Step 6b - Collecting Data

As mentioned above we need regular, continuous data collection. Your first impression may be that to collect data every week requires a huge amount of work. If you’d like to collect large volumes of data you are more than welcome to, but it is not required. ‘Little and often’ is best.


A good time frame to collect data is weekly. If however, you have the ability to easily collect data more often then that is also great. If you can only get data monthly, you may struggle to get sufficient points to allow a meaningful run chart analysis (unless you have the ability to gather retrospective data from a baseline). You may also find it most appropriate to have one point per ‘event’, regardless of the actual number of days between them – such as theatre list 1, theatre list 2 etc. or outpatient clinic 1, outpatient clinic 2 etc.


Regularity is more important than volume of data. But how much do you need? The answer to that question is not an exact one – “just enough data it takes to tell you what you need to know”. In practice even a handful of patients is enough. If you can review 8 randomly-selected patient notes every Friday, that is massively superior to 40 patient notes once a month. The former situation will give you 5 data points for your run chart, and the latter only 1. If you get what you feel is a meaningful sample by only surveying 8 patients, why do more? It will only take up more time for little benefit.

It might feel uncomfortable collecting such small samples, especially when the mantra in research is to get as large a sample as possible to give sufficient statistical power to the study. But remember in QI the detection of a significant change using run charts comes from the shift rules of the position of points in comparison to the median, NOT the sample size of those individual points.

Methods of Data Collection

Regular data collection is often a sticking point for projects as it can be quite difficult to obtain it in a format that allows run chart analysis. There are many ways to get hold of data:

Electronic systems – usually allow historical data to form a baseline and as most entries are time-stamped, allow easy format into data over time. They also easily give large volumes of data but remember this is a ‘nice-to-have’ and far less important than regularity. Investigation results, EMIS GP records and theatre systems are all examples where data can be pulled from.

Manual collection – it may be that you just need to go out and get the data yourself from the patient notes. This is where having willing team-mates can come in handy, as they can help when you are unable to collect it, such as on leave. You may even be able to introduce something new to the way you and your team works that helps get you the data you want – such as minor adaptations to existing forms/documents. If you need to collect data manually the “little and often” motto becomes very important, you don’t want to have to spend hours doing this, as your own motivation will begin to wane.

Nationally-mandated data – NHS Trusts are required to collect some types of data already such as the SSNAP stroke audit, WHO checklist completion rates etc. Speak to senior staff in your speciality to see what exists already, and if it can be formatted into ‘data over time’ (see the video below).

The slides below are from a workshop session delivered at the TIPSQI Showcase event on 12th June 2019. You may find them useful in helping you decide what data to collect and use in your project.

Making Run Charts

To actually make run charts all you need is the data in a spreadsheet format. See our video which talks you through the steps:

Step 6a

Step 7