A cluster randomised trial of a classroom communication resource program to change peer attitudes towards children who stutter among grade 7 students

Abstract

BackgroundClassroom-based stuttering intervention addressing negative peer attitudes, perceptions, teasing and bullying of children who stutter (CWS) is required as part of holistic stuttering management because of its occurrence in primary school. This study was conducted in 2017, in 10 primary schools in the Western Cape, South Africa within lower (second and third) and higher (fourth and fifth) quintiles.ObjectivesThe primary objective of this study was to determine treatment effect at six months after intervention of grade 7 participants (Classroom Communication Resource [CCR] intervention versus no CCR) using global Stuttering Resource Outcomes Measure (SROM) scores in school clusters. The secondary objective was to determine grade 7 participant treatment effect on the SROM subscales including Positive Social Distance (PSD), Social Pressure (SP) and Verbal Interaction (VI). The subgroup objective was to determine any difference in the primary outcome between schools between and across quintile clusters (lower and higher).MethodsOnce schools were stratified into lower and higher quintile (which are defined according to geographical location, fee per school and resources) subgroup clusters, schools were assigned randomly to control and intervention groups consisting of grade 7 participants who were typically aged ≥ 11 years. Teachers received 1 h of training before administering the single-dose CCR intervention over a 60–90-min session. The CCR intervention included a social story, role-play and discussion. All participants viewed a video of a CWS and stuttering was defined at baseline. The SROM measured peer attitudes at six months after intervention. Randomisation was stratified by quintile group using a 1:1 allocation ratio. Full blinding was not possible; however, the outcome assessor was partially blinded and the analyst was also blinded. Generalised estimating equations (GEE) was used assuming an exchangeable correlation structure to analyse the data adopting an intention-to-treat principle. Multiple imputation was used to handle missing data. Criterion for statistical significance was set at alpha = 0.05.ResultsTen schools were randomly allocated to control (k = 5) and intervention groups (k = 5), with n = 223 participants allocated to intervention and n = 231 to control groups. A total of 454 participants completed the SROMs in control (n = 231) and intervention (n = 223) groups and were analysed at baseline and six months after intervention. There was no statistically significant difference on the global SROM score (mean difference − 0.11; 95% confidence interval [CI] − 1.56–1.34; p = 0.88). There were also no significant differences on SROM subscales: PSD (mean difference 1.04; 95% CI − 1.02–311; p = 0.32), SP (mean difference − 0.45; 95% CI − 1.22–0.26; p = 0.21) and VI (mean difference 0.05; 95% CI − 1.01–1.11; p = 0.93). Additionally, there was no significant subgroup effect on the global SROM score (lower versus higher quintile subgroups) (interaction p value = 0.52). No harms were noted or repor...

Sensitivity of methods for analyzing continuous outcome from stratified cluster randomized trials – an empirical comparison study

“…The details about the Mallick et al study can be found elsewhere [ 5 , 15 ]. In brief, this was a cluster randomized trial aimed at examining the effect of Classroom Communication Resource (CCR) on peer attitude towards Children Who Stutter (CWS) in South African schools in the Western Cape.…”

“…Over the years, the number of adopting CRTs is increasing [ 2 ]. Diverse types of clusters can be allocated in CRTs including: geographical areas [ 3 ]; health care districts [ 4 ]; and schools [ 5 ]. Like trials on individuals’, most CRTs use one of the following three experimental design strategy such as: (a) completely randomized; (b) matched-pair; or (c) stratified.…”

“…In stratified designs, clusters are randomly allocated to the intervention and control groups within each stratum. For example, Mallick et al [ 5 ] conducted a school-based CRT to investigate the effect of the Classroom Communication Resource (CCR), vs Usual Care, to improve the peer attitude towards children who stutter (CWS). In this trial, schools were first divided into quintile (1–3: lower and 4–5: higher) and stratified as a high vs low school based on the socio-economic resources [ 5 ].…”

“…For stratified designs, we can also assess robustness by comparing the methods with or without adjusted for stratification. The GEE with exchangeable correlation structure was used as the primary method of analysis in the Mallick et al [ 5 ] study.…”

“…In this study, we empirically examined the sensitivity of methods for analyzing continuous outcome from the stratified CRT using the data from the Mallick et al [ 5 ] study, which in turn demonstrated the robustness of the results obtained using the primary GEE method.…”