Usefulness of seroconversion rates for comparing infection pressures between countries

Salmonella is a frequent cause of foodborne illness. However, since most symptomatic cases are not diagnosed, the true infection pressure is unknown. Furthermore, national surveillance systems have different sensitivities that limit inter-country comparisons. We have used recently developed methods for translating measurements of Salmonella antibodies into estimates of seroincidence: the frequency of infections including asymptomatic cases. This methodology was applied to cross-sectional collections of serum samples obtained from the general healthy population in three European countries. Denmark and The Netherlands had the lowest seroincidence (84,169 infections/1000 person-years), whereas Poland had the highest seroincidence (547/1000 person-years). A Bayesian method for obtaining incidence rate ratios was developed; this showed a 6·3 (95% credibility interval 3·3-12·5) higher incidence in Poland than in Denmark which demonstrates that this methodology has a wider applicability for studies of surveillance systems and evaluation of control programmes.