Developing a robotic colorectal cancer surgery program: understanding institutional and individual learning curves

Abstract

<div class="section"> <a class="named-anchor" id="S1"> <!-- named anchor --> </a> <h5 class="section-title" id="d8877351e159">Importance</h5> <p id="P1">Robotic colorectal resection continues to gain in popularity. However, limited data are available regarding how surgeons gain competency and institutions develop programs. </p> </div><div class="section"> <a class="named-anchor" id="S2"> <!-- named anchor --> </a> <h5 class="section-title" id="d8877351e164">Objective</h5> <p id="P2">To determine the number of cases required for establishing a robotic colorectal cancer surgery program. </p> </div><div class="section"> <a class="named-anchor" id="S3"> <!-- named anchor --> </a> <h5 class="section-title" id="d8877351e169">Design</h5> <p id="P3">Retrospective review.</p> </div><div class="section"> <a class="named-anchor" id="S4"> <!-- named anchor --> </a> <h5 class="section-title" id="d8877351e174">Setting</h5> <p id="P4">Cancer center.</p> </div><div class="section"> <a class="named-anchor" id="S5"> <!-- named anchor --> </a> <h5 class="section-title" id="d8877351e179">Patients</h5> <p id="P5">We reviewed 418 robotic-assisted resections for colorectal adenocarcinoma from January 1, 2009, to December 31, 2014, by surgeons at a single institution. The individual surgeon’s and institutional learning curve were examined. The earliest adopter, Surgeon 1, had the highest volume. Surgeons 2–4 were later adopters. Surgeon 5 joined the group with robotic experience. </p> </div><div class="section"> <a class="named-anchor" id="S6"> <!-- named anchor --> </a> <h5 class="section-title" id="d8877351e184">Interventions</h5> <p id="P6">A cumulative summation technique (CUSUM) was used to construct learning curves and define the number of cases required for the initial learning phase. Perioperative variables were analyzed across learning phases. </p> </div><div class="section"> <a class="named-anchor" id="S7"> <!-- named anchor --> </a> <h5 class="section-title" id="d8877351e189">Main outcome measure</h5> <p id="P7">Case numbers for each stage of the learning curve.</p> </div><div class="section"> <a class="named-anchor" id="S8"> <!-- named anchor --> </a> <h5 class="section-title" id="d8877351e194">Results</h5> <p id="P8">The earliest adopter, Surgeon 1, performed 203 cases. CUSUM analysis of surgeons’ experience defined three learning phases, the first requiring 74 cases. Later adopters required 23–30 cases for their initial learning phase. For Surgeon 1, operative time decreased from 250 to 213.6 min from phase 1–3 ( <i>P</i> = 0.008), with no significant changes in intraoperative complication or leak rate. For Surgeons 2–4, operative time decreased from 418 to 361.9 min across the two phases ( <i>P</i> = 0.004). Their intraoperative complication rate decreased from 7.8 to 0 % ( <i>P</i> = 0.03); the leak rate was not significantly different (9.1 vs. 1.5 %, <i>P</i> = 0.07), though it may be underpowered given the small number of events. </p> </div><div class="section"> <a class="named-anchor" id="S9"> <!-- named anchor --> </a> <h5 class="section-title" id="d8877351e211">Conclusions</h5> <p id="P9">Our data suggest that establishing a robotic colorectal cancer surgery program requires approximately 75 cases. Once a program is well established, the learning curve is shorter and surgeons require fewer cases (25–30) to reach proficiency. These data suggest that the institutional learning curve extends beyond a single surgeon’s learning experience. </p> </div>