Objectives The UK CF Registry annual reports include comparisons between centres on key outcomes such as FEV1 using rankings. While illustrating the distribution between centres, they promote the assumption that those with the highest measures provide “better” care. We hypothesised a more scientific approach based on statistical “process control” using funnel plots and adjustment for case-mix may help to identify exceptional CF care services in terms of clinically meaningful outcomes.
Methods We extracted data from annual reviews (2007–2012) on the CF Registry. Our outcomes included FEV1 (% predicted) at 15 years and change in FEV1 between 18 and 21 years. Funnel plots were generated with confidence limits at 2 and 3 standard deviations (SD). Centres with mean values outside these limits are said to display “special cause variation” -variability outside what one would expect. Outcomes were then adjusted for case mix (including gender, genotype, pancreatic sufficiency and socio-economic deprivation) and analysed using funnel plots.
Results 31 paediatric centres provided FEV1 data on 15 year olds between 2007 and 2012. Funnel plots of unadjusted FEV1 (% predicted) showed few centres with evidence of special cause variation (2SD limits). Initial case-mix adjustment reduced the number of centres outside these limits to 3. We also identified 28 adult centres providing sufficient data to calculate change in FEV1 (% predicted) between 18–21 years. While there was some evidence of special cause variation (at 2SD limits) in prior to case-mix adjustment, after adjustment none were outside the 2SD limits. None of the centres were outside the 3SD limits in either analysis.
Conclusion In conclusion the work to-date illustrates that funnel plots can be used to explore potential differences in FEV1 between specialist centres. Case-mix adjustment models should develop into a useful tool for making centre comparisons which can continue to be used by stakeholders. This is early work, however, and we need to bear in mind that by examining outcomes in small populations risk missing true differences due to low statistical power. Further work is required to assess whether any observed differences are due to chance or are related to the care patients receive.