Authors' response

We appreciate Bowler and colleagues'1 concerns and recognise that our findings2 are merely exploratory and require replication. We also recognise the sensitivities involved—the individual clinics were not specifically identified because the small number of patients would prevent meaningful deductions, instead we concentrated on the model of care. Nevertheless, individual clinics may feel aggrieved that they are lumped with others.

Bowler and colleagues point out potential weaknesses of our study. They rightly highlight the pernicious effect of social deprivation on survival in cystic fibrosis (CF),3 and we were unable to correct for these factors. There are areas of profound deprivation in the areas served by all three models of care and this undoubtedly warrants further exploration. Although our report contains a relatively small number of patients, this limitation is mitigated by the quality of data as we believe that all children and adolescents diagnosed with CF in South and Mid Wales were included.There are also potential differences generated by different individual lung function equipment being used in different clinics (including some full care patients), although there will be similar concerns over most national registry data reports.4 However, the percentage predicted forced vital capacity (FVC) between models of care are very similar, with only the percentage predicted forced expiratory volume in one second (FEV1; the more important prognostic indicator) being significantly different, suggesting the differences are real.

With the improved survival of patients with CF, better preservation of lung function through childhood and the inherent variability in measurement of FEV1, a large number of patients are needed to detect differences between groups in either observational or interventional studies.5 The standard deviation of our measurements compares favourably with previous reports, yet the numbers in our report are probably on the cusp to determine differences between groups. Subgroup analysis by stratification of patients by age or chronic Pseudomonas aeruginosa infection is thus unhelpful. Despite the title of their letter,6 Dryden and colleagues offer no evidence that ‘local clinics can deliver equitable care’, and we would argue that their small patient group will have insufficient power to exclude clinically important differences compared with their local CF centre.

We do not imply that patients receiving shared care did not see a CF multi-disciplinary team (MDT) on a regular basis—all CF clinics should have an MDT as recommended in the UK CF Trust.7 Rather we distinguish between those who had regular review by a specialist CF centre MDT and those who did not. Although seeing a specialist CF consultant for 1 or 2 h per year may comply with CF Trust standards for shared care, it is questionable whether a meaningful contribution can be made with such limited exposure. At no point do we suggest that patients should move away from local care, rather our findings suggest that specialist care can be delivered close to home and that regular review by a CF centre MDT is essential.