Baylor University Medical Center, Dallas, TX
Resident Training in Code Blue Execution in a Simulation Lab Improves Immediate Post-Code Survival

Abstract

Background: Internal medicine residents at teaching institutions often lead emergency resuscitation attempts without formal instruction in the practical elements of leading and executing a code blue in the hospital setting. Simulation training has been shown to improve resident comfort, but a mortality benefit has been established only in the pediatric population.

Methods: We implemented a simulation-based code blue training program with a 3G SimMan involving 21 internal medicine residents who were given lectures about roles/responsibilities and exposed to progressively more challenging code scenarios in which ACLS was implemented. Faculty provided feedback after each session. An internal review of code blue data was conducted comparing code-related outcomes during the 10-month intervention with a 12-month historical control. Primary outcomes were immediate post-code survival and survival to discharge. Secondary outcomes included post-code change to DNR status and post-code withdrawal of life-sustaining care.

Results: Of 287 emergency resuscitation attempts in the 22-month study period, 107 were control codes (8.9 per month) and 180 were intervention codes (16.4 per month). No statistical significance was noted between the groups with respect to age, gender, race, number of night codes, or number of weekend codes. The hospital census was stable during the study period. The Mortality Probability Model II was calculated for every patient. Mean scores were 0.323 (control) and 0.343 (intervention) (P=0.460). Primary analysis showed a trend toward increased immediate post-code survival in the intervention cohort: 72 control (67.3%) vs 128 intervention (71.1%) patients (P=0.496). This trend did not translate to increased survival to discharge: 25 control (23.4%) vs 40 intervention (22.2%) patients (P=0.823). Secondary analysis revealed a significant increase (P=0.013) in the number of patients in whom life-sustaining care was withdrawn after successful resuscitation between the control group (29 patients, 40.3%) and the intervention group (75 patients, 58.6%). No difference was found in the number of patients who changed to DNR code status after successful resuscitation (P=0.594).

Conclusions: Formal simulation-based code training of internal medicine residents may increase immediate post-code survival of adult inpatients. The improvement in our study was not statistically significant, possibly due to insufficient power. No improvement was seen in survival to discharge, although the rates in both groups are in the top decile of national hospitals and may reflect the ceiling for adult resuscitation mortality outcomes. The statistically significant increase in post-code withdrawal of life-sustaining care may reflect increased resident comfort in discussing end-of-life issues with patients' family members. Potential weaknesses of the study include insufficient power, lack of measured resuscitation-centered endpoints, no simulation training of ancillary staff, and observational bias.