Machine learning algorithms trained with pre-hospital acquired history-taking data can accurately differentiate diagnoses in patients with hip complaints

Background and purpose — Machine learning (ML) techniques are a form of artificial intelligence able to analyze big data. Analyzing the outcome (digital) questionnaires, ML might recognize different patterns in answers that relate types pathology. With this study, we investigated proof-of-principle ML-based diagnosis patients with hip complaints using digital questionnaire Kellgren Lawrence (KL) osteoarthritis score. Patients methods 548 (> 55 years old) scheduled for consultation were asked participate study fill an online questionnaire. Our consists 27 questions related gen- eral history-taking validated patient-related measures (Oxford Hip Score Numeric Rating Scale pain). 336 fully completed questionnaires their classified (either osteoarthritis, bursitis or tendinitis, other pathology). Different AI used diagno- ses. Resulting area under curve (AUC) classification accuracy (CA) reported identify best scoring model. The models was compared without radiologic KL scores degree osteoarthritis. Results most accurate model Random Forest (AUC 82%, 95% CI 0.78–0.86; CA 69%, 0.64–0.74) analysis addition Support Vector 89%, 0.86–0.92; 83%, 0.79–0.87). Interpretation Analysis self-reported can differentiate between basic pathologies. radiological further improves these outcomes.