Towards semi-automatic monitoring of delivery behavioral interventions

Abstract

Traditional methods for monitoring implementation of evidence-based programs require labor-intensive quality assessments. These assessments generally involve human observation, hence becoming a major bottleneck in the monitoring of implementation. We present the development of computer-based methods for implementation quality measurement. Firstly, we describe 7 principles of automatization necessary tosystematically identify critical dimensions for analysis.Secondly, we present an automatic classifier that uses linguistic patterns to classify sessions between high-fidelity versus low-quality. The classifier was trained on an initial corpus of 22 transcripts which were coded for high and low quality by humans and replicated by machine learning. Our goal is to develop efficient methods that reduce the burden of quality monitoring, and provide timely feedback for delivery supervisors.Overall, we want to explore to what extent quality monitoring can be done with automatic methods [1,3,4].

Files

Additional details