Design And Development Of An Intelligent Decision-making System For Advance Care Planning

The overall incidence rate of cancer in China continues to rise.The high incidence rate and high mortality of cancer have induced tremendous pressure on patients,their families and doctors.Especially,the life quality of cancer patients in their terminal stage has seriously declined,whereas there are still many communication problems between doctors and patients.Advance care planning(ACP)intelligent decision-making systems aim to alleviate the widespread problems of decision-making difficulties and lack of knowledge among cancer patients.It reflects the respect for the independent decision-making rights of cancer patients,so as to improve the quality of life of cancer patients.In this article,the implementation method is introduced with the intelligent aided decision-making as well as the design and application of We Chat applets,including three parts,i.e.,data processing,machine learning algorithm and data analysis,application of We Chat applets development.Data processing part deals with duplicate patient indicators and a large number of empty rows and columns in the original medical record indicator pool.The primary processing task is to delete these duplicate and blank data.Finally,it is necessary to check whether the data in all numerical type columns are of numerical type,and some non numeruical data should be converted to the numerical ones.In addition,String type data for cancer patients need special attention,such as whether the way of leaving hospital is death,marriage,medical insurance,etc..One can encode these data sequentially,and finally merge the coding results with the preliminary processed numerical type data so as to obtain a digitized patient index data.When dealing with classification problems,three conventional classification algorithms could be used,namely,logical regression algorithm,naive Bayesian algorithm and decision tree algorithm.Then,the three algorithms are used to train the data of cancer patients who had been preconditioned for cardiopulmonary resuscitation.The training result models of the three algorithms are tested through a test set.The first two models have the lower accuracy rates in judging patients with death.One of the reasons for the low accuracy rate would be the correlation between various chacrteristics of deceased patients,such as blood oxygen saturation and breathing.In the development of We Chat applets,the system is divided into the application layer,the data service layer and the data collection layer.The application layer is responsible for displaying the interface for user operations.The data service layer is responsible for processing data.And the data collection layer is responsible for collecting data.The design of We Chat applet includes system main interface,input interface,output interface and database.Finally,the obtained decision tree algorithm model is applied to the development of We Chat applets to establish an intelligent decision-making system for advance care planning.From the system,various medical record indicators of the patient can be retrieved through the input personal information in the database.Based on data analysis,a decision tree algorithm model is used to obtain the recommended pre order indicators and corresponding rescue success rate of the patient,which are displayed in the output interface.This auxiliary decision-making system can not only provide the medical choices for end-stage patients but also supply decision-making basis for the advance care planning.