Digital dermatitis is a widespread cause of lameness in dairy cows. A controlled field trial was conducted in 52 French dairy farms endemically affected by digital dermatitis (DD) to estimate the effect of treatment and management practices on DD incidence. The farms were quasi-randomly allocated to 1 of 4 treatment regimens for 6 months: no collective treatment (Control), walk-through footbath during 4 consecutive milkings every 4 weeks (FB/4W) or every 2 weeks (FB/2W) and collective spraying during 2 milkings fortnightly (CS/2W). DD was scored on all lactating cows during milking 7 times every 4 weeks by 14 trained investigators. During these farm visits, data related to cow characteristics and farm management also were collected. A cox frailty model including time-varying covariates was used to estimate, on the 5,598 feet free of lesions, the effect of potential risk factors and treatment practices on the time until the first case of an active DD lesion. A high initial DD prevalence strongly increased the hazard for DD occurence (RR=2.1), as well as absence of hoof-trimming (RR=1.8), and poor leg cleanliness (RR=2.4). Prim'Holstein (RR=2.0) and high-productive cows (RR=1.3) were found to be more likely to develop DD. Collective treatments tended to decreased hazard for DD occurence only when applied at least every fortnight (RR=0.7), compared to the control group. These results confirm the multi-factorial character of DD and provide useful data to design prevention programs.