A voluntary risk-based control program on paratuberculosis in dairy cattle was initiated in Denmark in 2006. Cows were categorized as high-risk (antibody-positive at least once within the last 3 tests) or low-risk animals based on the results of 3 to 4 annual milk ELISA detecting Mycobacterium avium ssp. paratuberculosis (MAP)-specific antibodies. High-risk animals require management practices aimed at decreasing calf exposure to MAP-contaminated colostrum and milk, and feces originating from these cows. Moreover, repeated test-positive cows are recommended for slaughter before next calving. The objective was to assess the effect of different management practices on the prevalence of MAP-specific antibodies. A questionnaire on management practices was distributed to 1,261 participating herds in December 2008. A total of 1,092 (87%) herd managers returned the questionnaire. Repeated prevalence data from 1,081 herds were available for a period up to 4.25 yr after the first test round. The changes in the prevalence of MAP-specific antibodies from the start of interventions were assessed using a hierarchical logistic model, where different management practices were assessed: a) culling of repeated test-positive cows, b) separation of high-risk from low-risk cows in calving areas, c) cleaning of calving areas after high-risk cows calved, d) removal of calves born to high-risk dams within 2h after calving, e) use of colostrum for feeding of heifer calves from low-risk cows only, f) use of waste milk for feeding of heifer calves from low-risk cows only, g) herd size, and h) proportion of purchased animals. Multivariable analyses suggested that only the proportion of purchased animals (>15% purchased animals as well as 0 to 15% purchased animals compared with no purchased animals in the herd), culling of repeated test-positive animals, and use of waste milk from specific cow groups influenced the decrease in prevalence of MAP-specific antibodies. The control program has been running for just 4.25 yr, and it is assumed that the full effect of the risk-based management practices will only be observed after 4 to 8 yr. Therefore, lack of association between some practices and decrease in prevalence may be a reflection of a short study period. Furthermore, decreases in the prevalence of MAP-specific antibodies may not reflect discontinued transmission of MAP in all age groups.