Web-based data warehouse in the osteoporosis community health information management system

Objective: To investigate the remote management system of osteoporosis community intervention and design the data warehouse. Materials and Methods: The basic principles and methods of data warehouse were applied to the osteoporosis community intervention to build the MySQL 4.5 relational database using PHP as the development tool. A web-based B/S Model remote management system was established for the high risk population of osteoporosis in the community. Results: The system can be used for data management, data query, online analysis, etc., in community health service center, specialist outpatient for osteoporosis, and health administration sectors. Conclusion: The remote management system and data warehouse can provide guidance for policymaking of health administrators, residents health information, and intervention suggestions for general practitioners in community health service centers, patients follow-up information for osteoporosis specialists in general hospitals, as well as large quantities of original research data and preliminary health statistic results.


Subject elements of community management for osteoporosis
The subject elements for community management of osteoporosis included dimension (content of business), subjects (includes data of subjects), particle size (dimension levels to extract data details), and storage limit (of data).Based on a comprehensive analysis of data conducted by physicians from community health service centers and the indicated departments of Changshu Hospital of Traditional Chinese Medicine, several dimensions were confi rmed; these included baseline information, osteoporosis-related highrisk factors, bone density, and assessments of interventions.Each dimension was divided into several levels, and particle size was used to confi rm and elucidate the dimension levels.
Osteoporosis-related high-risk factors are complex.Among males, 21 high-risk factors have been identifi ed, whereas 26 have been identifi ed among females. [7]We screened the high-risk factors and provided options for data entry, analysis, and mining.Evaluated osteoporosis-related high-risk factors comprised age, body mass, family history, and nutritional factors; intervention measurements included appropriate diet, exercise, suffi cient calcium intake, vitamin D intake, and correction of poor life-style habits; bone density examination items included quantitative computed tomography, dual energy X-ray absorptiometry, and ultrasound bone intensity examination; treatment measurements included calcium agents, vitamin D, alendronate sodium tablets, calcitonin, Aclasta, and estrogen replacement therapy.The dose and duration of each intervention, examination, and treatment were recorded.
The snowflake model was used to summarize the above multidimensional data relationships; the model consisted of a fact table and a group of dimension tables.This allowed the dimension factors to be further divided; for example, intervention factors included sports, amount of sunlight, appropriate diet, and medication.The "intervention" dimension could be snowfl aked, that is, the dimensions could be decomposed in terms of the attributes for sports, amount of sunlight, appropriate diet, and medication to form four-dimension tables.For the "female" dimension, time since menopause had to be recorded as the particle size.

Dynamic loading
Community management needs to be continuously updated so as to provide dynamic data loading when constructing a database.Traditional data warehouses store diachronic, resting, and integrated business data, which initially load data and then support business searches.However, data loading with a dynamic data warehouse can load data while simultaneously allowing users to conduct searches.Moreover, dynamic loading does not affect the use of the data warehouse, which allows the immediate analysis of loading data.The intervention measures and bone density measurements of subjects could be continuously recorded.

CONSTRUCTION OF DATABASE AND PHYSICAL ACHIEVEMENT Fact table of data warehouse
The fact table contains all the osteoporosis health data.The fact table was the largest table we constructed and its information was updated the fastest in the data warehouse.All attributes for each record depended on the primary key of the fact table, and a series of foreign keys was associated with each dimension table.With regard to the search function of the data warehouse, it is necessary to minimize connection operations among different tables.The fact table was designed as shown in Table 1.

DESIGN OF RELATED DIMENSION TABLES
For each attribute in the fact table, the dimension information was recorded using a special dimension table to confi rm the values of some dimensions. [4]The design of the dimension tables was based on the table name (main key word coding, name), and content in parentheses in The search function was improved by combining the small dimension tables.For example, the social dimension table is presented in Table 3.

Interactive data distributed structure
System management was achieved through interactive management of physicians from the community health service center and general hospital and health administration departments.Therefore, we used an interactive data mart structure.Although different data marts were achieved in specifi c departments, they were integrated and interlinked to provide a comprehensive data view for business scope.The administrator assigned different permissions for different grades of users.The user had to register the database for data entry, correction, export, and analysis, that is, the data warehouse for business scope.For example, the staff of different community health service centers could enter different data relating to their own communities.
User management is also a special dimension table.The user management was built as presented in Table 4.

PLATFORM DEVELOPMENT AND DATA WAREHOUSE
Hypertext Preprocessor was used to develop web application programs.The Linux operating system was employed as the server operating system, equipped with the Apache 2.0 operating platform.The MySQL 4.5 relational database was built and is accessible on the internet.The user can input http://www.cszlf.net/sycweb/ in the web browser, and the log-in page appears as seen in Figure 1.
The data entry interface appears after inputting the user name and password [Figure 2].
The system provides online analysis for real-time and online analysis of data in the data warehouse, including individual case analysis, group analysis, and global analysis.For individual case analysis, the user can search the target records through a conditional search and click the "analyze" button on the interface to obtain the individual case analysis by system and primary diagnosis and treatment suggestions; these do, however, require the confi rmation of clinical physicians.This process achieves computer-assisted diagnosis and treatment [Figure 3].

DISCUSSION
The re is a common lack of appropriate tool software and unified ordered organization in community public health service management, [8] and so it is difficult to analyze daily working data or to further community health management.Although some infor mation management systems for osteoporosis have been developed and used, most of them use databases rather than data warehouses; thus, they do not support data analysis functions or data mining. [9]We have been collecting data since 2010, and we have combined basic principles and methods for a data warehouse with community intervention in constructing such a warehouse that achieves dynamic loading.This data warehouse has several positive features: It effectively organizes data sources, and provides deep-level data mining and online analysis.In addition, users

Figure 2 :
Figure 2: Data entry interface of the changshu osteoporosis remote management system

Figure 3 :
Figure 3: Online analysis interface of the Changshu Osteoporosis Remote Management System

Figure 1 :
Figure 1: Log-in screen of changshu osteoporosis remote management system Table 2 represents the fi eld names.The main dimension tables are shown in Table 2. International Journal of Medicine and Public Health | Oct-Dec 2013 | Vol 3 | Issue 4