Table 1 Selected differences and similarities between the Keystone project and Matching Michigan

1. One cohort

1. Four cohorts (97% of English ICUs), including one pilot

2. Kicked off with 6 weeks of ‘immersion’ weekly teleconferences

2. Kicked off with data collection training

3/Whole-state workshops every six months—1.5 or 2 days (overnight), gradually becoming participant-led

3. Each cohort attended two ‘training events’ (0.5 or 1 day)—data collection and intervention

4. Continuous contact via teleconferences with 100~200

4. Teleconferences only at the beginning; discontinued after poor attendance. Webinars continued, but generally not well attended.

5. 5/6 months getting started with data collection & implementing the comprehensive unit-based safety program and daily goals; then Ventilator Acquired Pneumonia (VAP) and CVC-BSI interventions.

5. Initial period (3-6months according to cohort) of data collection only, then all interventions in any order. No VAP intervention.

6. Interactive web-based data entry tool allowing comparison with others

6. Interactive web-based data entry tool allowing comparison with others

7. Program team asked for infection rates to be reported by infection control practitioners independent of the ICUs.

7. ICUs allowed to determine method of data collection and reporting for themselves. Detailed definitions and guidance provided.

8. Targeted adult ICUs primarily

8. Targeted both adult and paediatric ICUs

9. Led by collaboration between prestigious out of state university and the state hospital association

9. Led by government agency