Is vaccination associated with better health in children? A secondary analysis using a large cross-sectional survey, the German KIGSS data file (Kindergesundheitssurvey - Children's Health Survey)

Abstract

We studied the question, whether vaccinated children are healthier than unvaccinated children. We used a large, representative cross-sectional survey of children (n = 16'662) aged 1 to 17 produced by the German Robert-Koch-Insitut (RKI), the children health survey (KIGGS). We compared the health status of n = 202 children without any vaccination with that of 15'344 children with vaccinations, using regression models. The principal outcome variable was a morbidity score, constructed out of the presence or absence of 11 diagnoses, which had been ascertained by doctors via a computer assisted diagnostic interview. We analyzed the influence of sociodemographic factors, previous infections, and, in a final step, of vaccinations on morbidity in a Poisson regression analysis. Secondary outcomes were the single diagnoses (atopic diseases, obstructive bronchitis, pneumonia, otitis media, heart disease, anemia, diabetes, thyroid disease, seizures, scoliosis, migraine), which we analyzed by logistic regression models. Unvaccinated children had a significantly higher likelihood to have contracted pertussis, measles, mumps and rubella. Vaccinated children had a higher likelihood to have contracted salmonella, scarlatina and chickenpox. After controlling for sociodemographic variables and previous diseases, the number of vaccinations received was an additional predictor (b = 0.004, p = .0007) to explain morbidity, i.e. children with more vaccinations had a higher morbidity score. The regression models explain between 0.5% to 8% of the total variance. The vaccinations associated with worse health outcomes are vaccinations against hemophilus influenza, mumps, diphtheria and tetanus. Vaccinations against polio, measles and hepatitis B are associated with beneficial health outcomes. Our data stem from a cross-sectional survey which does not allow for causal inferences and should be studied more carefully in large, three-armed, prospective longitudinal studies. We conclude from our analysis that vaccinated children are not healthier than unvaccinated children.