C. Marin1, R. Russi Ervilha4, S. Gao3, L. Cybulski1, J.C. Gooch1,2, A. Weiss1,2 4Federal University of Juiz de Fora, Department Of Pubic Health, Juiz De Fora, MINAS GERAIS, Brazil 1University Of Rochester, Division Of Surgical Oncology, Department Of Surgery, Rochester, NY, USA 2University Of Rochester, Wilmot Cancer Institute, Rochester, NY, USA 3University Of Rochester, Department Of Biostatistics And Computational Biology, Rochester, NY, USA
Introduction: There is ample evidence that same-day discharge following mastectomy for breast cancer is safe and of higher value than hospital admission. Despite this evidence, elderly patients are often admitted following breast surgery due to their age, comorbidities, perceived frailty, or lack of support at home. The aim of this project is to examine trends in length of stay (LOS) for elderly patients undergoing mastectomy and to identify patient factors associated with longer LOS in the elderly population to help expedite discharge after mastectomy in appropriate patients.
Methods: The National Cancer Database was queried for all breast cancer patients aged ≥18 years (y) who underwent mastectomy. We summarized clinical characteristics of the cohort, estimated LOS among all patients, those aged 18-64 y and ≥65 y, and performed multivariable analysis for factors associated with LOS using generalized linear models among the patients aged ≥65 y.
Results: A total of 1,205,297 patients met study criteria, with 34.8% aged ≥65 y (“elderly”), 23.8% with zero-day LOS, and 45.7% with a 1-day LOS. Most patients were female (98.5%), white (82.2%), non-Hispanic (89.5%), lived in a metro area (84.3%), had private insurance (56%), underwent unilateral mastectomy (55.6%), and did not have reconstruction (62.2%). Unadjusted mean LOS for the total population was 1.7 days; 1.81 for patients aged 18-64 y, and 1.51 for patients aged ≥65 y. LOS decreased in all patient groups over time. Focusing on the elderly population only, multivariable analysis revealed the following factors were associated with increased LOS: having Medicare insurance (coefficient [B]=0.027, 95% confidence interval [CI] = 0.006 – 0.049, p=0.014), living in a rural area (B=0.037, 95% CI = 0.019 – 0.055, p<0.001), and undergoing a bilateral mastectomy (B=0.165, 95% CI = 0.158 – 0.172, p<0.001). More recent year of diagnosis (every year p<0.001), not having a 30-day readmission (B=-0.166, 95% CI = -0.176 – -0.156, p<0.001), having a Charlson-Deyo comorbidity index of 0 (B=-0.343, 95% CI = -0.363 – -0.323, p<0.001), and not undergoing reconstruction (B=-0.327, 95% CI = -0.332 – -0.322, p<0.001) were all associated with shorter LOS.
Conclusion: In this large national data set, more recent diagnoses were associated with shorter LOS, suggesting that same-day discharge after mastectomy is becoming more common even among elderly patients. LOS is shorter among elderly patients who have less comorbidities; LOS is longer among elderly patients who live in rural areas and undergo more extensive surgeries (i.e., bilateral versus unilateral mastectomy). Additional support strategies for patients undergoing more extensive surgery, or for patients who live in rural areas should be developed so that all elderly patients can avoid hospital admission and successfully recover at home.