Abstract
INTRODUCTION: Lung cancer surgery with a lymph node kit improves patient-level outcomes, but institution-level impact is unproven. METHODS: Using an institutional stepped-wedge implementation study design, we compared lung cancer resection quality between institutions in preimplementation and postimplementation phases of kit deployment and, within implementing institutions, resections without versus with the kit. Benchmarks included rates of nonexamination of lymph nodes, nonexamination of mediastinal lymph nodes, and attainment of American College of Surgeons Operative Standard 5.8. We report institution-level adjusted ORs (aORs) for attaining quality benchmarks. RESULTS: From 2009 to 2020, three preimplementing hospitals had 953 resections; 11 implementing hospitals had 4013 resections, 58% without and 42% with the kit. Quality was better in implementing institutions and with kit cases. Compared with preimplementing institutions, the aOR for nonexamination of lymph nodes was 0.62 (0.49-0.8, p = 0.002), nonexamination of mediastinal lymph nodes was 0.56 (0.47-0.68, p < 0.0001), and attainment of Operative Standard 5.8 was 7.3 (5.6-9.4, p < 0.0001); aORs for kit cases were 0.01 (0.001-0.06), 0.08 (0.06-0.11), and 11.6 (9.9-13.7), respectively (p < 0.0001 for all). Surgical quality was persistently poor in preimplementing institutions but sequentially improved in implementing institutions in parallel with kit adoption. In implementing institutions, resections with the kit had a uniformly high level of quality, whereas nonkit cases had a low level of quality, approximating that of preimplementing institutions. Within implementing institutions, 5-year overall survival was 61% versus 51% after surgery with versus without the kit (p < 0.001). CONCLUSIONS: Surgery with a lymph node specimen collection kit improved institution-level quality of curative-intent lung cancer resection.