Trends in knee magnetic resonance imaging

OBJECTIVE. To assess trends in knee magnetic resonance imaging (MRI) use and accompanying management changes to determine whether indications for this test have changed over time. DATA SOURCES. Large administrative database containing health care information for 587,010 people living in 1 state who were enrolled in the Medicare or Medicaid programs. They all had used health services during 3 consecutive years between 1991 to 1995. METHODS. We performed a retrospective cohort study examining the rates of lower extremity MRI in successive years and calculated the proportion of patients who were seen by a knee specialist or underwent knee surgery subsequent to the MRI. RESULTS. The annual rate of knee MRI was 1.4 per 1000 person-years in 1991 and increased by 140% to 3.4 per 1000 person-years by 1995 (P = 0.001). Approximately half of patients who underwent a knee MRI in 1991 had a diagnosis of internal derangement of the knee in theprior year; this figure dropped to 35% in 1995 (P <0.001). The percentage of patients undergoing a knee MRI who had no record of any knee diagnosis in the prior year grew from 13% in 1991 to 33% in 1995 (P <0.001). Over the 5 years of the study, the percentage of patients whose knee MRI was followed by specialist care or knee surgery decreased from 68% to 58%, a relative change of 15% (P <0.005). CONCLUSION. Knee MRI use increased sharply during the study period, but the proportion of such patients who had a prior diagnosis of internal derangement or subsequently saw a knee specialist or underwent knee surgery decreased. This finding indicates that the criteria for knee MRI appear to have broadened substantially during this period.