These measures tell you about how the process of change is unfolding as your improvement and clinical teams work to comply with the bundles. Are changes being translated into actual practice as you intended them to be?  Process measures will allow you to identify whether you have created a reliable system that follows the timing, sequence, and goals mandated in the Sepsis Bundle.

Outcome Measures

These measures tell you whether changes are actually leading to the improvement you intended. The Surviving Sepsis Campaign  suggests that you set as a goal a 25 percent reduction in overall mortality due to sepsis from the time you begin your work to your specified end date. Faithful implementation of the bundles, combined with an unwavering focus on the above process measures, will help your team achieve this goal at your institution.

Data Collection

A hospital’s improvement team is likely to use one of two methods to collect data for the measures described above. Whichever method is selected, that approach should be maintained from month to month in order to accurately assess the degree of improvement over time. The methods are listed below. 

Concurrent Data Collection

Concurrent data collection is best suited to new improvement teams. That is, once a patient is placed on the hospital’s severe sepsis protocol, data can be abstracted from the patient chart in real-time or, as most teams have found, at some point during the first 24 hours of admission so that data collection is semi-concurrent with the patient's admission.  One convenient location for this effort is in the ICU where most patients will presumptively be admitted. There are two important advantages to this approach:

1. Concurrent collection of data serves as a prompt to execute the next phase of the bundles. Therefore, some teams may choose to begin the collection in the ED to encourage compliance. 
2. Concurrent collection of data allows teams to segment their population carefully so that they focus their initial efforts on patients for whom they are most likely to succeed. For example, an improvement team may wish initially to segment their patient population to only those patients on a hospitalist driven service to overcome resistance from multiple private practitioners. Overtime as the institutional culture has matured to understand and accept the protocol, the sphere of care can be expanded.
    

Retrospective Chart Review

Retrospective Chart Review is suitable for advanced improvement teams, or teams that have demonstrated success with concurrent data collection. Using this strategy, teams identify charts for monthly review with the assistance of the health information services department based upon discharge diagnoses such as septicemia (ICD9 038.9 and 038.47), systemic inflammatory response syndrome (ICD9 995.92, 996.64 and 999.3), septic shock (ICD9 785.52), or other appropriate diagnoses. As the clinical protocol is introduced and established, the success or deficiency of the improvement effort should be reflected in the results of the retrospective chart review. Advantages to this approach include:

1. Obtaining a more accurate reflection of the state of sepsis care at the institutional level by reviewing charts coded by reviewers unaware of the protocol.
2. The ability to use sampling to analyze only a portion of the charts coded as above. If there are a large number of charts, teams can select a reasonable sample to analyze, e.g., 20 charts per month.
     
   

Outcome Measures

Mortality Due to Severe Sepsis and Septic Shock 

Process Measures

Timing of Blood Cultures 
Timing of Antibiotics 
Central Venous Pressure Goal 
Central Venous Oxygen Saturation Goal 
Low-Dose Steroid Administration 
Recombinant ActivatedProtein C (rhAPC) Administration 
Glycemic Control Goal 
Inspiratory Plateau Pressure Goal 
Reliability: Sepsis Management Bundle 
Reliability: Sepsis Resuscitation Bundle