HIS Design&Research On Prediction Model Of Patient Waiting Time

With the construction of the development of hospital information, hospital informationsystem more widely in hospitals, hospital information system is a very complex managementsystem, involving all aspects of hospital management, hospital systems have a direct impacton the quality of service, this paper work to complete three parts, a simple analysis ofhospital information systems, registration systems design and implementation, and patientwaiting time prediction model studies, focusing on system design implementation andregistration waiting time prediction model.In this paper, the first hospital information system structure and function description,while outpatient registration, outpatient charges, hospital management, hospital charges,medical materials management, pharmacy management, storeroom management, outpatientmedical workstation, hospital medical workstation, nurses station and the medicaldepartments workstation11sub-systems are analyzed, the various subsystems of thehigh-level use case diagrams and use case description.Second, this paper out-patient registration system design and implementation process,given the overall workflow patient registration system, gives the first visit registration,referral registration, back registration, billing and other modules of the design andimplementation process, each module is given use case diagram and the overall chart, showsthe main modules of the interface and the implementation process to achieve.Again this paper influence the quality of patient hospital services, waiting time modelhas been studied, the queuing model of patient waiting, and waiting time analysis of theimpact of key parameters, in this based on BP neural network model, and collected samplesdata using the Matlab tool implementation model, and model training, simulation execution.In comparing with the existing methods, such as gradient descent, conjugate, variable learningrate algorithm, we analysis the impact of the results to find a suitable training algorithm andmodel number of hidden layers, transfer function and other parameters.Finally, the paper shows the inadequacy of the system and improved direction.