The clinical indications for temporal artery biopsy were explored using clinical decision analysis. The analysis indicated that biopsy is most useful in patients with a high likelihood of steroid side effects. It is nearly always useful when the pre-biopsy likelihood of temporal arteritis is low, but loses its value when temporal arteritis is likely on clinical grounds, except when there is a very high probability of steroid side effects. With the low biopsy sensitivity likely to exist in most institutions, the value of biopsy falls off particularly steeply with rising clinical likelihood of temporal arteritis.