T. Feeney2,3, S. Sanchez1,2, T. Brahmbhatt1,2, R. Schulze1,2, P. Burke1,2, G. Kasotakis1,2, T. Dechert1,2, F. T. Drake2,3 1Boston Medical Center,Section of Acute Care And Trauma Surgery,Boston, MA, USA 2Boston University School of Medicine,Department Of Surgery,Boston, MA, USA 3Boston Medical Center,Section Of Surgical Endocrinology,Boston, MA, USA
Introduction: Emergency General Surgery (EGS) represents a diverse set of operations performed on acutely ill and injured patients. The EGS patient population is diverse and increasingly from less advantaged cohorts, who have a higher likelihood of not speaking English (ENG). However, the effect of primary language on EGS outcomes has not been studied. We aimed to evaluate the effect of non-English primary language on outcomes following EGS operations.
Methods: The New Jersey Statewide Inpatient Database from 2009-2014 was used to evaluate cases identified as representing 80% of the national EGS case volume. Cases were restricted to ages >18, resulted from emergency department admissions, were noted to be emergent or urgent, and were performed between 0-2 days after admission. We evaluated outcomes for Spanish speakers (SPA) speakers and non-English/non-Spanish (NENS) speakers compared to ENG speakers. Outcomes were in-hospital mortality, 7-day readmission, and hospital length of stay (LOS). Generalized linear mixed models were used to account for hierarchy in the data, and logistic and negative binomial regression were used to estimate odds ratios and incident rate ratios respectively. Sensitivity analysis was performed to assess whether missing confounding variables could ablate effect estimates.
Results: 105,174 patients were included. A majority of ENG speakers were white and had private insurance; SPA speakers were younger and had fewer comorbidities. Where differences between SPA and NENS speakers existed, NENS speakers were more like the ENG speaking group. Readmission exhibited high between-hospital variability. Regression results are shown in the results table. After adjusting for operative risk, comorbid conditions, social determinants of health, and intra-hospital correlation, we found that non-English speakers had reduced LOS after appendectomy and lysis of adhesions. However, SPA speakers had an increased LOS following highest risk operations. There was no difference in mortality or short-term readmission in either language group compared to ENG. Sensitivity analysis indicated that the observed decreases in LOS were not robust to unmeasured confounding, while the increase in LOS among SPA speakers in the highest risk procedures was robust to additional confounding.
Conclusion: These data suggest SPA speakers are younger and healthier than ENG speakers, and that NENS speakers are more like ENG speakers than SPA speakers. Primary language does not have an independent impact on readmission or inpatient mortality following EGS operations. The effect of primary language on LOS after EGS depends on the type of operation. Future studies should focus on long term outcomes and generalizability to other regions of the United States.
