Many organizations involved in health care improvement are faced with the good problem of increasingly rare events, whether they are surgical site infections, ventilator-associated pneumonias, or other adverse events. In such cases, the standard approach of monitoring weekly or monthly rates on a time series chart is not very useful visually or statistically (essentially because these charts tend to plot long sequences of zeros when there are no cases each week or month, with no visible trends or variation patterns).
Instead, measuring the time or number of cases between occurrences, such as the number of surgeries between surgical site infections, is more effective for detecting changes and verifying improvements. To help verify improvements from tests of change, two simple rules and a new control chart have been developed for time-between data. These approaches are simple to use and for rare events have greater statistical sensitivity than standard methods.
This presentation and the related articles (below) illustrate several experiences applying these methods to test improvements in surgical site infections, ventilator-associated pneumonias, and bloodstream infections. These methods also are useful for organizations working on safety, reliability, and other adverse events, as well as the IHI 100,000 Lives Campaign.
