Background/Purpose:   A
number of composite outcome measures have been validated to quantify disease
activity in Rheumatoid Arthritis (RA).  Few studies have been published on
the use of a multi-biomarker disease activity (MBDA) blood test (Vectra DA®) in
a real world setting. The published articles that attempt to define the impact
of this test in clinical practice have been an exercise on the use of the MBDA
in RA case vignettes and a study evaluating the effect of the MBDA on treatment
decisions.  As there are no published studies using the MBDA in a true
clinical practice compared to a validated outcome measure, in this study we
compared the MBDA to the CDAI using the clinical registry JointMan® (JM).

Methods:  This
is a retrospective, observational study. JM captures RA diagnostic criteria and
selected disease features, formal joint counts, outcome measures, serology,
medication efficacy and safety and reason for discontinuation of medications.
Due to the retrospective analysis methodology not all data was available.

Results:

Over
5700 unique RA patients and 133,000 encounters have been recorded since
2009.  

The
total distinct Vectra patient count was 687 and the total number of Vectra
results was 1079.  The total JM entries where a Vectra and CDAI score are
within 20 days of each other was 436. Total distinct Vectra patients with more
than one Vectra score was 274.  

We
calculated three correlation coefficients: the Pearson, the Cohen-Kappa, and
the Weighted Kappa. The correlation coefficients are delineated in Table
1. The results demonstrate a Cohen Kappa estimate of 0.093
(p<0.006), a Weighted Kappa of κ = 0.298 (p<0.001) and a
 Pearson score, ρ = 0.4185 (P<0.001). Table 2.
Contains the distribution of the paired data.

    
Table 1.

Table
2.

Conclusion:  This
is the first study using the MBDA in a true clinical practice setting. In this
study the MBDA was compared to CDAI.  The Cohen Kappa indicates “slight”
agreement, whereas the weighted score of K = .298, is a “fair” correlation
according to Landis and Koch. The Vectra DA score achieved almost 3/10 of the
way between the expected accuracy (random) and perfect classification accuracy
(CDAI being the standard). The Pearson ρ score of 0.4185 (P<0.001)
falls into the category of a strong positive linear relationship between Vectra
and CDAI scoring metrics. The qualitative nominal interpretations of the scores
indicate that there is a valid, yet incompletely characterized relationship
between the metrics. Additional prospective studies are needed in real world
settings are needed to continue to define the role of the MBDA score.