Dropkin, G. Journal of Public Health, https://doi.org/10.1093/pubmed/fdx136
Published: 03 November 2017
Click here to view this article
Background
NHS RightCare is an NHS England programme describing itself as ‘a proven approach that delivers better patient outcomes’. It identifies opportunities for savings and quality improvements, comparing each Clinical Commissioning Group (CCG) with the ‘Best 5’ of a fixed set of ‘Similar 10’ CCGs chosen using equally weighted demographic and deprivation indicators. This article tests whether these indicators are sufficient and equal weighting is appropriate, and evaluates significance.
Methods
Robust public data on lung, colorectal, and breast cancer mortality is modelled using the indicators and incidence. Peers chosen using the preferred models are compared with the Similar 10. Confidence intervals are obtained for comparator group averages. RightCare significance is simulated.
Results
Preferred models have unequally weighted covariates. Incidence is the strongest predictor of lung cancer mortality. The ‘Similar 10’ are inappropriate comparators. RightCare significance ignores variability of comparator outcomes, causing 12% Type I errors. Whilst RightCare shows 1842 annual avoidable lung cancer deaths in 80 CCGs, only 168 deaths in 8 CCGs appear exceptional using appropriate peers and CIs.
Conclusion
CCGs cannot expect to match the average performance of the RightCare ‘Best 5’. Until the methodology is examined with data of known quality, claims that RightCare is a ‘proven approach’ are unsubstantiated.