Innovative Research Methods to Test Precision Research Questions

Precision home visiting research borrows concepts from precision medicine research to study what works, for whom, and in what contexts to achieve specific outcomes within early childhood home visiting. It focuses on developing precise and coherent models, identifying and testing active ingredients, engaging strong research-practice partnerships and using more efficient research designs to test impacts. This session will present three examples of innovative research methods that could be used to test precision research questions. The first speaker will introduce the idea of precision home visiting, show examples of precision research questions, and note study designs that may answer these questions. The remaining speakers will briefly discuss how their research could be used to explore precision research questions.

The second presenter will describe a web-based application that guides home visitors through an individualized data-driven decision-making model to help them know 1) when children are at risk for language delay, 2) what intervention strategies to use based on each child's language proficiency, and 3) the effect of those strategies on language outcomes and whether more intensive intervention is needed. Single case design research determined the effectiveness of the individual strategies to address specific language gaps. After the system was developed, two randomized control trials with Early Head Start demonstrated improved language growth for children whose home visitors used the web application compared to those given the same evidence-based strategies without the web-based application.

The third presenter will describe how the study combined participant feedback with secondary data analysis, to design and evaluate a precision-approach to Family Spirit, an evidence-based home visiting program for tribal communities. The study examines whether the program can reach, retain and actively engage more families. Partnering with home visitors and families, the team identified active ingredients of the intervention and data on impact variation to design more personalized pathways through the program to achieve stronger impacts at lower cost.

The final presenter will describe how data integration and machine learning methods could be used to use evidence to examine more individualized, effective strategies to serving families in home visiting. Predictive analytics, which combines the considerable power of machine learning methods and the critical expertise of analysts and practitioners, can be a useful tool for precision home visiting. In impact evaluations, predictive analytics can be used to identify subgroup profiles most likely to benefit from home visiting interventions. In continuous improvement efforts, the methods can be used to identify risk groups that help practitioners identify populations to potentially target and test service effectiveness. When paired with data integration that provides access to a wider variety of information on clients’ backgrounds, predictive analytics can provide new and valuable insights about clients’ risks so that services provided can better aligned with families’ needs.