Are appendectomy outcomes in level I trauma centers as good as we think?

Designated trauma centers improve outcomes for severely injured patients. However, major trauma workload can disrupt other care pathways and some patient groups may compete ineffectively for resources with higher priority trauma cases. This study tested the hypothesis that treatment at a higher-level trauma center is an independent predictor for worse outcome after appendectomy.

An observational study was undertaken using an all-payer longitudinal data set (California State Inpatient Database 2007-2011). All patients with an ICD-90-CM diagnosis of "acute appendicitis" (International Classification of Diseases, Ninth Revision, Clinical Modification code 540) that subsequently underwent appendectomy were included. Patients transferred between hospitals were excluded to minimize selection bias. The outcome measures were days to the operating room, length of stay, unplanned 30-d readmission (to any hospital in California), and in-hospital mortality. Logistic and generalized linear regression models were used to adjust for patient- (age, sex, payer status, race, Charlson comorbidity index, weekend admission, and generalized peritonitis) and hospital-level (teaching status and bed size) factors.

There were 119,601 patients treated in 278 individual hospitals. Patients in level I trauma centers (L1TCs) reached the operating room later (predicted mean difference 0.25 d [95% confidence interval 0.14-0.36]), stayed in hospital longer (0.83 d [0.36-1.31]), and had higher adjusted odds of generalized peritonitis (odds ratio 1.63 [95% confidence interval 1.13-2.36]) than those in nontrauma centers. There were no differences in mortality or unplanned 30-d readmissions to hospital; or between level II trauma centers and nontrauma centers across any of the measured outcomes.

Odds of generalized peritonitis are higher and hospital length of stay is longer in L1TCs, although we found no evidence that patients come to serious harm in such institutions. Further work is necessary to determine whether pressure for resources in L1TCs can explain these findings.