# Community-Based Care for Minority Adolescents with ADHD: Improving Fidelity with Machine Learning-Assisted Supervision and Fidelity Feedback

> **NIH NIH R34** · SEATTLE CHILDREN'S HOSPITAL · 2020 · $359,812

## Abstract

Project Summary
 The proposed study will develop and evaluate a technology-assisted supervision protocol to promote
treatment fidelity and more efficient delivery of evidence-based treatment (EBT) for ADHD in under-resourced
community settings. The protocol (LC4S) will be grounded in the principles of measurement-based care. LC4S
will task-shift basic supervision tasks to an existing machine learning tool that analyzes and codes therapy
recordings for fidelity (Lyssn; lyssn.io; Tanana et al., 2019). These data will be automatically integrated into an
established HIPAA-compliant online clinical dashboard that produces performance reports for therapists and
supervisors (Care4; www.care4soft.net). Care4 also will be loaded with online facilitation resources for
therapists, while supervisors will be trained to lead weekly in-person supervision sessions according to best
practices in measurement-based care (Martino et al., 2016; Schoenwald et al., 2009; Webster-Stratton et al.,
2014). This trial will be conducted in three community mental health agencies (Psych Solutions, Inc, Behavioral
Aid Solutions, Jewish Community Services) in Miami-Dade County, FL that serve low income ethnic/racial
minority youth and previously partnered with the research team in R01 MH106587. In Y01 of the project, we
will conduct a stakeholder-focused development process for the LC4S intervention that begins with a thorough
capacity and needs assessment using the Consolidated Framework for Implementation Research (CFIR;
Damschroder et al., 2009) and is guided by the Knowledge to Action (K2A) Implementation Science
Framework (Graham et al., 2006). In this phase, we will hold monthly videoconference meetings between
development teams at each agency, Care4, and the investigative team to iterate the final Care4 interface with
an emphasis user-centered design. In year 1, we also will conduct a small open trial to obtain user feedback
(12 cases, 3 supervisors, and 6 therapists) on LC4S and we will make final adaptations to the technology and
protocol. In year 2, we will conduct a pilot Hybrid Type 3 Implementation- Effectiveness randomized controlled
trial (Curran et al., 2012; N=72 youth; 24 youth, 8 therapists, and 2 supervisors per agency) that will randomly
assign both therapists and youth to the LC4S condition or enhanced supervision as usual (ESAU). Both
supervision conditions will be delivered by endogenous, agency supervisors who are trained in delivery of the
EBT and the core elements of its supervision and have access to monthly phone consultation from experts.
The proximal target in this R34 pilot RCT is session by session fidelity scores. The distal service delivery
outcomes in this trial are: (1) quality of EBT implementation, (2) quantity of EBT delivered in standard
intervention time frame, and (3) speed of delivery (i.e., number of sessions and days to EBT completion) for
each case. In support of a future R01 that measures the impact of LC4S on patient outcomes acro...

## Key facts

- **NIH application ID:** 10112643
- **Project number:** 1R34MH125037-01
- **Recipient organization:** SEATTLE CHILDREN'S HOSPITAL
- **Principal Investigator:** Margaret Harper Sibley
- **Activity code:** R34 (R01, R21, SBIR, etc.)
- **Funding institute:** NIH
- **Fiscal year:** 2020
- **Award amount:** $359,812
- **Award type:** 1
- **Project period:** 2020-09-15 → 2022-08-31

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/10112643

## Citation

> US National Institutes of Health, RePORTER application 10112643, Community-Based Care for Minority Adolescents with ADHD: Improving Fidelity with Machine Learning-Assisted Supervision and Fidelity Feedback (1R34MH125037-01). Retrieved via AI Analytics 2026-05-22 from https://api.ai-analytics.org/grant/nih/10112643. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*