Rationale, aims and objective:
We previously demonstrated that annual review %FEV1 under-estimates lung health of adults with CF compared to %FEV1 captured during periods of clinical stability. This has implications in the comparisons against registries with encounter-based FEV1, such as the US. It is uncertain whether this bias affects between-centre comparison within the UK. Previous funnel plot analyses have identified variation in annual review %FEV1 according to centre size, hence we investigated whether paired differences between annual review and best %FEV1 also vary according to centre size.
Methods:
This registry analysis included 18 adult CF centres in the UK with ≥80% completeness for best FEV1 data in 2014. Mean discrepancy between annual review and best %FEV1 is a surrogate for the extent by which annual review %FEV1 underestimates lung health; and was plotted against centre size. A Local Polynomial Regression (LOESS) curve was used to explore the relationship between the two variables. An appropriate model is fitted based on the LOESS curve to determine the strength of relationship between discrepancies in %FEV1 and centre size.
Results:
There is an inverted U-shaped relationship between mean discrepancies in %FEV1 and centre size. A regression of the paired mean difference in %FEV1 against centre size showed a significant improvement in the goodness of fit for a quadratic model (R2 = 23.8% for a quadratic model compared with 0.4% for a linear one; p = 0.048 for the quadratic term).
Conclusions:
Annual review %FEV1 under-estimated lung health of adults from small and large centres in the UK to a greater extent compared to medium-sized centres. A plot of %FEV1 against centre size (e.g. funnel plot comparison) would be affected by systematic bias in annual review %FEV1. Therefore, annual review %FEV1 is an unreliable metric to compare health outcomes of adult CF centres within the UK.
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)