Objective. We aimed to evaluate whether a cheap and less labour-intens
ive regional implementation strategy for guidelines was sufficient to
change knowledge and behaviour among GPs. The model studied was the im
plementation of anticoagulant therapy to prevent stroke in atrial fibr
illation. Method. The intervention took place in the county of Viborg
(149 GPs), Denmark, with the county of Ringkobing (166 GPs) as control
. A local interdisciplinary steering group modified national college-b
ased guidelines, followed by a regional dissemination and implementati
on strategy. The effect of the intervention was evaluated during a fol
low-up period by a repeated questionnaire and by monitoring prescripti
ons for oral anticoagulants in 1993 and 1995 in the Danish National He
alth Service. Results. Adherence to the guidelines was higher after th
e intervention but, considering secular trends and baseline difference
s, the guidelines had no significant effect. The use of oral anticoagu
lants increased substantially in both counties during the 2-year follo
w-up period, but the difference in relative change between the countie
s was negligible. Adherence to the guidelines could not be predicted b
y any of the reported practice characteristics or attitudes to guideli
nes. Conclusion. Despite solid scientific documentation and regional m
odification to establish ownership of nationally agreed guidelines, th
e impact of guidelines on GPs' knowledge and behaviour was disappointi
ng.