This study evaluated the safety performance of two types of climbing lane configurations-pocket and overtaking-that are implemented on South Korean expressways. For the evaluation, empirical Bayes (EB) and turning point analyses were sequentially applied to compare crash occurrences before and after the configuration change from pocket to overtaking types on 30 climbing lanes. First, the EB analysis with a meta-analysis indicated that the overall safety performance was not significantly degraded after the configuration conversion. However, because some sites were associated with substantial increases in crash occurrences, 11 sites were identified with crashes above the upper bound of a confidence interval estimated from the meta-analysis. A turning point analysis was applied to those sites to validate whether the increase in crash occurrences and the configuration conversions were consistent in the temporal trend. The results show that changes in crash occurrence at three study sites did not coincide with the configuration change, suggesting no evidence for causality. This implies that in a comparative analysis, evaluation of time series of crash occurrences is important, in addition to a comparison of crash frequencies before and after the implementation of roadway facilities. This study suggests that the developed method can be used to evaluate the safety performance of implementing a roadway facility and for selecting sites where the facility degraded the safety performance and required further evaluation by field engineers.