Over the next few weeks I’ll be releasing some dashboards looking at patient safety data in the UK, specifically that gleaned from the National Patient Safety Agency’s NRLS system. The information is made available to the public, and I’ve done some collation and basic further analysis to create the dashboards. The idea was inspired by a dashboard originally created by Carl Plant [edit: site now defunct, unfortunately]. The first board, below, lets you compare stats between ‘clusters’, i.e.groups of organisations within the NHS which fulfil a similar role. The second dashboard to be released will allow you to compare data between the organisations making up each individual cluster, and the third will concentrate on the performance of single organisations over time. There are a few important points to remember when viewing all the dashboards. They can be read below – please do take a look. Finally, the dashboards are definitely best viewed on a large screen rather than mobile devices.

 Click here to view Dashboard 1: Compare Clusters

I don’t intend to give any great analysis of the data, but the comparisons between clusters do seem to back up what I suspect would be most people’s intuition. The number of reports has remained broadly stable per cluster over time (I have no explanation for the PCO – No Inpatient cluster in 2010-03!). The Degree of Harm profiles don’t hold any big surprises either. Average time taken to report is interesting, with Ambulance Services taking noticeably longer than other areas. I should point out that the measure of time taken to report is the median of all the medians presented by the NPSA over all time periods; statistically that might be shaky ground so if I find a better measure I’ll update the dashboard accordingly. The fact that ‘time taken to report’ is calculated based on all data since 2009-09 means that it doesn’t necessarily reveal any recent improvements or drops in performance, but I think is still an interesting indicator of general trends for the last couple of years.

Notes Regarding All Dashboards:

The Data
All the data used to make these dashboards is publicly available on the NPSA’s NRLS website. They have some comments relating to treatment of the data on the front sheet of each workbook; I recommend having a read before drawing too many conclusions from the dashboard data. The information is released on a six-monthly basis, and I’ve consolidated those individual reports into one big spreadsheet. I then spent some time adding bits and working it up to get some of the extra stats that have been included. The file will be made available to download when the final dashboard is released (i.e. when I’ve finished altering it to add more analysis!).

 Dates
The dates given relate to the date of release of the data by the NPSA. For example, ‘2011-09’ refers to the most recently published dataset (this month). Each release contains data for a six-month period.

Changes to Organisations
Over the time period that these data relate to, there have been some changes to the organisations concerned. For example: Welsh LHBs have appeared; some organisations have merged with others;  some organisations have moved from one cluster to another. I’m sure there are plenty of these spread around, and my apologies to the organisations concerned if it makes it harder / impossible to compare your figures with everyone else.

Disclosure / Disclaimer
I work in the NHS, as a manager of a risk management software system for a large acute teaching Trust. Data from this system feeds into the NRLS. This dashboard work has been carried out in my own time, and the work / opinions expressed in this blog post are entirely my own and not those of my employer. While I’ve made some general observations about the data, each organisation is different and there will be a large number of factors affecting any of the organisations listed here. The information presented is as released by the NPSA, and it should be borne in mind that a Trust may have perfectly valid reasons for reporting a very low or high rate of incidents, including IT or staffing issues. These analyses are provided for interest only and may contain errors and omissions.