A0940
Title: Distributed Cox proportional hazards regression using summary-level information
Authors: Dongdong Li - Harvard Medical School (United States) [presenting]
Abstract: Individual-level data sharing across multiple sites can be infeasible due to privacy and logistical concerns. A general distributed methodology is proposed to fit Cox proportional hazards models without sharing individual-level data in multi-site studies. Inferences are made on the log hazard ratios based on an approximated partial likelihood score function that uses only summary-level statistics. This approach can be applied to stratified and unstratified models, accommodate discrete and continuous exposure variables, and permit the adjustment of multiple covariates. In particular, stratified Cox models can be fitted with only one file transfer of summary-level information. The asymptotic properties of the proposed estimators are derived, and the proposed estimators are compared with the maximum partial likelihood estimators using pooled individual-level data and meta-analysis methods through simulation studies. The proposed method uses a real-world data set to examine the effect of sleeve gastrectomy versus Roux-en-Y gastric bypass on time to first postoperative readmission.
