We develop and evaluate the Handoff Communication Assessment, using actual handoffs of patient transfers from emergency department to inpatient care. This was an observational qualitative study. We derived a Handoff Communication Assessment tool, using categories from discourse coding described in physician-patient communication, previous handoff research in medicine, health communication, and health systems engineering and pilot data from 3 physician-hospitalist handoffs. The resulting tool consists of 2 typologies, content and language form. We applied the tool to a convenience sample of 15 emergency physician-to-hospitalist handoffs occurring at a community teaching hospital. Using discourse analysis, we assigned utterances into categories and determined the frequency of utterances in each category and by physician role. The tool contains 11 content categories reflecting topics of patient presentation, assessment, and professional environment and 11 language form categories representing information-seeking, information-giving, and information-verifying behaviors. The Handoff Communication Assessment showed good interrater reliability for content (kappa=0.71) and language form (kappa=0.84). We analyzed 742 utterances, which provided the following preliminary findings: emergency physicians talked more during handoffs (67.7% of all utterances) compared with hospitalists (32.3% of all utterances). Content focused on patient presentation (43.6%), professional environment (36%), and assessment (20.3%). Form was mostly information-giving (90.7%) with periodic information-seeking utterances (8.8%) and rarely information-verifying utterances (0.4%). Questions accounted for less than 10% of all utterances. We were able to develop and use the Handoff Communication Assessment to analyze content and structure of handoff communication between emergency physicians and hospitalists at a single center. In this preliminary application of the tool, we found that emergency physician-to-hospitalist handoffs primarily consist of information giving and are not geared toward question-and-answer events. This critical exchange may benefit from ongoing analysis and reformulation.