Estimating hospital admissions due to rotavirus gastroenteritis from hospital episode statistics

Objectives. To determine the proportion of hospital admissions, in children < 5 years old, coded for intestinal infectious disease or non-infectious gastroenteritis, using ICD-10 codes, that were due to rotavirus infection. To assess how many children admitted with rotavirus gastroenteritis were given the specific ICD-10 code (A080) for this disease. Methods. Sixteen-month prospective, observational study of children < 5 years old, admitted to district general hospital with: acute gastroenteritis (greater than or equal to 3 loose stools/day), proven rotavirus infection and those coded as intestinal infectious disease or noninfectious gastroenteritis. Results. Four hundred and twenty children < 5 years old were admitted with acute gastroenteritis. Rotavirus, was detected in 170 children's stools. Acute rotavirus gastroenteritis accounted for 81/397 (20%) children coded as having non-infectious gastroenteritis and 32/81 (40%) coded for intestinal infectious disease. Only 18 children were coded for rotavirus gastroenteritis. Potentially preventable rotavirus gastroenteritis occurred in 122 children; 78 coded as non-infectious gastroenteritis (20%) and 26 coded for intestinal infectious disease (34%). Conclusions. The proportion of children coded with diarrhoeal diseases and found to have rotavirus is less than previously estimated. Using the specific code for rotavirus infection to estimate hospital admissions would be a gross underestimate. Hospital episode statistics cannot reliably estimate the burden of disease due to rotavirus. (C) 2004 The British Infection Society. Published by Elsevier Ltd. All rights reserved.