Skip Navigation
Funding Opportunities | Search Funded Research Grants and Contracts

IES Grant

Title: Adaptive Response to Intervention (RTI) for Students with ADHD
Center: NCER Year: 2017
Principal Investigator: Pelham, William Awardee: Florida International University
Program: Social, Emotional, and Behavioral Context for Teaching and Learning      [Program Details]
Award Period: 4 years (07/01/2017–06/30/2021) Award Amount: $3,298,918
Type: Efficacy and Replication Award Number: R305A170523
Description:

Co-Principal Investigators: Fabiano, Gregory; Schatz, Nicole

Purpose: The purpose of this project is to investigate the efficacy of adaptive, evidence-based classroom interventions (i.e., Tier 1 and Tier 2 interventions delivered through a Response to Intervention (RTI) framework) for children with attention-deficit/hyperactivity disorder (ADHD). The research team will experimentally evaluate the efficacy of well-developed and evidence-based behavioral interventions within a problem-solving framework such as RTI. The findings from this project will significantly inform practice within school-based behavioral intervention teams.

Project Activities: This study will employ a sequential multiple assignment randomized trial design (SMART). Prior to the beginning of the academic year, students will be randomly assigned to one of two conditions: (1) Business as usual in which children receive whatever sequence of academic supports and interventions their teachers, school, and parents would typically put into place throughout the entire academic year and (2) an RTI approach which begins with Tier 1 classroom-wide management strategies and includes opportunities to add Tier 2 strategies for youth who do not respond to the initial Tier 1 approach.

Products: The products of this project will include intervention procedures and manuals and evidence for the efficacy of an adaptive behavioral Response to Intervention framework for supporting children with ADHD in schools, peer-reviewed publications, and presentations.

Structured Abstract

Setting: This study will take place in elementary schools in Western New York and Southern Florida.

Sample: A total of 300 children, grades 1–5, with ADHD who have not been classified as special education students and who have not been prescribed psychoactive medication will participate.  Half of the participants will be recruited in Florida and half will be recruited in New York. This sample is selected to represent school-age children with ADHD who are at elevated risk for referral for special education services.

Intervention: The intervention is an RTI approach for child behavior, and includes Tier 1 classwide-behavioral management strategies, a Tier 2 Daily Report Card (DRC) that targets the child's behaviors needed to promote successful social and academic outcomes, as well as enhanced RTI for students needing additional behavioral support beyond Tier 2. Enhanced RTI has two arms: (1) DRC enhanced with additional behavior management strategies (e.g., differential attention, delayed punishments to reduce escape-maintained behaviors), or (2) DRC plus the student receives stimulant medication.

Research Design and Methods: In this study, researchers will randomly assign one-third of children to the BAU condition and two-thirds to the RTI approach. Children assigned to the RTI approach who do not respond to the Tier 1 strategies will receive a Tier 2 intervention, namely a daily report card (DRC). Children who demonstrate non-response to the DRC will be randomly assigned to one of two additional treatment arms: (1) enhanced RTI (RTI-E) or (2) stimulant medication. Researchers will assign children evenly across these two conditions. This SMART design has three embedded treatment protocols, or sets of decision rules that together define an adaptive intervention. Those three protocols are as follows: (1) Business as usual (BAU), (2) Tier 1 strategies, followed by Tier 2 strategies in the event of non-response (i.e., a DRC), followed by enhanced RTI in the event of non-response, and (3) Tier 1 strategies, followed by Tier 2 strategies in the event of non-response, followed by medication management in the event of non-response.

Control Condition: The students assigned to the control condition will receive typical services provided to students with ADHD.

Key Measures: Measures of student behavioral functioning include direct observations of behavior in the classroom using the Student-Behavior Teacher-Response observation system, teacher ratings of behavioral behavior on the Disruptive Behavior Disorders Rating Scale, the Impairment Rating Scale, the BASC-2 Behavioral and Emotional Screening System, and the Domain-Specific Improvement Rating. Student academic functioning will be measured using the AIMSweb Progress Monitoring and Response to Intervention System Secondary, and the Academic Performance Rating Scale.  Additional outcome measures will include potential mediators and moderators of treatment response including: office discipline referrals, number and type of parent contacts due to problems in academic or behavioral functioning, teacher adherence to Tier 1 and Tier 2 strategies, and student referrals for special education evaluation.

Data Analytic Strategy: With Phase 1 randomization, the researchers will use regressions with group membership as a predictor to determine the effect of randomizing participants to either (a) the RTI problem-solving framework approach or (b) business as usual. In Phase 3 randomization, the researchers will also use regressions with group membership as a predictor to determine the effect of randomizing non-responders to both Tier 1 classroom management strategies and Tier 2A basic RTI to either (a) enhanced RTI or (b) medication management. To compare the three different embedded protocols, the researchers will use pairwise comparisons. They will use multiple regression methods to test moderation of the main effects of Phase 1 and Phase 3 randomizations. In addition, the research team will test moderators of the effectiveness of the three embedded treatment protocols using Q-learning, a regression-based means of estimating the optimal treatment protocol given a specific pattern of values on moderator variables. In addition, they will use a recently developed machine-learning based approach to moderation analysis in a SMART design.  Researchers will conduct exploratory mediation analyses using a structural equation modeling framework and the joint significance test for the mediated effect.


Back