Monday, 15 August 2011

#NHS stats - what's the real story?

Data on NHS waiting times is available by PCT.  I've just got round to taking a look at them.

You can look at the latest data and an archive and it's quite illuminating as you can look by procedure, by PCT and if you really want by PCT and proceedure together, although since the numbers get small there they become less statistically meaningful.

As you would expect the picture is complex - more so than the national headlines suggest.

I was interested in Berkshire Wet as it is where I live (this PCT covers Reading and other Western parts of Berkshire, as the name suggests!).

The picture seems to be an increase in numbers waiting over 18 weeks locally, which reflects the national picture.  I looked at May 2011 which is the latest data compared with May 2010, to give a fair annual comparison (seasonal factors can affect numbers needing treatment in some care types I'm told).

The local stats that worry me are that there has been an increase from 74 to 176 patients waiting over 18 weeks in Berkshire West.  If I wanted a headline I would say 'Patients waiting over 18 weeks more than doubles' which is obviously true but as a fan of 'More or less' it would perhaps be better to say that those waiting over 18 weeks has increased from 4% to 8.4% which is still pretty dramatic.

Of course the 18 week target was one of the bits of 'red tape' that the Tories abolished when they took power, and the local and national picture shows that when you take away focus on a particular target then that measure will slip.  It seems unfair to me that even though the median wait has fallen locally (but not nationally) that substantially more people are waiting a very long time for treatment.

Note for spreadsheet aficionados:  the figure of 74 is calculated by taking 4% of the 1835 procedures with a known start in May 2010 , there were a further 2 without data.  The figure of 176 in May 2011 is calculated by taking the number of completed pathways completed within 18 weeks 1926 from the total with a known start  - 2102.  There were a further 8 without data in May 2011.  The reason for the difference in methodology is that the data is presented differently.  As a check I calculated the May 2011 number using the same method as May 2010 and got 176.   Unfortunately because the data wasn't available I wasn't able to calculate the May 2010 data by simply taking two numbers from each other.

I would be grateful if a fellow spreadsheet obsessive could try to reverse engineer my calculations!

UPDATE: second link now works,  have corrected one mention of 'West Berkshire' to 'Berkshire West'