Do hospital attributes predict guideline-recommended gastric cancer care in the United States
Department of Surgery
Medical Subject Headings
Adenocarcinoma; Aged; Carcinoma, Signet Ring Cell; Clinical Competence; Female; Gastrectomy; Hospitals; Humans; Lymph Node Excision; Male; Middle Aged; Neoplasm Staging; *Practice Guidelines as Topic; Stomach Neoplasms; Survival Rate; Treatment Outcome
Oncology | Surgery
BACKGROUND: Hospital attributes have been shown to impact short- and long-term outcomes after cancer surgery. However, the effect of hospital attributes on processes of cancer care in terms of delivery of guideline recommended care has not been evaluated. We examined the impact of hospital attributes (volume and type) on guideline-recommended care in patients treated for gastric cancer.
METHODS: We identified patients who were surgically treated for gastric cancer at Commission on Cancer (CoC) hospitals from 2001 to 2006. Patient, tumor, and treatment factors were compared separately by hospital volume and type. Multivariable analyses were used to evaluate the impact of hospital attributes on delivery of guideline recommended gastric cancer care: adequate lymphadenectomy (>/=15 lymph nodes), and adjuvant multimodality therapy (for AJCC Ib-IVM0), controlling for covariates.
RESULTS: More than 1,490 CoC hospitals performed 37,124 gastrectomies. High-volume and teaching CoC hospitals were more likely to treat younger patients, non-whites, patients with lower AJCC stage, and to perform adequate lymphadenectomy than low-volume and community CoC hospitals (p ≤ 0.001). Hospital volume and type, however, were not associated with receipt of adjuvant multimodality therapy. These associations persisted in our multivariable analyses to show that CoC hospital attributes were associated with adequate lymphadenectomy, but marginally predictive of receipt of adjuvant multimodality therapy.
CONCLUSIONS: The strong association between CoC hospital volume or type and guideline-recommended care diminishes after gastric cancer surgery. Variations in referral, insurance, and documentation patterns are potential explanations for these findings. These results highlight some limitations of using hospital attributes as a sole predictor of optimal cancer care.