There are a lot of numbers floating around in the organisation and it is easy to get overwhelmed by them, so the first step of data analysis is being clear on what data you have got, what you need to have and what is going to help. Use your problem statement and future state vision to help frame the data and information that you will find helpful.
You can use data in a couple of ways; you can use it to indicate where issues and problems may be and you can use it to test out hypotheses about your challenge. There are multitudes of ways in which you can analyse data so we advocate asking some simple questions to get you started and help you get curious with your data and what it is telling you. Ask questions such as:
- Are there differences in performance at different times of the day / days of the week / week of the month / time of the year? What are these differences? Are there patterns?
- Are there particular times when an issue is really prevalent? What is different or unusual about that time?
- Are there differences between the results of different wards / teams / practitioners? Are there different practices seeking to achieve the same result or the same practice generating different results?
- How much variation is there within the process? What are the causes of it?
- Are there particular categories of issue (e.g. certain type of datix incident or complaint type that are more prevalent)? Which are more prevalent?
- Are there any corresponding things happening in other data you have? (e.g. are 28 day breaches rising at the same point as staffing numbers fall?) Is there a genuine correlation between these two data?
There are some tools we would suggest that can help with being able to see what your data is telling you (literally, by turning numbers into charts and visualisations), including the following:
Run (trend) charts
A tool used with continuous data that helps you see whether there are patterns over time
A special type of bar chart that helps you to focus on components of a problem that are having the biggest effect. Used with discrete or attribute data.
Histograms (frequency plots)
Used either with continuous data or counts of attributes (discrete) data to help see variation and where the majority of measurements are occurring (distribution)
Click on the links to be taken to an external site that explains about each of these tools in much more detail.
We would advise starting off with simple run charts as these can give you some powerful information and allow you to keep track of progress as you move forward into testing solutions.
As you become more skilled at looking at and interpreting data you may look further afield for your analysis tools, but for now use those few simple tools. If you have someone in your team that is knowledgeable about other techniques then feel free to use them if they are helpful in gaining you a better understanding of the challenge.
Finally, don't forget that the data will tell you something about the challenge and the problems you are facing, but it won't tell you everything! You will need to delve into the actual work, where it actually happens to get into the causes of what you are seeing in the data.