Robotic-assisted versus laparoscopic hysterectomy in Stage I endometrial cancer: a real-world causal mediation analysis of operating time and length of hospital stay

This study estimated the impact of robotic-assisted surgery (RAS) versus laparoscopy (LPS) on operating time (OT), length of hospital stay (LOHS), and examined mediation through perioperative factors. A secondary analysis of individual patient data from a retrospective multicenter study was conducted. Women with histologically confirmed stage I endometrial cancer (EC) undergoing hysterectomy between January 2010 and December 2023 were included. After inverse probability of treatment weighting (IPTW), baseline covariates achieved standardized mean differences below 0.10. The IPTW-weighted effect estimate of LPS relative to RAS, interpreted as an adjusted association under the assumption of no unmeasured confounding was, 31.4 min for OT (95% CI: 17.8, 46.2) and 0.93 days for LOHS (95% CI: 0.59, 1.34). Blood loss accounted for 11.3 min (36%) of the OT difference and 0.46 days (49%) of the LOHS difference, while other mediators contributed minimally. Direct effects not explained by the measured mediators accounted for 18.6 min (59%) of the OT difference and 0.43 days (46%) of the LOHS difference. At an annual volume of 100 cases, the IPTW-weighted difference between RAS and LPS corresponded to an estimated difference of approximately 52.4 operating-room hours and 93 inpatient bed-days. RAS was associated with shorter OT and LOHS than LPS after adjustment for measured baseline differences. Reduced intraoperative blood loss explained a substantial proportion of these differences, although a considerable residual effect remained unexplained by measured perioperative mediators. Annual resource utilization difference should be interpreted as an illustrative planning estimate rather than a guaranteed operational saving.