Improving patient safety in intensive care units in Michigan

Purpose: The aim of this study was to describe the design and lessons learned from implementing a large-scale patient safety collaborative and the impact of an intervention on teamwork climate in intensive care units (ICUs) across the state of Michigan. Materials and Methods: This study used a collaborative model for improvement involving researchers from the Johns Hopkins University and Michigan Health and Hospital Association. A quality improvement team in each ICU collected and submitted baseline data and implemented quality improvement interventions. Primary outcome measures were improvements in safety culture scores using the Teamwork Climate Scale of the Safety Attitudes Questionnaire (SAQ); 99 ICUs provided baseline SAQ data. Baseline performance for adherence to evidence-based interventions for ventilated patients is also reported. The intervention to improve safety culture was the comprehensive unit-based safety program. The r_wg statistic measures the extent to which there is a group consensus. Results: Overall response rate for the baseline SAQ was 72%. Statistical tests confirmed that teamwork climate scores provided a valid measure of teamwork climate consensus among caregivers in an ICU, mean r_wg was 0.840 (SD = 0.07). Teamwork climate varied significantly among ICUs at baseline (F_{98, 5325} = 5.90, P < .001), ranging from 16% to 92% of caregivers in an ICU reporting good teamwork climate. A subset of 72 ICUs repeated the culture assessment in 2005, and a 2-tailed paired samples t test showed that teamwork climate improved from 2004 to 2005, t(71) = -2.921, P < .005. Adherence to using evidence-based interventions ranged from a mean of 25% for maintaining glucose at 110 mg/dL or less to 89% for stress ulcer prophylaxis. Conclusion: This study describes the first statewide effort to improve patient safety in ICUs. The use of the comprehensive unit-based safety program was associated with significant improvements in safety culture. This collaborative may serve as a model to implement feasible and methodologically rigorous methods to improve and sustain patient safety on a larger scale.