| Correctness of medical work is critical.Careless would cause fatal consequences.So we must do prescription comments.But at present,many hospitals do comments manually.Some have taken prescription comment systems,however,the systems are function-fixed,difficult to develop in custom.Or they are expensive,not suitable for average institutions.We have done research on demand and feasibility of developing prescription comment systems,based on B/S structure.And we described what we have done.The system we developed has such functions: import and export drugs and prescriptions data,assign tasks by monitors,do comment by operators,then verify by monitors again.It can assign records randomly,saving time of manually works.Also,we can comment automatically by drugs and rules in database,while those rules could be modified either online or offline.We have provided a platform service.Other applications can post regular prescriptions to our system’s interface,to get generated comments.Besides,we attempted in creating auto-comment rules(mainly dosages)by machine learning.This system is based on LAMP services.We used improved MVC developing mode,saved complex data by JSON format.We took HTML,CSS,Java Script pages for interfaces,and PHP for server scripts.Besides,some few modules like import and export are developed by Java.This application are small-sized,lightweight,which can be adopted on simple devices.The system has some features of architecture of large scale websites,like transit server and distributed processing.For transactional,we ensured idempotence of all,part of them have atomicity and consistency.We used Bootstrap to create standard and responsive web pages.Also we ensured correct operations at most times.For security,we did work on avoidance of external attacks and SQL injections.At testing period,we took thousands of prescriptions authorized by tens of doctors,did works including random assignment,comment automatically,etc.All functions are available.In the article,we have told theories and technologies in development.Designs like databases,storing format and adoption are also described.Finally,we discussed what we can do in the future. |
