Machine learning-based improvement of an online rheumatology referral and triage system

Abstract

IntroductionRheport is an online rheumatology referral system allowing automatic appointment triaging of new rheumatology patient referrals according to the respective probability of an inflammatory rheumatic disease (IRD). Previous research reported that Rheport was well accepted among IRD patients. Its accuracy was, however, limited, currently being based on an expert-based weighted sum score. This study aimed to evaluate whether machine learning (ML) models could improve this limited accuracy.Materials and methodsData from a national rheumatology registry (RHADAR) was used to train and test nine different ML models to correctly classify IRD patients. Diagnostic performance was compared of ML models and the current algorithm was compared using the area under the receiver operating curve (AUROC). Feature importance was investigated using shapley additive explanation (SHAP).ResultsA complete data set of 2265 patients was used to train and test ML models. 30.5% of patients were diagnosed with an IRD, 69.3% were female. The diagnostic accuracy of the current Rheport algorithm (AUROC of 0.534) could be improved with all ML models, (AUROC ranging between 0.630 and 0.737). Targeting a sensitivity of 90%, the logistic regression model could double current specificity (17% vs. 33%). Finger joint pain, inflammatory marker levels, psoriasis, symptom duration and female sex were the five most important features of the best performing logistic regression model for IRD classification.ConclusionIn summary, ML could improve the accuracy of a currently used rheumatology online referral system. Including further laboratory parameters and enabling individual feature importance adaption could increase accuracy and lead to broader usage.

Prediction of the acceptance of telemedicine among rheumatic patients: a machine learning-powered secondary analysis of German survey data

“…In the field of rheumatology, various ML and deep learning algorithms have been employed for tasks such as clinical profiling, patient classification, diagnosis identification, image analysis and predicting treatment response, primarily for rheumatoid arthritis and systemic lupus erythematosus [ 10 , 19 , 20 , 23 – 25 ]. However, to the best of our knowledge, this study represents the first attempt to utilize ML techniques to predict TM try among RMD patients and identify the key contributing factors based on nationwide survey data.…”

“…Nested cross-validation was used for each model to limit overfitting [ 10 ]. More specifically, each of the models underwent a tenfold cross-validation and the classifier hyperparameters were tuned.…”

Classification of Osteoarthritis from Rheumatoid Arthritis based on Stacked Ensemble Machine Learning

“…Six features-patient's age, gender, RF, ACCP, anti-14-3-3, and anti-carbamylated protein-are included in the model. Among others, Iain S. Forrest [30] In spite of the fact that systemic autoimmune rheumatic disorders (SARDs) patients frequently go through protracted diagnostic processes before receiving a diagnosis and treatment, SARDs can have catastrophic consequences if left untreated.Knitza and others [31] Rheport is an online referral tool for rheumatology that permits automatic appointment triage of new referrals for rheumatology patients based on the possibility that they might develop an inflammatory rheumatic disease (IRD). Koo et al proposed machine learning approach successfully identified clinical traits related to remission in each of the bDMARDs.…”

“The Simpler, the Better.” A Qualitative Study on Digital Health Transformation in Early Adopter Rheumatology Outpatient Clinics

“… 7 , 8 Digital services, such as telehealth, 9 , 10 mHealth, 11 , 12 specifically digital health applications (DiGA) 13 or app-based ePRO monitoring, 8 and wearables 14 are valuable measures that can support classic therapeutic options. 15 Furthermore, assistive digital systems like voice assistants, 16 systems for obtaining medical anamnesis, 17 and questionnaire-based need-driven appointment management 18 support organizational processes in medical practices, which in turn might be connected in relieving rheumatology workforce 6 ( Figure 1 ).…”

“…19 In 2020, the German Society of Rheumatology (DGRh) launched their position paper of the commission on digital rheumatology of the German Society of Rheumatology. 20 In addition, eHealth received further impetus from Covid-19 and the urge to reduce infections through less face-to-face contact 2 - 23 that also resulted in European recommendations for remote care in rheumatology. 24 Despite advancements, the perspectives of healthcare professionals (HCPs) on digital transformation and its effects within rheumatology outpatient clinics remain largely unexplored.…”