Management of suspected giant cell arteritis: a decision analysis.

We used decision analysis to compare 4 strategies for managing suspected giant cell arteritis (GCA): treat no patients, treat all patients, perform temporal artery biopsy and only treat positive cases, or biopsy and treat all cases irrespective of the result. The "treat no patients" strategy is preferred only if the likelihood of disease is less than 2%. The "biopsy and treat positive cases" strategy predominates for intermediate likelihood of disease (50%) and is insensitive to altering the probability of any key variables. The "biopsy but treat irrespective of the result" strategy becomes preferred at likelihood of disease of 81%. However, this result was sensitive to alteration of the key variables. Empirical therapy without biopsy would only be preferred if the likelihood of disease exceeded 90% and the disutility one attaches to treating under uncertainty is low at high probability of disease. Management of suspected GCA should be guided by the results of temporal artery biopsy in the majority of cases. At high likelihood of disease, temporal artery biopsy may be of value independent of its usefulness in determining management strategy. Further research should be directed towards improving the accuracy of diagnosis, defining the risk of iatrogenic complications and estimating utilities for the outcomes.