BACKGROUND: Since drug shops play an important role in treatment of fever, introducing rapid diagnostic tests (RDTs) for malaria at drug shops may have the potential of targeting anti-malarial drugs to those with malaria parasites and improve rational drug use. As part of a cluster randomized trial to examine impact on appropriate treatment of malaria in drug shops in Uganda and adherence to current malaria treatment policy guidelines, a survey was conducted to estimate baseline prevalence of, and factors associated with, appropriate treatment of malaria to enable effective design and implementation of the cluster randomized trial. METHODS: A survey was conducted within 20 geographical clusters of drug shops from May to September 2010 in Mukono district, central Uganda. A cluster was defined as a parish representing a cluster of drug shops. Data was collected using two structured questionnaires: a provider questionnaire to capture data on drug shops (n=65) including provider characteristics, knowledge on treatment of malaria, previous training received, type of drugs stocked, reported drug sales, and record keeping practices; and a patient questionnaire to capture data from febrile patients (n=540) exiting drug shops on presenting symptoms, the consultation process, treatment received, and malaria diagnoses. Malaria diagnosis made by drug shop vendors were confirmed by the study team through microscopy examination of a blood slide to ascertain whether appropriate treatment was received. RESULTS: Among febrile patients seen at drug shops, 35% had a positive RDT result and 27% had a positive blood slide. Many patients (55%) had previously sought care from another drug shop prior to this consultation. Three quarters (73%) of all febrile patients seen at drug shops received an anti-malarial, of whom 39% received an ACT and 33% received quinine. The rest received another non-artemisinin monotherapy. Only one third (32%) of patients with a positive blood slide had received treatment with Coartem(R) while 34% of those with a negative blood slide had not received an anti-malarial. Overall appropriate treatment was 34 (95% CI: 28 -- 40) with substantial between-cluster variation, ranging from 1% to 55%. CONCLUSION: In this setting, the proportion of malaria patients receiving appropriate ACT treatment at drug shops was low. This was due to the practice of presumptive treatment, inadequate training on malaria management and lack of knowledge that Coartem(R) was the recommended first-line treatment for malaria. There is urgent need for interventions to improve treatment of malaria at these outlets.