Bob Luttman, Robert Luttman & Associates
Multi-collinearity
Operational Issues: Documentation and Reporting
Traditional Pathway / Variance Documentation
Multi-collinearity
Multi-collinearity is a statistical term for the lack of independence between variances. In other words, whether Variance A occurs depends somewhat on whether Variance B occurs, and vice versa.
Why is this important?
Because all those neat little statistical tools and analyses you learned in research methods are based on variables being independent.
Now you probably learned some neat things for getting around this issue. But they are NOT something you can do on the back of an envelope (or in Excel) during a variance meeting. They require a higher degree of statistical knowledge and computer power than most people have at their finger tips.
And the purpose of doing variance analysis is not to gainfully employ statisticians; it is to gainfully employ you and your staff in providing excellent and cost effective patient care. To do that they need simple tools that do allow them to make quick decisions about what - and how - to improve.
Why does multi-whatever happen?
The cascade effect (next page) is the biggest reason. You can see why in the following (highly simplified) cardiac surgery process:
As the flowchart indicates, one variance ("Get Patient out of Bed") may cause other variances ("ETT", ... , "Discharge") that are sequentially related to it. Alternatively, some variances have common causes (patient age, co-morbidities, etc.).
