Predictors of emergency ventral hernia repair: Targets to improve patient access and guide patient selection for elective repair

Background. Emergency operations are associated with worse outcomes than elective operations. If not repaired electively, ventral hernias are at risk of strangulating and requiring emergency repair. We sought to identify patient- and hospital-level factors associated with emergency ventral hernia repair in a nationally representative, United States sample. Methods. We abstracted data from the 2003-2011 Nationwide Inpatient Sample for adults (>= 18 years) who underwent inpatient ventral hernia repair. Our primary outcome was emergency repair. We assessed differences in patient- and hospital-level factors as possible predictors of emergency repair using multivariable logistic regression. We examined secondary outcomes (mortality, total hospital cost, duration of stay) using multivariable logistic and generalized linear (family gamma; link log) regression. Results. After weighting to the United States population, we included 453,161 adults (39.5% emergency). Independent predictors of emergency repair included payer status (uninsured: odds ratio 3.50, [3.10, 3.96]; Medicaid: 1.29 [1.20, 1.39] compared with private insurance), race/ethnicity (black: 1.77 [1.64, 1.92]; Hispanic: 1.44 [1.28, 1.61] compared with white), age (>= 85 years: 2.23 [2.00, 2.47] compared with <45 years), and comorbidities (Charlson Comorbidity Index >= 3: 1.68 [1.56, 1.80] compared with 0). After risk-adjustment, emergency repair was associated with greater odds of in-hospital death, greater costs, and longer hospital stay. Conclusion. Inpatient ventral hernia repairs are frequently performed emergently, with worse outcomes in this group. Independent predictors of emergency repair include factors that may limit access to and/or selection for an elective operation. These predictors provide targets for interventions to improve access to elective care and inform patient selection with the goal of improving patient outcomes.