154 / 2024-04-22 10:57:43

AI's Assistance in Disease Diagnosis: When Patients Cannot Access Advanced Tests

diagnostic accuracy,simultaneous imputation and prediction,deep learning,high-dimensional data,advanced information,healthcare operations

The accuracy of diagnostic processes significantly influences patient outcomes and carries considerable implications for service efficiency and quality within healthcare systems. However, this accuracy can often be compromised by financial, psychological, or physical barriers that restrict access to advanced diagnostic tests, especially among rural patients. In response to these constraints, our study explores the deployment of artificial intelligence to maintain high diagnostic accuracy, regardless of the availability of advanced test results. This paper introduces SIP-HD, a novel deep learning model that performs simultaneous imputation and prediction on high-dimensional data. Notably, our model utilizes the imputation process to enhance the accuracy of predictions. When benchmarked against various two-step models, one-step models, and doctors' preliminary diagnoses, SIP-HD exhibited superior performance in terms of accuracy, area under the curve, and F1 score. The proposed model offers significant potential to enhance diagnostic decision-making in real-world scenarios, particularly when advanced tests are inaccessible. Its adaptability amplifies its value, allowing it to serve as both a supportive diagnostic tool and a telemedicine aid, with a potential extension to an online self-diagnosis platform. These applications hold promise for improving service efficiency and quality within healthcare while also offering substantial cost savings.