(IJCSIS) International Journal of Computer Science and Information Security,Vol. 8, No. 2, 2010
information like unique patient id, name, sex, date of birth,blood group etc. As per the patient-id, patient details liketreatment prescriptions, test reports, images like MRI, CT-scanetc. have been stored in the database of the hospital server. Onthe basis of stored details of the patient, doctor can prescribethe proper medicine. SQL server tuning has been used whiledesigning the application.
Fig. 1: Smart card based health care system
3.2
To increase the performance of the application of online health care system, the following issues &technical issues will be taken:3.2.1
Issues:
•
Design an application that can serve maximumnumber of hospital and provide a platform fortheir interaction.
•
A system generated patient id that is unique formaximum possible time interval.
•
Store patient information optimally.
•
Speed up the DML (insert, update, delete)operations.
•
Store patient information for maximum possibletime.
3.2.2 Technical issues
:
•
Normalization of tables.
•
Maximum utilization of memory.
•
Imposing necessary constraints on tables.
•
Selecting best possible structure for queries.
•
Tuning and optimizing queries.
•
Optimal selection of join order and joinalgorithm.IV MethodologyApproach for above issues:
1.
The patient will provide basic information through smartcard (having 10 digit patient id and personal information) atregistration counter where it will be uploaded to server.2.
In smart card we are storing patient id and personalinformation like name, address, DOB etc. of a person. As perthe size of smart card we can store the information of a person.3.
Patient’s files may be one of the two forms image or PDF.Images are stored separately in image tables and PDF in PDFtable .These tables contain an Image ID or PDF ID and theirdescription. Image ID and PDF ID are stored in theirrespective tables as foreign keys that may be in patient doctoror Patient Lab or in PatientRoom Tables.4.
Since we are storing all the patient related information’son a centralized database server. So movements of files is notrequired .The required files for a particular patient & their labtest report & images, prescriptions, scan document &diagnosis reports can be retrieved from centralized databaseserver. Accessing the information’s related to patient will becontrolled by various level of access control and it will help usto prevent the unauthorized access from the data base.5.
As we have normalized our database we have removes allpossible redundancies6.
We have covered optimization at application design level,database design level, memory utilization & optimization of queries for accessing the data base. Few points about theoptimization are as follows:
•
Application is studied well and requirements areidentified. These requirements are categorized onobject level and respective tables to store thatinformation are created. After that database isnormalized up to BCNF (Boyce Codd Normal Form)level and 14 tables are created.
•
Size of attributes (their respective data types) isfurther minimized in order to compact record size, Sothat maximum number of records can be stored in asingle data page. Select queries are designed tohandle all possible search criteria.
•
Indexes are created to speed up above read access.First data base ids well studied to find out whatpossible values different attributes will take and whatwill be the size of tables. Considering both the thingsHash, B-tree (clustered and non clustered) andBitmap indexes are proposed.
•
Order of attributes in composite indexes are studiedand justified.
•
Every hospital will have day to day patientinformation locally to improve the performance
280http://sites.google.com/site/ijcsis/ISSN 1947-5500