Synergic deep learning based preoperative metric prediction and patient oriented payment model for total hip arthroplasty.

In recent days, total hip arthroplasty (THP) become a very victorious process which relieve the pain and enhances its functioning. The main motive of the paper is to design and assess a synergic deep learning (SDL) for learning and predicting various metrics like duration of stay (DOS), discharge disposition (DD) and inpatient expenses for THA. The next motive is to develop a patient specific payment (PSP) model reporting the complexity level of the patient. By the use of 15 preoperative parameters from 78,335 THA patients for osteoarthritis from National Inpatient Sample and OME databases, the prediction of DOS, DS and inpatient expenses takes place. A set of two evaluation parameters namely accuracy and receiver operating characteristic curve is used for experimentation. In addition, a predictive uncertainty is employed. All patient refined comorbidity cohort for establishing the PSP model. The presented SDL model exhibited better learning with high trustworthiness, receptiveness, and validity in its prediction outcome. The presented model can be applied for the implementation of PSP model for tiered payments depending upon the case complexities.