Studies of total joint arthroplasty (TJA) have not evaluated the costs and outcomes in the context of expected arthritis worsening. Using a cost-consequence approach, the objective was to examine changes in direct health care costs and arthritis severity after TJA for hip/knee arthritis compared with contemporaneous changes in matched controls. The research design was a case control study nested in a population-based prospective cohort.

In a population cohort with disabling hip/knee osteoarthritis followed from 1996 to 2003, primary TJA recipients were matched with cohort nonrecipients on age, sex, region of residence, comorbidity, and inflammatory arthritis diagnosis. Pre- and postoperative total and arthritis-attributable direct health care costs, arthritis severity, and general health status were compared for cases and matched controls.

Of 2,109 participants with no prebaseline TJA, 185 cases received a single elective TJA during the follow-up period; of these, 183 cases and controls were successfully matched. Mean age was 71 years, 77.6% were female, 35.5% had >/=2 comorbidities, and 81.5% had >/=2 joints affected. At baseline, controls had less pain and disability and lower total and arthritis-attributable health care costs than cases. After surgery, although overall health care utilization was unchanged, cases experienced significant decreases in arthritis-attributable costs (mean decrease $278 including prescription drugs) and pain and disability (P < 0.0001 for all). Over the same time period, controls experienced a significant increase in total health care costs (mean increase $1978 including prescription drugs, P = 0.04) and no change or worsening of their arthritis status.

Compared with matched controls, arthroplasty is associated with significant reductions in pain, disability, and arthritis-attributable direct costs.