Introduction: American Society of Clinical Oncology / National Comprehensive Cancer Network quality guidelines recommend the examination of ≥12 lymph nodes (LN) following colorectal cancer (CRC) resection. Compliance with this 12-LN standard may be a function of patient, surgeon, pathologist, or hospital factors. We sought to identify factors associated with adherence to the 12-LN guideline to better inform targeting of quality improvement (QI) efforts.Methods: SEER-Medicare data (1998-2002) were used to identify patients with Medicare Part A and Part B coverage undergoing a first CRC resection. The treating surgeons, pathologists, and hospitals were also identified. Multilevel mixed-effects logistic regression was used to evaluate adherence to the 12-LN standard.Results: Study criteria identified 31,815 eligible patients with median age 76 years, of whom 54% were female. Most patients had colon tumors (80%). CRC surgery was performed by a general surgeon in most cases (79%), and less commonly by a colorectal surgeon (13%) or other specialist (8%). The majority of patients were treated at teaching hospitals (54%), and 23% were treated at ACOSOG-participating hospitals. A median of 10 LN were evaluated per patient, and only 41% of patients had ≥12 LN examined. In the multivariable model (Table), predictors of increased adherence to the 12-LN standard included colon vs. rectal primary, resection by a colorectal vs. general surgeon, treatment at a teaching or ACOSOG hospital, and hospital CRC resection volume. Notably, neither surgeon nor pathologist case volume was associated with 12-LN examination. Of the variation in guideline compliance that was not explained by the factors in the model or by other patient-related factors, 82% was hospital-related, 17% pathologist-related, and 1% surgeon-related.Conclusions: Adherence to the 12-LN standard for CRC resection is poor. Non-compliance is largely attributable to differences among hospitals and pathologists rather than variation among surgeons. QI efforts should be targeted at underperforming hospitals.
Predictors of Adequate Lymph Node Assessment

Variable	Odds Ratio	95% Confidence Interval	P-Value
Age (per decade)	0.94	0.91 — 0.97	0.001
Female vs. male	1.19	1.13 — 1.25	< 0.001
Year of operation (per year)	1.10	1.08 — 1.12	< 0.001
Colon vs. rectum	1.60	1.50 — 1.71	< 0.001
Colorectal vs. general surgeon	1.11	1.00 — 1.22	0.04
Teaching hospital	1.29	1.13 — 1.48	< 0.001
ACOSOG hospital	1.48	1.22 — 1.80	< 0.001
Surgeon case volume (per 10x ↑)	1.07	0.99 — 1.17	0.09
Pathologist case volume (per 10x ↑)	0.98	0.88 — 1.08	0.6
Hospital case volume (per 10x ↑)	1.19	1.03 — 1.37	0.02