National hospital data for intussusception: Data linkage and retrospective analysis to assess quality and use in vaccine safety surveillance

Publication date: Available online 5 December 2015
Source:Vaccine
Author(s): Lamiya Samad, Mario Cortina-Borja, Alastair G. Sutcliffe, Sean Marven, J. Claire Cameron, Haitham El Bashir, Richard Lynn, Brent Taylor
ObjectivesTo assess the quality of national Hospital Episode Statistics (HES) data for intussusception, and evaluate this routinely collected database for rotavirus vaccine safety surveillance by estimating pre-vaccination trends in intussusception hospitalisation.MethodsData linkage was performed between HES and prospective intussusception data from the British Paediatric Surveillance Unit (BPSU), followed by capture–recapture analysis to verify HES data quality. Inclusion criteria were infants aged less than 12 months and admitted for intussusception to National Health Service (NHS) hospitals in England from March 2008 to March 2009. To estimate pre-vaccination incidence rates of intussusception, we performed a retrospective analysis of HES data. Infants aged less than 12 months and admitted for intussusception to NHS hospitals in England between 1995 and 2009 were included.ResultsData linkage between 254 cases of intussusception identified in HES data and 190 cases reported via the BPSU resulted in 163 cases common to both data sources. Of remaining 91 cases in HES, 37 had confirmed intussusception. HES data accuracy was 78.7% (200 confirmed/254 cases) and completeness for intussusception was 86% (163 matched/190 BPSU cases) compared to 81.5% (163 matched/200 HES cases) for BPSU. A total of 233 (95% CI: 227.4 to 238.8) intussusception cases were estimated for the infant population (2008 to 2009). For retrospective analysis, of 6462 intussusception admissions in HES data (1995 to 2009), 1594 (24.7%) were duplicate admissions. A declining trend in intussusception incidence was observed in the infant population, from 86/100,000 in 1997 to 34/100,000 in 2009 (60% reduction, P<0.001). Cosinor modelling showed an excess of cases among infants in winter and spring (P<0.001, n=4957, 1995 to 2009).ConclusionNational hospital data capture the majority of admissions for intussusception and should be considered for the post-implementation surveillance of rotavirus vaccine safety in England.



from #Medicine via ola Kala on Inoreader http://ift.tt/1TtDiu5
via IFTTT