Monthly herd disease incidence rate or prevalence estimates in 196 Swedish commercial dairy herds from 1988 to 1995 were collected retrospectively from the official milk-recording scheme and merged with county administrative and farmers' data on housing and management. To study the effects of changes in housing system on the occurrence of veterinary-treated foot/leg disorders, clinical mastitis, teat injuries and high milk somatic cell counts (MSCCs), four marginal Poisson or negative-binomial regression models were applied to the data (6011-7063 herd-month records), using the generalized estimating-equations method. Monthly observations were treated as repeated measures within herds. There were significant transitory increases in the incidence of clinical foot/leg disorders when changing from tie-stalls to cubicles and decreases in the incidences of clinical mastitis and teat injuries when changing from tie-stalls to cubicle or straw-yard systems. Effects on foot/leg health generally lasted for <18 months after building finish, while udder-health improvements persisted >18 months. Reductions in the incidence of clinical mastitis were not accompanied by any clear changes in the prevalence of high MSCCs.