OBJECTIVES: To describe how multiple goals theory can be used as a reliable and valid measure (i.e., coding scheme) of the quality of conversations about end-of-life issues. METHODS: We analyzed conversations from 17 conversations in which 68 participants (mean age=51years) played a game that prompted discussion in response to open-ended questions about end-of-life issues. Conversations (mean duration=91min) were audio-recorded and transcribed. Communication quality was assessed by three coders who assigned numeric scores rating how well individuals accomplished task, relational, and identity goals in the conversation. RESULTS: The coding measure, which results in a quantifiable outcome, yielded strong reliability (intra-class correlation range=0.73-0.89 and Cronbach’s alpha range=0.69-0.89 for each of the coded domains) and validity (using multilevel nonlinear modeling, we detected significant variability in scores between games for each of the coded domains, all p-values <0.02). CONCLUSIONS: Our coding scheme provides a theory-based measure of end-of-life conversation quality that is superior to other methods of measuring communication quality. PRACTICE IMPLICATIONS: Our description of the coding method enables researches to adapt and apply this measure to communication interventions in other clinical contexts.