Sir, the letter of Neurath et al. (BDJ 2014; 217: 55) commenting on our paper (BDJ 2014; 216: E10) draws attention to the apparent variation in reporting of dental general anaesthesia which may affect the validity of hospital episode statistics (HES). We addressed this issue quite clearly in our paper and agree that this makes comparison of areas problematic, particularly comparing relatively small individual health economies such as individual NHS Primary Care Trusts or local authorities. Their interpretation of the analyses of HES data published by Robertson, Ní Chaollaí et al.1,2 does not address those studies finding that dental GA provision is substantially under-recorded in HES.

What is more difficult to explain is the difference in rates of hospital admission between larger groupings of areas, such as in our study and, more recently, by Public Health England who found quite dramatic differences between fluoridated and non-fluoridated areas. It seems inherently unlikely that arrangements for reporting dental GA and child hospital admission are systematically different in fluoridated areas compared with non-fluoridated areas, thereby causing a greater degree of under-reporting in the former. If there is, as seems likely, a general degree of under-reporting, then the differences between fluoridated and non-fluoridated populations could be far greater than is currently appreciated.

It is further suggested by Neurath et al. that dental care is possibly more extensive in a fluoridated area or that dental professional behaviour is different. It again seems inherently unlikely that dentists in fluoridated areas are behaving systematically differently to those in non-fluoridated areas regarding referral, or have better clinical skills than their peers in non-fluoridated areas with higher levels of decay. The most likely explanation for the slightly higher care index in young children in the West Midlands (10.6% compared with 8.0%) is the lower prevalence of decay, meaning that what remains is more readily managed by treatment services.

Neurath et al. feel that an individual-level randomised controlled trial (RCT) of water fluoridation using bottled water would not be difficult, although they do concede the difficulty of a community-level water fluoridation RCT. It is hard to see how a bottled-water study could be practically achieved or complied with over a number of years to establish effect across all age groups. However, publication of their proposed methodology would enable the practicalities to be assessed.

The potential use of HES data to assess the impact of water fluoridation schemes needs further investigation but variations in reporting, or variations in dentist behaviour seem unlikely reasons for the magnitude of the differences seen between larger communities served by multiple service providers.