A MEDLINE search of the 1996 and 1997 literature yielded two reviews of the methodology of cost-effectiveness analyses (CEAs) in the field of rheumatology and eight original CEAs evaluating slow-acting antirheumatic agents (two studies), total hip arthroplasty for patients with osteoarthritis (one study), educational intervention for patients with fibromyalgia (one study), interventions for patients with low back pain (three studies), and hormone replacement therapy for women with osteoporosis (one study). Most CEAs relied on modelling of data that were assembled from various sources; however, in one CEA a cost-effectiveness comparison was based on data collected prospectively within a randomized clinical trial. Extensive sensitivity analyses were conducted by a majority of the studies. The CEAs of slow-acting antirheumatic agents provided results that were difficult to interpret given extensive modelling and reliance on expert estimates. Only one CEA was supported by an industry grant. Most of the studies were of high quality and met current methodologic requirements, including the use of discounting, sensitivity analysis, and incremental analysis. However, the major shortcoming of these recent studies continues to be the use of estimated costs instead of prospectively measured resource utilization data presented in terms of separate physical units.