The aims of this study were to estimate the proportion of asymptomatic Kenyan preschool children using anti-malarial drugs, to identify factors associated with chloroquine use, and to assess the validity of frequency of febrile episodes and drug use reported by mothers or carers. Of 318 children studied, 38% (95% confidence interval [CI] = 30-47%]) tested positive for chloroquine or sulfadoxine. Of chloroquine-positive children, 15% had concentrations exceeding the estimated minimum therapeutically effective values. Among those testing negative for sulfadoxine, chloroquine-positive children were more frequently parasitemic (odds ratio = 2.6, 95% CI = 1.3-5.2), and had lower mean hemoglobin concentrations (6.1 g/L, 95% CI = 2.1-10.1) than chloroquine-negative children. Mothers over-reported the frequency of malaria or fever episodes as usually defined in medical studies, and underreported anti-malarial drug use. We conclude that anti-malarials are frequently given for treatment of malaria or malaria-associated illness, rather than prophylactically or for symptoms unrelated to malaria. Questionnaire surveys cannot replace biochemical markers to obtain information on anti-malarial drug use.