Previous research on collaboration posits collaboration process as a key factor for team performance. However, it is not fully understood which characteristics of a process make collaboration more efficient. In this research, we investigate the effect of collaboration process patterns on teamwork efficiency (e.g. time cost) in the software development setting. We propose a framework to identify frequent interaction structures referred to as collaboration process patterns and study their impact on the efficiency of software development. For purposes of pattern extraction, we propose an algorithm to extract sub-structures from software development processes stored in a software project tracking system. To analyze the effect of different collaboration process patterns, we conduct an empirical study to examine their correlation with issue resolution time using data from an open source software community. As a result, we identified several collaboration process patterns that are positively (or negatively) correlated with issue resolution time. We also found that this correlation may change with task complexity.