Background Unfortunately, response to Methotrexate (MTX) is not universal and nonadherence may partially explain poor response, therefore identification of patients who are likely to be nonadherent prior to MTX commencement would present an early opportunity to intervene and optimise treatment response.
Objectives 1)To evaluate adherence to MTX over the first six months 2)To identify potential predictors of nonadherence.
Methods Patients were recruited to the Rheumatoid Arthritis Medication Study (RAMS). This is a UK multicentre prospective cohort of incident MTX users with a diagnosis of RA. At baseline, clinical history, disease activity scores (DAS28), disease duration, Health Assessment Questionnaire (HAQ), visual analogue scales (VAS) general well-being, pain and fatigue, demographic, alcohol and smoking data are collected. Patients completed The Beliefs about Medicines Questionnaire (BMQ), Brief Illness Perceptions Questionnaire (BIPQ), Hospital Anxiety and Depression Scale (HADS). Adherence was measured during the first 6 months after MTX commencement using a patient completed weekly MTX diary, including missed doses and reasons. Proportional adherence was calculated as the percentage of weeks patients took MTX as directed. Nonadherence was defined as ≥1 dose missed against medical advice. Baseline variables with a near significant association with adherence in univariate logistic regression analysis (defined as p<0.2) were entered into a backwards stepwise logistic regression to identify independent factors associated with nonadherence.
Results Of the total 1014 eligible patients recruited between 31/08/08 and 01/09/14, 661 (65%) returned a 6-month diary. Comparison of baseline characteristics showed diary returners were younger, less deprived, and had lower disease activity than those who did not return a diary. Patients included in this study were on average 60.2 (13.0) yrs old, had a median DAS28 score of 4.4 [3.4-5.3] and 68% were female. Overall 156/594 (26.2%) reported nonadherence (≥1 nonadherent week), although proportional mean adherence was high (97%). Of the 438 who were adherent 17% were medically advised to miss ≥1 dose. Reasons for nonadherence (%nonadherent patients) included; feeling unwell (38%), side effects (33%), no reason (18%), forgot (14%), drug holiday (11%), changed dose (7%), ran out (6%), delayed commencement (5%) & lack of effect (1.5%). Higher tender joint count (OR 1.032 95%CI:1.007-1.056), fatigue (OR 1.097 95%CI 1.026-1.172), higher concerns (OR 1.133 95%CI:1.037-1.238) and emotional distress (OR 1.069 95%CI:1.002-1.140) attributed to RA and lower mood at baseline predicted nonadherence at six months. 12 baseline variables were entered into multivariate analysis (Table 1), and a higher fatigue score (OR 1.071 95%CI:0.996-1.151) and higher concerns attributed to RA (OR 1.113 95%CI:1.013-1.223) associated with nonadherence (ROC analysis AUC=0.6027).
Conclusions Initial adherence to MTX amongst diary returners was high, higher than previously reported. Further exploration of the way higher RA concern, a modifiable component of the self regulation model of illness, is influencing patient nonadherence may provide insight into early interventions to optimise patient adherence.
Disclosure of Interest None declared