BACKGROUND: A non-blinded, cluster randomised controlled trial to test whether footwear reduces prevalence and intensity of hookworm infection in school-aged children on Pemba Island, Zanzibar. METHODS: Six schools were randomised to receive shoes and standard care (deworming with albendazole and health education) and six control schools to receive standard care only. Children provided a stool sample to assess prevalence and intensity of infection with soil transmitted helminthiases (Ascaris lumbricoides, Trichuris trichiura and hookworm). Shoes were then distributed to pupils in the intervention schools; deworming took place as part of the government's mass drug administration programme and a further round of stool samples were collected six months later. RESULTS: Nine hundred and fifteen children were traced at follow-up (1056 at baseline). As many children wore shoes in the control arm as the intervention arm. There was no difference in hookworm prevalence (23.4% for intervention schools, 21.3% for control schools, p=0.48), and no difference in mean hookworm infection in eggs/gram of stool (18, 1-36 in intervention schools, 18, 7-29 in control schools, p=0.23). Shoe-wearing increased across all schools, from 47.4 to 82.4%. If a child wore shoes at the end of the study, the relative risk of hookworm infection was 0.7 (CI 0.53-0.91). CONCLUSION: Due to contamination, the trial could not conclude that shoes were protective against hookworm infection but the intervention led to behavioural change, and observational data suggest that shoes are protective against hookworm. This trial is registered at ClinicalTrials.gov, NCT01869127.