During the classification of the useful/not useful comments by the interviewees, we looked for signals that could be used to automatically classify comment usefulness. We observed that useful comments very often trigger changes in close proximity to the comment-highlighted lines (recall that a comment is associated with a highlighted portion of the code) in subsequent iterations. We refer to these comments that appear to induce a change in the code as change triggers. Therefore, automatically identifying these change triggers can provide a valuable feature for classifying the usefulness of review comments.