Table 3

Summary of complementary information given by the combination of funnel plots and how this might be used to inform quality improvement strategies

		Position of hospital in funnel plot around the median DNT
		Within control limits	Above upper control limit
Position of hospital in funnel plot with proportion above 90th percentile	Within control limits	No immediate actions needed.	Performance on DNT might be improved by targeting processes affecting all patients.
	Above upper control limit	Performance on DNT might be improved by investigating which factors are causing delays in the subpopulation of patients with substantial delay. This might result in further reduction of median DNT.	Improvement strategy depends on whether the shape of the DNT distribution is: (1) similar as nationwide or (2) skewed and stretched towards substantial delays. The first suggests the entire distribution may have shifted to higher values and that improvement on DNT might be obtained by targeting processes of all patients. The second suggests that the subpopulation of patients with substantial delay may be causing the higher median and that these patients should be targeted to examine which factors might be causing the delay to improve performance on DNT.