At this point you will have identified a system you wish to improve, brought together a team, formulated an aim, created a family of measures and then generated a series of interventions you would like to test. Before running into the plan-do-study-act cycle, take a moment to consider the data collection that will be required and begin to collect a baseline measurement. Using this data, we will be able to generate information and knowledge through inquiry and analysis.
Before embarking on the collection of data, take some time to consider what data you will need and how you will collect it. The goal is to spend as little time collecting data as possible while ensuring you have enough to inform yourself and others that your change is having an effect (positive or otherwise). Think about why you need the data and what you will do with it once you have it. Satisfy yourself that you can answer the following questions:
- What is your aim?
- Which change will you test?
- How will you test your change?
- Where, on whom and at what scale will you test your change?
- How, where and with whom will you review the findings?
Data Collection Plan
With the task at hand fresh in your mind, you can now establish a collection plan. Prevent wasting time by outlining the equipment you may need and creating a deliberate and specific collection plan. Your goal is to collect useful data, not perfect data. In the ideal world, you would integrate measurement into the daily routine as much as possible. If your plan addresses the following points, you are ready to begin collecting.
- How did you arrive at this numeric goal?
- Who determined the goal?
- When will it be achieved by?
- What are you trying to evaluate?
- What measures make the most sense for this purpose? Can you define what is the numerator and what is the denominator? What should be included or excluded from the measure?
- What is your definition for the terms used within your measure? (i.e. how would you define a hospital acquired pneumonia?)
- What are the specific data sources? Is the data available?
- Who will collect it?
- How will they collect it? When will they start? How often do they collect it? When will they stop? How much data is needed?
- Where are you starting from? How do you know?
- How will you determine the amount of baseline data needed?
- Over what timeframe will you collect the data?
Recall that your aim is to collect only enough data to make an informed decision. A small sample enables you to reach this point and helps you to learn efficiently through better use of your time. Rather than collecting more data, you will be able to redirect energy and efforts to improving processes. If a change is not working, rather than collect more data, spend your time understanding why it didn’t work and what could be done to make it better. This is made easier if data is collected at the end of the day so you can connect with those individuals who have done the work that day. Even better, let the people who are performing the work collect the information. This way they will be part of the project and they will own it. There are three ways you can sample:
Simple Random Sampling
Select data from a sample of the population using a random process (i.e. random number generator).
Proportional Stratified Random Sampling
Divide the population into separate categories and take a random sample from each category. Your samples should be proportionate in size depending upon the size of the stratum.
Read more from NHS Improvement on data collection.
Uses of Data
Data comes in many forms. It can be qualitative, quantitative or a combination of the two. It can be used for a variety of purposes, not only to inform others of the effect your change is having. Using data in more than one way helps produce a compelling argument for the adoption of your successful changes. Data can be used in the following ways:
- Understanding variation in a process.
- Monitoring processes over time (is it consistent or is it changing).
- For use as a reference point (this can be used as a discussion point).
- For use as an accurate basis for prediction.
- Understanding the effect of a change we test.
Stratifying data is a method of classifying and separating data according to specific variables. It helps identify specific patterns within the data which may not have been apparent from the data as a whole. Use stratification to understand the causal factors at work and identify opportunities for improvement. It will often help you answer questions you may have about who, when or where is best to target an intervention. Try stratifying your data by time, organisational unit, demographics or location.