Background Anti-nuclear antibodies (ANA) are found in patients with (non)-rheumatic diseases and in healthy controls, with 30% positives in the latter group.1 ANA testing is useful in diagnosing auto-immune diseases, especially in Systemic Lupus Erythematodus (SLE), Systemic Sclerosis (SSc), polymyositis (PM), dermatomyositis (DM), Mixed Connective Tissue Disorder (MCTD) and Sjögren Syndrome (SS).1 For the correct use of diagnostic tests knowledge about disease prevalence and test characteristics is essential. Many physicians find this difficult to apply, often leading to overuse of tests.2,3
Objectives To assess the characteristics of ANA testing in patients visiting the rheumatology outpatient clinic of the Sint Maartenskliniek, the Netherlands, before and after a targeted intervention.
Methods The number, result and final diagnosis of all ANA tests done by rheumatologists working at the clinic between 1-1-2010 and 31-1-2012 (25 months) were compared with the ANA’s done after the intervention (1-7-2012 and 31-10-2012; 4 months). To make comparison possible the absolute ANA count was corrected for the number of new patients seen at the outpatient clinic in the same period. The intervention consisted of a one-hour, group wise training in which the results from the pre-intervention period were given together with general background information on the correct ANA use. Directly afterwards all doctors received individual information on their own ANA tests in comparison with their peers.
Results In the pre-intervention period 2696 ANAs were requested by 14 rheumatologists in 2673 unique patients (mean age 53 ±15 years, 74% women); 62% were negative. The most frequent final diagnoses were undifferentiated artralgia/myalgia (20%), rheumatoid arthritis (19%) and fibromyalgia (8%); 7% had an ANA-associated disease (SLE, SSc, PM, DM, MCTD, SS). In the four months after the intervention 127 ANA’s were requested by the same rheumatologists in 125 unique patients (mean age 48 ±17 years, 81% women); 56% were negative. When corrected for number of new patients seen, on average 0.42 (SD ±0.15) ANA per new patient was done in the pre-intervention period and 0.15 (SD ±0.15) in the post-intervention period (p= 0.00), resulting in a decrease of 64% (figure 1). The distribution of the final diagnoses did not change.
Conclusions The finding that our single session intervention resulted in a sizable reduction in number of ANA’s requested, while inter-individual variation and final diagnoses remained unchanged, suggests that excessive usage was reduced without a reduction in appropriate use.
Solomon DH et al. Arthritis Rheum 2002;47:434-444.
Phillips B et al. Arch Dis Child.2009;94:178–179.
van Walraven C et al. JAMA.1998;280:565-6.
Disclosure of Interest None Declared