Health policy analysis and magnetic resonance imaging. The case of the New York State Demonstration Project

Abstract

In the absence of controlled clinical trials, the diffusion of magnetic resonance imaging (MRI) has been driven by market forces and the perceived benefits of this technology. To date, all projective needs for MRI use are based on a consensus impression of a medical panel on the role of MRI for DRG or International Classification of Diseases, Ninth Revision, codes. Since an impression of future utilization is not particularly precise, the focus of The New York State MRI Demonstration Project, which approved the acquisition of MRIs at 14 medical centers in 1983, was to determine the actual use of MRI in a medical setting. In a 3-year period, all sites performed 16,095 MRI examinations with 6647 subjects also receiving computed tomography (CT). The results of this study were as follows: (1) 88% of MRIs performed were of the central nervous system (CNS), (2) low level of utilization in the chest and abdomen reflects both a problem with MRI motion artifacts and the failure of MRI to compete with established diagnostic modalities such as mammography, CT scanning, and ultrasonography, (3) for the CNS 18% (1037/5876) studies were positive on MRI but negative by CT, (4) only 1.4% (n = 85) of cases were lesions detected by CT and missed by MRI, and (5) for 81% of the 4754 examinations, MRI and CT were in agreement. Based on the number of lesions observed, the image contrast, and the overall radiologist's impression, MRI was rated superior to CT in 50-60% of the CNS cases. The projected need, based on this study, is for one MRI per 430,000 population in New York State. Also as newer MR imaging protocols evolve, patient use should increase, with the technical cost per study becoming approximately +250 for a scanner performing 3900 studies per year with a +1 million operating expense. At the present time, the best predictive index of MRI utilization is the need for CNS examinations.