Measles, mumps, and rubella vaccine coverage in 2 year old children in East Lancashire--better than it looks.

A study population of 1850 children resident in East Lancashire born between 2 July and 1 October 1994 was obtained from local child health information systems (CHIS) and family health service (FHS) general practitioner registration data in March 1997 to determine the accuracy of reported measles, mumps, and rubella (MMR) vaccine coverage in 2 year old children registered with East Lancashire general practitioners. The reported MMR immunisation coverage was 89.7%, but the observed coverage was higher at 95.5% (95% confidence interval (CI) 93.9-97.3%). Small practices (3 Pounds GPs) achieved significantly higher MMR coverage (97.1%; 95% CI 95.8-98.0%) than large practices (93.5%; 95% CI 91.5-95.1%). Significantly higher MMR coverage was found in practices that used CHIS recall systems (96.2%; 95% CI 95.1-97.1%) than those using their own recall methods (91.2%; 95% CI 87.0-94.4%). Logistic regression showed that MMR vaccine coverage was independently predicted by practice size (odds ratio (OR) 2.5; 95% CI 1.5-4.0), recall method (OR 2.3; 95% CI 1.4-3.8), and relative deprivation (OR 1.7; 95% CI 1.0-2.6). Actual MMR coverage in 2 year old children in East Lancashire was significantly higher than reported, mainly because of inaccuracies in the CHIS database. Methods to improve the transfer of data on immunisation from practices to population databases should be explored.