Systems approach for classifying the response to biological therapies in patients with rheumatoid arthritis in clinical practice

Abstract Introduction Biological therapies have greatly improved the treatment efficacy in rheumatoid arthritis (RA). However, in clinical practice a significant proportion of patients experience an inadequate response to treatment. The aim of this study is to classify responding and non-responding rheumatoid arthritis patients treated with biological therapies, based on clinical parameters and symptoms used in Western and Chinese medicine. Methods Cold and Heat symptoms assessed by a Chinese medicine (CM) questionnaire and Western clinical data were collected as baseline data, before initiating biological therapy. Categorical principal components analysis with forced classification (CATPCA-FC) approach was applied to the baseline data set to classify responders and non-responders. Results In this study, 61 RA patients were characterized using a CM questionnaire and clinical measurements. The combination of baseline symptoms (‘preference for warm food’, ‘weak tendon severity’) and clinical parameters (positive rheumatoid factor/anti-cyclic citrullinated peptide antibody, C-reactive protein, creatinine) were able to differentiate responders from non-responders to biological therapies with a positive predictive value of 82.35% and a misclassification rate of 24.59%. Adding CM symptom variables in addition to clinical data did not improve the classification of responders, but it did show 8.3% improvement in classifying non-responders. Conclusions No significant differences were found between the three classification models. Adding CM symptoms to the clinical parameters in the combined model improved the classification of non-responders. Although this improvement is not significant in the current study, we consider it worthwhile to further investigate the potential of adding symptom variables for improving treatment efficacy.