How to Analyse and Present Data

Introduction

Data analysis is a process of examining and presenting useful information. Run charts are the preferred charts to use for projects using the model for improvement which have collected quantitative data. There are other types of charts which may help you to analyse and present your data. This is a high level guide for those who already have an understanding of the content, however more detailed guidance is linked to throughout the document.

Understanding the Different Types of Data

There are two types of data, qualitative and quantitative.

Qualitative data

This is non-numeric data such as written comments in response to an open question, which can be difficult to analyse. If you have free text data you should group comments into themes or categories. You might also want to consider reproducing some comments verbatim in your report if they are particularly pertinent.

Quantitative data

Quantitative data is numerical. The data can be analysed manually, or using software packages and results displayed using tables and charts.

Quantitative data analysis can be summarised using measures of central tendency and dispersion:

A measure of central tendency is a value that describes the data. The three main measures of central tendency are the mean (average), median and mode

Measures of dispersion look at how spread the data is: the range, interquartile and standard deviation.

Averages

Mean, median and mode are the three types of average. It is important to be aware of the differences in results that they can produce. The mean is the value you get when you share everything equally, the mode is the most frequently occurring value and the median is the middle value in a data set.

Mean

The mean is the most common measure of average. To calculate the mean add all the numbers together and divide the total by the amount of numbers. The mean is not the appropriate measure of average in all instances as it is more sensitive to outliers such as exceptionally high or low numbers.

Mode

The mode is the most frequently occurring value.

Median

The median is the mid-point of all the values. To calculate, take the values you are using to create your median and put them in numerical order from lowest to highest. The median is the middle value, or the two middle values added together and divided by 2. If your data is skewed by outliers, as is often the case with clinical audit data, then the median should be used. The median is less affected than the mean by extreme values.

The median is used when creating run charts. The addition of the median line to a run chart enables you to apply the run chart rules. The median line creates a relationship between it and the data, which then allows you to see if improvements are being made through the application of the run chart rules.

Run charts

What are they?

Run charts are commonly used in Quality Improvement. A run chart is a line chart which displays data over time, allowing you to determine if your change is leading to improvement.

The addition of a median line turns your line chart into a run chart. The median should be added once there are 10-12 data points, and should be calculated using all available data points.

Another key element to a run chart is annotations. Annotations tell the story behind the data points. An unusual data point can be annotated to describe the action that was taken that had an impact on the process.

How to interpret them

There are 2 types of variation in a system: Random and Non Random.

Random Variation is inherent in a system, and should be expected.

Non Random Variation can be attributed to something in the system causing the variation. Run charts allow us to determine what type of variation is at play exists within the process we are looking at, and therefore whether any actions should be taken.

Run charts can be analysed using 4 simple rules, known as ‘Run Chart Rules’.

Run charts rules:

A trend: 5 or more data points all going in the same direction. If the data points for 2 or more consecutive points is the same, these points should be counted as one.

A shift: A shift is based on statistical probability, and tells you that something has changed in the process. A shift is 6 or more data points that all fall on the same side of the median. You should be aiming to see shifts in your data that move you towards your overall aim.

An astronomical data point: An astronomical data point is one which anyone looking at the data would describe as unusual, and different from the other data points.

Too few/too many runs: Too few or too many runs is an indication that there is non-random variation within a system. A run is defined as the number of data points on one side of the median, before the values line crosses the median.

NHS Education for Scotland have published a more detailed explanation, including examples.

Hospital at Home