I read with great interest the research of Suh et al. [1] published in the Korean Journal of Internal Medicine on January 2018 regarding the status of medication nonadherence in Korean patients with rheumatoid arthritis (RA) and related factors. A total of 292 patients with RA participated in this study. The subjects were divided into two groups, intentional and unintentional, based on the results of a self-reported questionnaire. Each patient completed two iterations of a brief illness perception questionnaire, a belief about medicines questionnaire, and a patient health questionnaire, which provided demographic and clinical data, patient perceptions of illness, treatment beliefs, and mood. Statistical differences of related factors were calculated among different groups; multinomial regression analysis was applied to determine risk factors. The authors also compared the prevalence of unintentional and intentional non-adherence medication adherence, which was stratified by variants of anti-rheumatic medication. The rationale of the study is clear but left me with some thoughts.

Previous studies have rarely used unintentional nonadherence and intentional nonadherence as the criteria for grouping. This distinction is one of the highlights of the paper and it reflects the authors’ deeper understanding of medication nonadherence. However, if a patient is both intentional and unintentionally nonadherent, it raises the question how do we accurately classify these behaviors? This is a crucial distinction that I believe needs to be addressed further.

Another pertinent issue of this study is that the questionnaires did not provide verification of the reliability or validity. This lack of methodological vetting raises difficulty as the reader cannot judge whether the questionnaires provide accurate responses, nor can we be certain of the reliability of the results.

Finally, the authors compared the rates of intentional and unintentional nonadherence among different drug types, but there are repetitions in samples between methotrexate group and nonsteroidal anti-inflammatory drugs and/or disease-modifying antirheumatic drugs group. This discrepancy may have affected the results and subsequent conclusions. I would suggest that samples be separated into different group by number [2] or name of drugs such as methotrexate, leflunomide, or hydroxychloquine [3].