Module Components

• Overview

Topics

• Common Parametric and Nonparametric Procedures for Testing Changes over Time
• Quality Improvement: Measurment
• Quality Improvement: Variation
• Run Charts
• Statistical Process Control Charts

NRSG 795: BIOSTATISTICS FOR EVIDENCE-BASED PRACTICE

Module 11: Testing Differences Over Time

Quality Improvement: Variation

Data Variation is Useful for Identifying Areas that Need Improvement:

The identification of variation in structures, processes, and outcomes is one way a QI leader identifies opportunities for action. For example, if the central line infection rate in hospitals A, B, and C are much lower than they are in hospital D then the QI leader further analyze the situation to determine the reasons why. Are there best practices in hospitals A, B, and C that can help guide improvements in clinical practices in hospital D? Are there differences in the population of patients being cared for in hospital D versus hospitals A, B, and C? A QI leader can also use variation in data to explore the differences among clinical practices between the night and day shifts or among different clinicians. Variation is the first clue that improvements may be needed.

Types of variation

In quality improvement initiatives there are two types of variation that may be present in data that is plotted over time.

Common Cause Variation	Special Cause Variation
is random or unassignable variation is due to regualr, natural or ordinary causes results in a "Stable" process that is predictable	non-random variation is due to irregular or unnatural causes that are not inherent in the design process results in an "unstable" process that is not predictable
If the process only shows common cause variation then the appropriate improvement strategy is to change the underlying process	If a process shows special cause variation the the appropriate improvement strategy is to investigate the origin of these special causes.

Visualizing Variation

Charts are a great visual aid that lend themselves to identify normal random variation or fluctuations in data from special cause variation.

• A run chart graphically depicts how the process is running. It can reveal shifts and trends, but it shouldn’t be used to assess process stability.
• A run chart can be turned into a control chart by adding upper and lower control limits. Control charts are used to monitor the stability of the process. In other words, they not only graphically depict the process over time but they allow you to see if the results consistently fall within control limits.
  

Required Readings and Videos:

• Gupta M, Kaplan HC. Improving Quality Improvement in Neonatal-Perinatal Care. Clin Perinatol. 2017 Sep;44(3):xvii-xix. doi: 10.1016/j.clp.2017.06.002. pages 964-5.