ObjectiveTo pilot a single‐patient trials (SPTs) service in general practice, designed to improve decision‐making about long‐term medications for chronic conditions. Design12‐week within‐patient, randomised, double‐blind, placebo‐controlled, crossover comparison of ibuprofen with paracetamol for osteoarthritis, involving three pairs of two‐week treatment periods for each participating patient. Setting and patientsPatients attending an academic general practice with a clinical diagnosis of osteoarthritis, with pain of at least a month's duration severe enough to warrant consideration of long‐term non‐steroidal anti‐inflammatory drug (NSAID) use. Main outcome measuresPain and stiffness; measures of overall arthritis compared with previous fortnight; preference for NSAID at the end of each twoweek treatment period; use of escape analgesia; side effects; and management changes as a result of the SPTs. ResultsEight of 14 patients completed SPTs. One was a clear responder to NSAIDs, five were non‐responders, and two were indefinite. Of the five who were using NSAIDs before the SPT, two continued and three ceased using them. Clinically useful information assisted decision‐making for all eight participants. Medication management changed for six. ConclusionsSingle‐patient trials can be successfully implemented in general practice and might be a valuable method for GPs to identify patients who respond to medication for chronic stable conditions such as osteoarthritis, in which individual response to medication is variable.