Continuous Quality Improvement Initiative

Sampling and Sample Sizes

There are several methods to decide how much data to collect for measurement.  Measurement should speed improvement, not slow it down.

Projects can become caught up in collecting data and measurement, thinking a large volume of data will give the most accurate picture of the process.  This in turn leads to significant delays to making a change, as it takes longer to collect the specific data, especially if it is low volume procedure or process. 

To move improvement projects forward, a team needs enough data to make a sensible judgment about what the next steps should be.  Sampling saves time while accurately tracking a process.

Systematic Sampling is collecting data at fixed, predetermined times or intervals – for example, at the end of each day for three weeks, or every fourth patient.  This method is useful when collecting data on a high volume process, and it should be used when data can be collected over a short period of time.

Block Sampling is a method used to select data from a block of predetermined size.  Instead of measuring at a fixed time or interval, measurements are done sequentially for a limited time or until a pre-designated number of observations are met.  Block sampling is used to establish a picture of measurement that is time or process dependent.

The number of observations needed is the same regardless of what sampling process is selected.  To identify if there are trends in processes over time, the general principle is that 25-30 data points is the smallest sample size the can be reliably used.