Investigating discussions on social networks, particularly on sensitive topics such as politics or healthcare, can be challenging. However, understanding the opinions and trends within these debates is crucial for identifying user cohesion on specific topics. In this work, we proposed a multilayer network approach for modeling discussions on a social network. We applied our model to a Twitter dataset containing tweets about COVID19 vaccines. We compared our approach to traditional single network analysis, and found that it is more effective at identifying influential users with larger networks and more interactions with their neighbors.