension intervention programs focusing on extended discourse. These studies investigated extended discourse intervention on reading comprehension in young children. Fricke et al. (2013, 2017) reported a 30-week Nuffield Early Language Intervention program focusing on narrative skills, vocabulary, active listening, and expressive language to 4-year-old children with DLD. Fricke et al. (2013) reported a significant generalization effect of the intervention on reading comprehension in the maintenance test, whereas Fricke et al. (2017) reported no significant difference in reading comprehension in the maintenance test compared to the waiting control group at Grade 1. Fricke et al. (2017) also included a 20-week Nuffield Early Language Intervention program for 5-year-old children with DLD at kindergarten and found no significant difference in reading comprehension compared to the 30-week intervention group and the waiting control group at Grade 1. Revees et al. (2019) implemented Talk Boost KS2, which focuses on active listening, vocabulary, narrative skills, conversational skills, and problem-solving, to 162 Grades 3 to 4 students with DLD and no significant difference in reading comprehension in post-test compared to the waiting control group.

Overall, there were limited findings regarding how extended discourse interventions affect reading comprehension. There is a need to investigate the potential difference in procedure (i.e., the order of technique delivery) given the wide scope of the intervention programs.

Reading comprehension outcome characteristics

Nineteen of the 24 included studies (79 %) adopted standardized reading comprehension tests to measure intervention outcomes. Four studies (17 %) used experimental reading comprehension tasks to measure reading outcomes in which all involved reading narrative text(s) and answering literal and inference questions verbally. None of the experimental reading comprehension tasks provided reliability information. One study adopted reading book level as the reading comprehension intervention measure.

Risk of bias assessment

Sixteen of the included studies (67 %) reported measures in ensuring treatment fidelity, including observation and feedback, interrater review, checklist review, etc. Three different risk-of-bias assessments were used to evaluate the included studies based on their intervention design (Supplementary Material 2 shows the full results). Quality Assessment of Controlled Intervention Studies was used for 14 studies, including RCT studies and studies with a control group. Most studies had a similar baseline across groups (13/14; 87%), low dropout rate (i.e., less than 20 %, 14/14; 100%), and valid and reliable outcome measures (11/14; 79 %). However, only half of the studies had indicated blinded assessment (7/14), and only two of them (14 %) reported power analysis to justify the effect size. Four of five RCT studies included (80 %) provided adequate information on randomization method and concealed treatment allocation.

Quality Assessment Tool for Before-After (Pre-Post) Studies With No Control Group was adopted for the three included group design studies. All of them had low dropout rates (i.e., less than 20 %), and statistical tests were used to report the pre- and post-treatment change. Two of them (66.67 %) had a clear definitions of inclusion criteria and valid and reliable outcome measures. However, none of them reported on blind assessment and power analysis to justify the effect size.

The Quality Assessment Tool for Case Series Studies was adopted for the seven included single-case studies. All of them had clear research questions. Over half of them had a clearly described intervention (5/7; 71%), valid and reliable outcome measures (5/7; 71 %), comparable subjects (4/7; 57 %), and well-described results (4/7; 57 %). However, only two of them (29 %) had adequate length of follow-up.