ABSTRACT
OBJECTIVE: To assess characteristics of deployment of MRI of the SI joints (MR-SI) in patients with suspected axial spondyloarthritis (SpA) before and after a targeted intervention. METHODS: In a retrospective chart review study, all MR-SI performed in the period 1 April 2004 to 31 December 2010 were collected. Inclusion criteria were complete patient data and MR-SI ordered by a rheumatologist for suspicion of axial SpA. MR-SI reports were graded as normal, suspected sacroiliitis or sacroiliitis. In April 2007 an intervention was made to improve deployment. Rheumatologists were provided with data on ordering behaviour, patient characteristics and MRI outcomes. An introduction on the effect of pretest chance on positive and negative predictive value was given; the burden for patients and costs was illustrated. An alternative behavioural strategy was offered in the form of a simple diagnostic algorithm. Percentages of MRIs and positive MRI for sacroiliitis were compared before and after intervention. RESULTS: From April 2004 to April 2007, 198 MR-SIs were performed, of which 166 (83.9%) were normal, 5 (2.5%) were suspicious and 27 (13.6%) were positive. After the intervention, patients displayed significantly more SpA features. More optimal patient selection resulted in 79 MR-SI requests, a decrease of 60.1%. Fifty-seven (72.2%) reports were normal, 0 were suspicious and 22 (27.8%) were positive. CONCLUSION: A simple, one-time, five-step feedback intervention resulted in a 60% reduction in MR-SI requests with a doubling of the percentage of MR-SI positive for sacroiliitis. This approach may benefit future research in areas with diagnostic uncertainty and suboptimal testing.
