Studies of the nonpharmacologic treatment of osteoarthritis (OA) have fallen behind that of pharmacologic therapy for a variety of reasons. The design of studies that involve a therapist-delivered physical intervention presents several problems with respect to patient and observer blinding, the beneficial effect of therapist contact alone, and separation of the effects of individual components of the "package" of delivered care. Important developments in the design and reporting of OA trials are discussed. Recent studies have demonstrated the modest but definite benefits of exercise therapy for OA of the knee, delivered either in hospital, primary care, or community settings. The reduction in pain and disability through an arthritis self-management group education program may extend to at least 1 year. Recent data that low intake of micronutrients (vitamins C, E, and D, and beta-carotene) may adversely influence the progression of knee OA and the incidence of hip OA suggest potential avenues for primary and secondary prevention of large joint OA.