Variation in lymph node assessment after pancreatic cancer resection: Patient, surgeon, pathologist, or hospital?

Abstract

Background

We sought to define individual contributions at the patient, surgeon, pathologist, and hospital levels on lymph node assessment after pancreatic cancer resection.

Methods

SEER-Medicare beneficiaries who underwent pancreatic cancer resection were identified. Multi-level multivariable regression was performed to assess the proportion of variance explained by patient, surgeon, pathologist, and hospitals on lymph node assessment (≥12 versus <12).

Results

2872 patients underwent pancreaticoduodenectomy by 646 distinct surgeons and 1063 distinct pathologists across 308 hospitals. Patient-related characteristics contributed the most to the variance in adequate lymph node assessment (71.0 ​%). After accounting for all explanatory variables in the full model, 5.5 ​% of the residual provider-level variation was attributed to the pathologist, 35.2 ​% to the surgeon, and 59.3 ​% to the hospital.

Conclusions

Patient-to-patient variation was the greatest underlying contributor to variations in adequate lymph node assessment related to pancreatic cancer surgery. Variation among hospitals was greater than among surgeons or pathologists.