Identification and prediction of difficult-to-treat rheumatoid arthritis patients in structured and unstructured routine care data: results from a hackathon

Abstract

Background The new concept of difficult-to-treat rheumatoid arthritis (D2T RA) refers to RA patients who remain symptomatic after several lines of treatment, resulting in a high patient and economic burden. During a hackathon, we aimed to identify and predict D2T RA patients in structured and unstructured routine care data. Methods Routine care data of 1873 RA patients were extracted from the Utrecht Patient Oriented Database. Data from a previous cross-sectional study, in which 152 RA patients were clinically classified as either D2T or non-D2T, served as a validation set. Machine learning techniques, text mining, and feature importance analyses were performed to identify and predict D2T RA patients based on structured and unstructured routine care data. Results We identified 123 potentially new D2T RA patients by applying the D2T RA definition in structured and unstructured routine care data. Additionally, we developed a D2T RA identification model derived from a feature importance analysis of all available structured data (AUC-ROC 0.88 (95% CI 0.82–0.94)), and we demonstrated the potential of longitudinal hematological data to differentiate D2T from non-D2T RA patients using supervised dimension reduction. Lastly, using data up to the time of starting the first biological treatment, we predicted future development of D2TRA (AUC-ROC 0.73 (95% CI 0.71–0.75)). Conclusions During this hackathon, we have demonstrated the potential of different techniques for the identification and prediction of D2T RA patients in structured as well as unstructured routine care data. The results are promising and should be optimized and validated in future research.

Difficult-to-treat rheumatoid arthritis: Current concept and unsolved problems

“…The international survey estimated the prevalence of 5-20% based on the British Rheumatology Biologics Registry data (11,14). Subsequently, the proportion of patients who fulfill the criteria of D2T RA or who show multiple b/tsDMARD resistance in patients with RA has been reported in several cohort studies worldwide, mostly in the 5-20% range (15)(16)(17)(18)(19). In our cohort, 7.9% of 672 patients with RA fulfilled the definition of D2T RA (20).…”

Artificial Intelligence in Rheumatoid Arthritis: Current Status and Future Perspectives: A State-of-the-Art Review

“…One-class SVM showed the best performance with a sensitivity and specificity of 89% and 83%, respectively. Despite the step-up therapy and trying several regimens, response failure persists in some patients (i.e., difficult-to-treat patients) [ 119 ]. An extreme gradient boosting algorithm [ 119 ] was able to identify these patients with a comparatively high accuracy (AUC = 0.73, sensitivity = 79%, specificity = 50%).…”

Big data analyses and individual health profiling in the arena of rheumatic and musculoskeletal diseases (RMDs)

“…SLE, Sjogren syndrome) from other RMDs using available sociodemographic, clinical, and biomarker data, addressing the need for improved identification and classification of patients with RMDs, as well as identifying novel avenues for drug repurposing. 34 , 57 – 61 The accuracy of available models varies from moderate-to-excellent in the training datasets. More successful studies compare different ML models to identify the best performing model.…”

“…One of the recent studies used gradient boosting methods to identify and predict difficult to treat RA in structured and unstructured routine care data, using clinically classified RA as a validation set, and achieved area under the curve (AUC) 0.88 for identification and AUC of 0.73 for prediction of difficult to treat RA. 58 An innovative unsupervised automatic phenotyping algorithm (PheVis) combining diagnostic codes with medical record data showed excellent performance for identification of RA patients (cross-validated AUC 0.94) and can be evaluated for other chronic conditions. 73 Another study used support vector machines to classify patients with RA using free-text EHR data validated against manual chart review applying 1987 and 2010 RA classification criteria, resulting in accurate classification of patients with RA (positive predictive value, PPV, 0.86, negative predictive value, NPV, 0.99) and enabling fast patient data extraction from the huge EHR resource.…”