This study proposes a system for predicting the recovery status of patients with shoulder damage by estimating the results of the Disabilities of the Arm, Shoulder, and Hand (DASH) questionnaire using an Adaptive Neuro-Fuzzy Inference System (ANFIS). The study aimed to answer two primary research questions: First, is it possible to accurately predict the recovery status of patients with shoulder
... [Show full abstract] damage using the proposed system during treatment? Second, how does this estimation contribute to the treatment process? A literature review indicates that artificial intelligence is often used in rehabilitation to help patients perform exercises correctly. However, previous studies have typically focused solely on exercise execution, without addressing recovery prediction. In contrast, this study aims to predict the recovery status of patients and integrate it into a physiotherapy application, allowing for real-time observation of patient progress. To develop the recovery prediction model, we collected data on the treatment processes of 105 shoulder patients at Bilecik State Hospital and estimated the results of the DASH questionnaire using an ANFIS-based model. The developed model has a mean square error of 9.4E − 3 for the training data and a mean square error of 2.56E − 2 for the test data. The proposed model was integrated into a physiotherapy application using the best weight values from 1000 runs. In this way, it is ensured that successfully predicted recovery status can be observed in real-time. The findings of this study have important implications for shoulder injury rehabilitation. By integrating recovery prediction into a physiotherapy application, healthcare providers can monitor patient progress more effectively and make more informed decisions about the timing and intensity of rehabilitation exercises.
