Usage of rapid slide-culture for streptococci was observed in health centres one year after its introduction and first evaluation. The proportion of culture sore-throat patients returned to the original 55% level after an initial period of enthusiasm. The proportion of antibiotic prescriptions remained unchanged at 40%, and rationality of treatment did not increase. Doctors' decisions were based mostly on clinical assessment, while patients' opinions were unpopular decision aids. Evaluation of new tests should extend beyond specificity and sensitivity to their effect on diagnostic and treatment decisions in actual working environments. An observation done shortly after a method has been introduced may be biased by the enthusiasm for novelty, which could be called the "honey-moon effect".