Development of a nomogram to predict medication nonadherence risk in patients with rheumatoid arthritis

Poor adherence among patients with chronic diseases including inflammatory rheumatic diseases (IRDs) is a complex and serious global health care problem. This study aimed to develop an intelligent nomogram using retrospectively collected patient clinical data for predicting nonadherence to biologic treatment in rheumatoid arthritis (RA) patients.

The clinical characteristics of 102 RA patients were collected from outpatients and inpatients at the Orthopedic Departments of Ningxia General Hospital of Ningxia Medical University and Ningxia Hui Autonomous Region People's Hospital from October 2020 to September 2021. Adherence was evaluated using the proportion of treatment days covered within 6 months as the outcome event. A least absolute shrinkage and selection operator (LASSO) regression analysis was used to identify risk predictors, and then multivariate logistic regression analysis was applied to construct the risk prediction model. Furthermore, the nomogram was plotted by multivariable logistic regression.

Among the 102 patients analyzed, 43 patients did not adhere to biologic therapy for various reasons. LASSO regression analysis identified age, sex, education level, disease activity, monthly income, medical insurance, and adverse drug reactions as the significant risk predictors. By incorporating these factors, the nomogram was plotted which showed good discrimination, calibration, and clinical value. The C-index was 0.759 (95% CI: 0.665-0.853), and the area under the receiver operating characteristic (ROC) curve was 0.7416 with a good calibration ability. Decision curve analysis showed that the prediction effect of this model could benefit about 75% of the patients without compromising the interests of other patients.

This nomogram could help medical staff identify patients with higher risk of nonadherence early, so that intervention measures can be taken in time.