Implementation of daily goals in the ICU reduces length of ICU stay and errors of omission in patient care

Introduction

In the ICU, daily goals have improved the effectiveness of communication within care teams [1]. The study objective was to investigate the effect of explicit daily goals in ICU practice on the length of intensive care stay, length of hospital stay, duration of mechanical ventilation and errors of omission in patient care.

Methods

A multicenter, sequential cohort, pretest-post-test study was performed. During the pretest phase, daily goals were formulated by staff-physicians, evaluated by a research physician and kept blinded for the rest of the ICU team. Subsequently, daily goals were broadly implemented. During the post-test phase, daily goals were routine practice. Patient data were collected one full year before (as baseline reference) and during the pretest and post-test phases. During pretest and post-test, medical charts (random subsets of 10 patients a week) were reviewed for errors of omission, defined as deviations from our care protocols. Patient data were compared with data from nonparticipating academic ICUs, extracted from the National Intensive Care Evaluation database.

Results

Two academic ICUs participated in the study. Patients included: baseline reference n = 3,641; pretest n = 3,401, and post-test 3,426. ICU mean (SD) length of stay (days) differed significantly between the pretest phase 5.7 (2) and the post-test phase 4.3 (2.3), difference 1.4 days (95% CI 1.3/1.5). The post-ICU length of hospital stay decreased in the post-test phase, mean difference 2.3 days (95% CI 1.2/3.4). Median (IQR) duration of mechanical ventilation during pretest was 2 (1 to 4) days and during posttest was 1 (1 to 4) days, P < 0.001. The probability of an error of omission was significant larger in the pretest phase then in the post-test phase, RR 6.5 (5.7/7.4).

Conclusions

The results showed a significant decrease in ICU LOS, duration of mechanical ventilation and length of post-ICU stay following DGF implementation. Also the chance of errors of omission to occur was six times lower, which indicates a significant increase in protocol adherence. The drawback of the study design (historic control group and a pretest and post-test for comparison) was limited by comparing the study results with other academic centers during the same time frame.