Ca. Mustard et al., THE INFLUENCE OF REFERRAL PROTOCOLS ON THE UTILIZATION OF MAGNETIC-RESONANCE-IMAGING - EVIDENCE FROM MANITOBA, Canadian Association of Radiologists journal, 45(2), 1994, pp. 117-123
The relative scarcity of magnetic resonance imaging (MRI) capacity in
most Canadian jurisdictions has created pressures to develop guideline
s on appropriate clinical indications for examination by this method.
The objective of this study was to assess the influence of referral pr
otocols on the utilization of MRI services. Three neuroradiologists an
d one radiologist reviewed the indications for MRI for 198 referrals t
o the facility at St. Boniface General Hospital in Winnipeg (99 from o
utside Manitoba and 99 from within the province), selected at random f
rom patients seen between Jan. 1 and Dec. 31, 1991, for suspected diso
rder of the brain or the spine. Out-of-province referrals had not been
subject to referral protocols, whereas those from within Manitoba had
been subject to such protocols. At least three of the four radiologis
ts agreed on whether an examination was appropriate in 175 of the 198
cases (88.4%). Out-of-province referrals were significantly more likel
y to be considered inappropriate for MRI: 24 (24%) of the referrals fr
om outside Manitoba were judged inappropriate by three or more reviewe
rs, whereas only 10 (10%) of the referrals from within Manitoba were j
udged inappropriate (chi2, p < 0.01). Of the 19 cases that were judged
inappropriate for MRI by three or more reviewers and that yielded abn
ormal findings, unique clinical information was derived from the exami
nation in only 1 (5%); in contrast, of the 85 cases that were judged a
ppropriate for MRI by three or more reviewers and that yielded abnorma
l findings, unique information was found in 47 (55%). The authors conc
lude that the differences in case mix between within-province and out-
of-province MRI referrals were related to the appropriateness of the r
equest for an examination. They estimate that the combined effect of i
nstituting protocols and reviewing each referral before the examinatio
n could result in a 16% to 31% reduction in the demand for MRI service
s without significantly compromising diagnostic information.