Hypospadias birth prevalence may be increasing with maternal exposure to endocrine disrupters. Yet hypospadias registers are hindered by data quality concerns. We compare the birth prevalence per thousand male births (BP) and ascertainment of hypospadias in the South East of England between 1st January 1997 to 31st September 1998 in a population-based hypospadias case register (surgeons' register) 731 cases, BP 3.8, (95% confidence interval 3.7–3.9), the National Congenital Anomaly System (NCAS) 645 cases (88% of surgeons' register), BP 3.4, (95% confidence interval 3.29–3.5) and Hospital Episode Statistics (HES) 221 cases (30% of the surgeons register), BP 1.15 (95% confidence interval 1.1–1.2). There were 191,438 male births (National Statistics). We found considerable differences in birth prevalence estimates and ascertainment according to data source. Compared with the Surgeons' register, the HES under-ascertainment of 10% was possibly explained by procedural weaknesses. NCAS's 70% under-ascertainment is explained by exclusion of so-called mild hypospadias (since 1990). This exclusion does not appear warranted since mild and severe hypospadias probably share aetiology and all require surgery. NCAR's utility could be improved if it returned to including “minor” hypospadias. Meanwhile routine data such as HES are potentially suitable for surveillance of this condition. Birth Defects Research (Part A), 2007. © 2007 Wiley-Liss, Inc.