Learning curve for robotic surgery in children: a systematic review of outcomes and fellowship programs

The learning curve (LC) of a new technique is fundamental for its application and assessment, and for the training. Literature was analysed to define the LCs of different surgical procedures and the impact of fellowship programs. MEDLINE, EMBASE and paediatric surgical journals’ databases from January 1995 to December 2018 were systematically analysed. Two independent residents sought for articles providing description of robotic-assisted procedures’ LCs in paediatric age/population. Seventeen articles were selected, describing LC of robotic-assisted pyeloplasty (n = 9), fundoplication (n = 4), cholecystectomy (n = 2), choledochal cyst resection (n = 1) and lingual tonsillectomy (n = 1), with 721 procedures. Ten studies refer to one single surgeon; six to more than one; one does not specify the number of operators. Eleven studies are unicentric retrospective, two multicentric retrospective, three prospective and one is a comparative analysis between a retrospective case series and a prospective cohort. The most recruited parameter is operative time alone in 3 articles, associated with complications in 12, length of hospital stay in 6, blood loss in 3, resolution in 4 and narcotic use in 2. The LC is described as impacting procedural planning (n = 17), training (n = 9) and economic costs (n = 2). To date, operative time is the most reported outcome to measure LC and proficiency. Efforts are needed to consider measures of surgical expertise and patient status. Robotic training should be standardized on targeted programs planned upon well-defined LCs.