Classification of adverse events occurring in a surgical intensive care unit.

INTRODUCTION: We aimed to describe the preventability and provider specificity of surgical intensive care unit (SICU) deaths and complications compared with those in a cohort of trauma patients. METHODS: Data were collected on all trauma and SICU admissions from July 1, 2001, to June 30, 2004, from administrative (Trauma Base and Project Impact) and morbidity databases. Services were protocol driven and staffed by in-house attendings. Performance improvement assessments were made by consensus. Deaths and complications were classified as preventable, potentially preventable, or nonpreventable, and provider-specific or not. Statistical significance was established at the P < .05 level. RESULTS: One hundred sixty-eight deaths (5.6% rate), 464 procedure-related, and 694 non-procedure-related complications were noted in 2969 SICU patients compared with 166 deaths (3.6% rate), 178 procedure-related, and 261 non-procedure-related complications in 4,655 trauma patients. Thirty-one percent of SICU deaths were preventable/potentially preventable compared with 14% of trauma deaths, but only 1.9% was attributable to the SICU provider. SICU complications were less frequently preventable/potentially preventable than in trauma patients (52% versus 61%) and less often provider-specific (5% versus 19%). CONCLUSIONS: SICU complications are deemed preventable less often than in trauma patients and, if so, infrequently incriminate the SICU provider. Preventable and potentially preventable SICU deaths are rarely attributed to SICU care. These data suggest that SICU performance improvement should focus on systems solutions and pre-SICU care.