Trendy! How do we know whether improvement has worked?

The trigger

We have recently tried to strengthen community hospitals to deliver more care locally with the hope to increase turn over and release pressure on secondary care beds. One day we are sitting in a meeting and everybody is rubbing their chins because there is activity data from the community hospitals – the number of patients seen and their length of stay, but it is a squiggly line – it is going up and down a lot. How do we know whether a) the number of patients seen has increased and b) the length of stay is getting better?

The problem

First a small secret: just measuring before and after a change is often not a good idea; the numbers always vary from day to day, week to week or month to month, if you plot them the resulting graph goes up-and-down and all over the place. We need to find a way to work out whether the up-and-down has an actual trend and that trend goes in the right direction.

Part of the solution

To analyse whether there is change in the data you can use a run-chart. If the data is in an Excel or csv format it can be copied into a spreadsheet: For a run-chart you need 15 data points – they could be days, weeks of months, but should be similar i.e. compare Mondays with Mondays, or week and weeks, but not Mondays with Sundays. In a 5 easy steps:

· Put the data into a graph.

· Calculate the median and add it to the graph.

· In a good chart the line will cross the median a few times. If it doesn’t you might need a bit of extra help.

· Look for 6 or more points that are on one side of the median – that is a shift: something has happened. 5 points on one side is good: it is a trend!

· If there is no run: Sorry - nothing happened, you have not achieved change yet.

Here is a quick read-up from our Scottish colleagues.

and here a spreadsheet that will build run-charts automatically for you.

P.S. Sorry: The activity in the community hospitals has not gone up (yet)!

Chris Subbe Consultant Physician Ysbyty Gwynedd

67 views0 comments

© 2018 BCUQI

Privacy statement/GDPR:

 BCUHB is not responsible for external link content nor external Twitter comments.

 Several external organisations' logos are used: they all link to their respective websites.

BCUHB staff should not include any patient identifiable information or use any health board system passwords including nadex passwords when using any areas of this website.

Unless otherwise indicated, the design and content of this site is the property of BCUQI which is part of Betsi Cadwaladr University Health Board. Permission should be sought prior to any use of the design and content from this or other pages contained in this site.

The BCUQI Team (part of the Betsi Cadwaladr University Health Board) may share your personal information within the team and relevant health board staff for the purposes of administration of Members' Area/Forum as well as booking onto events.  By booking a place at events using this site, or applying for membership, you agree to this policy and for your data to be retained indefinitely for the purpose of Quality Improvement work and consent to us contacting you in future. This booking may involve an external provider, in which case their Privacy Policy will apply. You may contact us for more information about any data  or photographs we hold, including asking us to delete these, by contacting us by email (