The prevalence rates of Echinococcus multilocularis in foxes (n = 5600) evaluated in several Austrian surveys conducted between 1991 and 2004 were analysed for spatial and temporal differences. Data from early studies (1993-1997) in which the intestinal scraping technique (IST) was utilized were compared with data from recent (1999-2004) investigations, which made use of the shaking in a vessel technique (SVT), and it was assessed whether or not the infection rates of Austrian foxes had increased between the investigated intervals. In total, data from 85 districts are presented and both the retrospective and recent data are available from 39 of these districts. A Bayesian hierarchical model of parasite prevalences is presented which (i) accounts for differences in the sensitivity of IST and SVT, (ii) incorporates spatial auto-correlation between neighbouring districts, (iii) investigates the possibility of a temporal shift in the infection status of foxes, and (iv) quantifies uncertainty at each level of the model. The national average prevalence rates in the mid-1990s and at the turn of the millennium were 2.4% (95% confidence intervals 1.1-4.8) and 3.9% (95% confidence intervals 1.5-8.4) respectively. Above average prevalence rates were observed in the western and the northern parts of the country. Evidence is also presented for a temporal augmentation of the prevalence rates in some districts in the northern and eastern parts of the country. These findings are in concordance with several investigations in other European states where both newly emerged areas and elevated levels of transmission in existing endemic areas have been found. None of the districts investigated here showed significant evidence of a drop in prevalence.