In our previous blog we discussed whether performance can be an indicator of care quality and CQC rating. One of the key findings was that no single measure existed that gave a clear and up to date insight into how an NHS Trust is performing, as a whole. To meet this challenge and support local leaders, Public View have created the ‘Hospital Combined Performance Score’ (HCPS).
This score is made up of 10 key performance indicators spanning quality, operational, finance and workforce domains which is adjusted and aggregated using an adaption of the formula for the decathlon in athletics.
How closely does the Hospital Combined Performance Score match CQC rating?
From reviewing the Public View trend graph with CQC peer groups there is a visible correlation between the Hospital Combined Performance Score and CQC rating, with more than 8,000 points suggesting ‘outstanding’ performance, 7,000 ‘good’, 6,000 ‘requires improvement’ and less than 5,000 ‘inadequate’.
But as discussed in our previous blog, no metric can replace an in-person assessment to determine care quality and a more in-depth investigation of the correlation suggests that this remains the case even with the HCPS algorithm. The graph below shows the ranking chart for the Hospital Combined Performance Score with the columns shown in green representing Trusts rated ‘Good’ by CQC.
This shows that whilst there is a congregation of ‘Good’ rated organisations towards the top (rightmost) part of the graph, there are a few organisations in the lowest quartile, with the second worst HCPS being attributed to an organisation rated ‘Good’. This may suggest a limitation of the HCPS or be an early warning/prediction of potential rating changes to come. Only time will tell.
But exactly how closely does HCPS match CQC rating?
The statistical correlation using the Pearson Correlation Coefficient shows a linear correlation of R2 = 0.446. In other words, differences in the HCPS accounts for 45% of the variance in CQC rating. The box and whiskers graph below show the overlap in HCPS and CQC ratings.
Is 45% good?
If CQC ratings were a purely mechanical process, then the HCPS formula at 45% would have insufficient strength. However, in the fields of social science where attempts are made to predict human behaviour, a 45% correlation would be considered ‘strong’. The reason for this is that humans are much harder to predict than machines with many more complex variables to be taken account of. For more information about correlation strengths click here.
But the HCPS can do much more than just predict CQC rating. By using the HCPS we can answer questions such as;
- Where is the best performing trust in the NHS?
- Which Trust has shown the most improvement?
- Do ‘Integrated’ Trusts perform better than Acute only?
- any many more …
Keep an eye out for blogs to come answering these questions and many more.
To view your organisation’s Hospital Combined Performance Score (HCPS), see how it compares to peers and whether it is improving click here to register for a free trial account.
- Updated Stroke Metrics 15 Oct 2021
- User Interface Updates 11 Oct 2021
- 287 Metrics 07 Oct 2021
- Automated Testing 20 Sep 2021
- New Release 09 Sep 2021
- New User Defined Reports 04 Sep 2021
- NHS Template Errors 26 Aug 2021
- Templates Direct Database Connexion 26 Aug 2021
- Statistical Process Control (XmR) Tool 25 Aug 2021
- Public View Database Service 23 Aug 2021
- User Defined Pages Beta 22 Aug 2021
- New Top Level SPC Icons 18 Aug 2021
- Public View World Wide 14 Aug 2021
- Public View G Cloud 13 Aug 2021
- Public View Database Service 06 Aug 2021
- More SPC Chart Goodness 05 Aug 2021
- A New Blog Site 04 Aug 2021
- Release 1.4, Database Access 29 Jul 2021
- New SPC Charts 02 Jul 2021
- New Maternity Indicators 02 Jul 2021