The capacity of heel inserts of different plastic and rubber materials to attenuate heel-strike forces was analysed. The heel-strike response was recorded by accelerometers fastened on to the skin. Thirteen rheumatoid patients with a knee arthroplasty and five healthy controls were tested. Inlay material of 6 mm thickness did not significantly reduce the tibial deceleration produced by heel-strike. However, shoes with heels of polyurethane plastic of 20 mm thickness significantly lowered the amplitude of the tibial response compared with rubber-heeled shoes. The possible implications of this for patients with degenerative joint diseases or with joint implants is discussed.