A core social-communication deficit in children with ASD is limited joint attention behaviours which are considered to be a powerful predictor of later language ability. The concept of joint attention has been described as the ‘simultaneous engagement of two or more individuals in mental focus on one and the same external thing’. The systematic review and meta-analysis by Murza et al (2016) «Murza KA, Schwartz JB, Hahs-Vaughn DL ym. Joint at...»1 aimed to provide a quantitative assessment of the efficacy of joint attention interventions in improving joint attention abilities in children with ASD. Meta-analysis procedures were used also to determine the overall effects of several comparisons including treatment type, treatment administrator, intervention characteristics and follow-up.
Six electronic databases were identified and searched from the earliest indexed date through to 28 January 2015. Finally, 12 studies were included in the meta-analysis (the original studies). Two of these studies were continued as a follow-up study and these two additional studies were included in the follow-up analyses. The number of participants in the studies ranged from 11 to 152.
Patients: Participants (N=787) in the original studies were aged 11 to 152 months with an average age of 55 months. All participants were diagnosed with ASD. 404 children were included in the joint attention intervention groups and 383 children in the comparison groups.
Interventions: The different joint attention interventions were defined as: (1) appeared to include at least one aspect of joint attention training and (2) demonstrated a joint attention outcome(s) 3) involved instruction in at least one aspect of joint attention (e.g., sharing attention with others, initiating joint attention, gaze alteration, using gesture to direct attention to object/ entity/event), and (4) had at least one outcome measure of joint attention. There was a major variation between interventions regarding the duration of each session, intensity and length of intervention.
Comparisons: Eight of the 12 original studies (67%) examined a joint attention intervention to a control group, two examined joint attention to a comparison joint attention intervention, one compared joint attention to symbolic play, and one had a combination of comparisons (joint attention versus control and joint attention versus symbolic play).
Outcomes: The joint attention outcomes measured across studieswere diverse and included for example shared attention, joint attention with gaze shift, coordinated joint attention with shared affect, child initiated joint attention, conventional gestures, and child responding to joint attention acts.
In this meta-analysis nine studies implementing a joint attention treatment versus non-treated controls design were aggregated for treatment effect. A random effects analysis of these studies yielded a positive significant effect (Hedges g = .660, 95% CI [.395, .925], p < .001).
Five studies compared joint attention intervention administrated by a parent with non-treated controls. It yielded an overall effect of Hedges g = .678, 95% CI [.313, 1.043], p < .001. Analyses in which the treatment administrator was a non-parent (e.g., clinicians, researchers, teachers) yielded an overall Hedges g effect of .654, 95% CI [.184, 1.124], p = .006.
Analysis of two studies compared joint attention treatment (administrated by a non-parent) with symbolic play intervention. It yielded an overall Hedges g effect of .527, 95% CI [.077, .978], p = .022. Five studies examined the effects of interventions which combined a behavioural, discrete trial training approach with a developmental social interactive approach. An analysis of these studies yielded an overall positive and significant effect with Hedges g = .762, 95% CI [.337, 1.187], p < .001. Comparisons with interventions that used a social interactive approach revealed an overall Hedges g effect of .589, 95% CI [.194, .983], p = .003.
Five studies included follow-up data collected at 4 and 8 weeks, 3 months, 6 months, and 6 and 12 months. An analysis of the follow-up data yielded an overall Hedges g effect of .349, 95% CI [.042, .657], p = .026 providing evidence that there is some lasting impact of the joint attention intervention after the treatment concludes.
Risk of bias: there was a low risk of bias across studies with a majority of the ratings (60%). Only 7% of the judgments were classified as high-risk primarily due to a greater than 20% attrition rate without a description of how missing data was accounted for in the analysis or a lack of outcome assessment blinding. 33% of the possible ratings were designated as ‘unclear risk of bias' in the areas of randomization, allocation concealment and blinding procedures.
The methodological quality of the included studies was generally judged to be acceptable despite the finding that most of the studies were judged to be ‘unclear' in their randomization and allocation concealment process. The unclear category indicates that while the randomization and assignment process was reported, the details included in the studies did not allow for the determination of the rigor by which the authors carried out these tasks.
Summary: The results of this meta-analysis provide strong support for explicit joint attention interventions for young children with ASD. All comparisons resulted in statistically significant effects. Treatment administrator, dosage and design (control or comparison, etc.) characteristics of the studies do not appear to produce significantly different effects.