Download View PDF

A study on building data warehouse of hospital information system
LI Ping†,WU Tao† ,CHEN Mu,ZHOU Bin and XU Wei-guo*
School of Economics and Management, Tongji University, Shanghai 200092, China(Li P)
Xinhua Hospital, Shanghai Jiaotong University School of Medicine, Shanghai 200092, China(Wu T, Chen M, Zhou B and Xu WG)
†These authors contributed
equally to the article.
*Correspondence to: Dr. XU Wei-guo, Xinhua Hospital, 1665 Kongjiang Road, Shanghai 200092, China(Tel:86-21-6579000.
Email: [email protected])
This study was supported by the Scientific Research Plan Projects of Science and Technology Commission Foundation of Shanghai ( No.
09dz1500306).
Keywords: hospital information management; hospital information system; data warehouse; online analytical
processing
Abstract:
Objective Existing hospital information systems with simple statistical functions cannot meet current management
needs. It is well known that hospital resources are distributed with private property rights among hospitals, such as in
the case of the regional coordination of medical services. In this study, to integrate and make full use of medical data
effectively, we propose a data warehouse modeling method for the hospital information system. The method can also
be employed for a distributed-hospital medical service system.
Methods To ensure that hospital information supports the diverse needs of health care, the framework of the
hospital information system has three layers: datacenter layer, system-function layer and user-interface layer. This paper
discusses the role of a data warehouse management system in handling hospital information from the establishment of
the data theme to the design of a data model to the establishment of a data warehouse. Online analytical processing
tools assist user-friendly multidimensional analysis from a number of different angles to extract the required data and
information.
Results Use of the data warehouse improves online analytical processing and mitigates deficiencies in the decision
support system. The hospital information system based on a data warehouse effectively employs statistical analysis and
data mining technology to handle massive quantities of historical data, and summarizes from clinical and hospital
information for decision making.
Conclusions This paper proposes the use of a data warehouse for a hospital information system, specifically a data
warehouse for the theme of hospital information to determine latitude, modeling and so on. The processing of patient
information is given as an example that demonstrates the usefulness of this method in the case of hospital information
management. Data warehouse technology is an evolving technology, and more and more decision support information
extracted by data mining and with decision-making technology is required for further research.
In recent years, hospital information systems have been widely used in major hospitals for the management and
accumulation of clinical data1–4. Current hospital information systems with simple statistical functions cannot meet the
increasing demands of hospitals5. In general, the development of the hospital information system has gone through two
stages6–8.
(1) The development of local area networks, such as charging system. Information is shared within a system but not
among systems.
(2) The application of modern hospital information systems. Through the establishment of a hospital information
system, an overall framework is built to facilitate the flow of information and sharing of information among various
systems.
Current research focuses on the information-based integrated database design mode for a local hospital
information resource9–11. However, hospital resources are distributed with private property rights among hospitals, such
as in the regional coordination of medical services. It is impossible to integrate all hospital resources into a centralized
database. It is crucial to determine how to find proper information from enormous, isomeric and even distributed
repositories, and then ally them with these correlative medical services.
Data such as those on patients, drugs, and doctors are needed for complex statistical analysis and for data mining
techniques to summarize new knowledge or information from massive quantities of historical clinical and hospital data.
Therefore, like management information systems, hospital information systems need to be based on data
warehouse12–16. However, most previous works have focused on the management of medical information15. At this time,
data integration using data warehouse technology is an urgent problem. The lack and distribution of medical resources
is a bottleneck in the treatment of a patient with a complex condition. First, it is difficult to make medical resources
reusable and open such that they can be used efficiently. Second, it is presently not possible to integrate distributed
medical resources for the benefit of cooperative regional medical services. This paper focuses on this problem.
METHODS
Benefits of the data warehouse
A data warehouse is a central repository for all or significant parts of the data that an enterprise's various business
systems collect17. Some of the benefits that a data warehouse provides are as follows18–21.
1. Thematic
A data warehouse provides a common data model for all data regardless of the data's source.
2. Consistency
Prior to loading data into the data warehouse, inconsistencies are identified and resolved, which greatly simplifies
reporting and analysis.
3. Safety
The information in the data warehouse can be stored safely for extended periods of time.
Overall architectural design of a hospital information system
In current systems, with the development of medical affairs, managers require that the health information
management system not only assists the processes involved in medical affairs but also provides data analysis to support
decision making and medical diagnostics. The current hospital information system database is divided into eight
functional databases according to function as show in Figure 1: an outpatient information database, clinical information
database, remote medical information database, electronic medical records information database, inpatient
information database, medical imaging information database, pharmacy information database, and management
information database.
Patient information
database
Clinical information
database
Electronic medical
records information
database
Remote medical
information database
Database
Inpatient information
database
Medical imaging
information database
Figure 1.
Pharmacies information
database
Management
information database
Current hospital information databases.
Traditional database technology, however, cannot meet the above requirements for two main reasons. On the one
hand, there is inadequate storage. On the other hand, current software systems are not suitable to work out with data
mining, data analysis, decision support and medical diagnostic support.
A hospital information system that has multiple data sources, even distributed hospitals, provides data and
information of all medical aspects, and different data sources may have different data formats. To integrate these data
sources effectively requires effective technology, such as data warehouse technology. Theme-oriented data warehouse
issues, and the integrated data collection of medical information can be employed to support decision making. In
hospital information systems, data warehouses play an important role in the data center, working with doctor, patient
and process-related data and information integration. In this way, data warehouse can be integrated as shown in Figure
2.
Doctor
Patient
…
Nurse
Data warehouse
Operator
information
Doctor
information
Nurse
information
Patient
information
Hospital
department of
information
Medical
records
Basic
information
of patients
In-patient
information
Emergency
work log
…
Drug
information
Diagnostic
information
of patients
Outpatient
work log
Surgery the
patient
information
…
Figure 2. Hospital information system data integration based on data warehouse.
On the basis of the overall objective of the system, the system’s overall architecture is a multi-layered structure
that comprises a datacenter layer, system-function layer, and user-interface layer, as shown in Figure 3.
Data warehouse
server
Datacenetr layer
Application Server
System function layer
Data warehouse
server
Datacenetr layer
Web Server
Web Server
System function layer
Intranet/Internet
Application Server
Intranet/Internet
...
...
User interface layer
User interface layer
Hospital A
Hospital B
……
……
Hospital …
Figure 3. Overall system architecture design.
Microsoft Visual Studio .NET is used as a development tool, and the database uses Microsoft SQL Server software,
which is widely used for the management of large database systems.
 Datacenter layer
The first layer is a datacenter layer, which stores information.
 System-function layer
The second layer is a system function layer, which handles the request of the client to complete some logic task. The
function layer also includes a security controller. The data warehouse information system is designed to address the
integration and analysis of medical data. A data warehouse can obtain and collect a high load of data from many
different sources in an integrated and consistent manner.
 User-interface layer
The third layer is the user-interface layer, where a web browser or application interface is used as the client to
complete the interoperation among systems.
Hospital information system for the regional coordination of medical services
Figure 3 shows that hospitals A and B have set up their own datacenter, data warehouse center and corresponding
hospital information system. Hospitals A and B share their information through their web servers. Hospital resources
are often distributed across hospitals, such as in the regional coordination of medical services. The hospital information
system is based on the hospital requirements. The representation of information through a web server is thus proposed
to support the web-based regional coordination of medical services. It is impossible to integrate all hospital resources
into a centralized database. An authorized worker at hospital A can visit the web server of hospital B to exchange
information between the two hospitals. It is easy to transform the distributed medical resource into a reusable and
open web-based resource that can be used efficiently within the system architecture. Additionally, it is convenient to
integrate the distributed medical resource for collaborative medical services using the above method.
Data security strategy
There are two parts to the management of system security.
 Software security

Users use the same username and password to log in to different software systems or modules.
Management of database security
Software prevents users from bypassing logging in directly into the database system to use data.
Data warehouse modeling
The data warehouse design of the hospital information system requires close cooperation among the various
departments and hospitals in determining the data theme, and then detailed design and construction. The creation of
the process of data warehouse model is from a relational, prescriptive model to multidimensional data model
transformation process.
 Creation of a theme
The fact table is the main entity in any future data warehouse. In this paper, the data warehouse of the hospital
information management system includes the management of patient information, management of information on
medical workers, management of medicine information and management of hospital business information, as shown in
Figure 4.
Hospital information
system
Hospital business
information
management
Medicines
information
management
Information
management of
medical affairs
workers
Patient information
management
Hospital business
process information
Health care cost
information
Hospital
management
information
Figure 4. Theme of the data warehouse.
The patient information management system includes the management of patient registration information and
patient information. The management system for information on medical affairs workers includes the management of
doctor and nurse information. The medicine information management system includes the management of drug
information management, and drug storage information management, etc. The hospital business information
management system includes the management of information on hospital business processes, health care costs, and
hospital management.
The theme of the data model is shown in Table 1. The data model includes a fact table and latitude table. The fact
table gives the latitude of the table model of the basic form and information on the storage and management of
medical information directly related to the data. The latitude of the association table is the fact table.
Table 1: Patient summary
Data Management Project
Fact table
Latitude table
Patient
information
management
Medical affairs workers
information management
Patient
information
table
Medical
staff
information table
Medicine
Medicine information
Patient registration latitude table,
medical records table, etc.
Medical information form, nurses
information table, information table
carers, etc.
Pharmaceuticals entering the table,
information
management
sheet
Hospital
business
information management
Hospital
business
information form
splitting
tables
drugs,
drug
sub-delivered, drug inventory table,
patients
receive
medication
information form, etc.
Out-patient service information table,
emergency service information table,
etc.
There are many fields in the latitude table, and the fact table with a reasonable correlation, as shown in Figure 5.
Patient registration table
PatientID
PatientName
Sex
Age
Source
Quantity
Paytype
DoctorID
Diagnoses
Result
DrugID
...
Figure 5. Patient information fact table.

Detailed design and construction
We take patient information as an example to explain the star model of the composition of the latitude model. The
patient information management theme indicators as entities in the fact table of the star model include patientID,
patientName, sex, and age indicators. The relevant latitude information then includes DoctorID, DoctorName, Sex,
age, and SectionID, etc. The information constitutes the latitude entities in the star model shown in Figure 6. Fact tables
and dimension tables set the connection so that each primary key of a dimension table is connected to foreign keys
through the fact table.
Doctor latitude table
Patient registration table
PatientID
PatientName
Sex
Age
Source
Quantity
Paytype
DoctorID
Diagnoses
Result
DrugID
...
DoctorID
DoctorName
Sex
Age
SectionID
Title
Specialty
...
Drug table
DrugID
DrugName
Supplier
Cost
Stock
...
Hospital section table
SectionID
SectionName
Number
Tel
...
Figure 6. The star model for part of the theme of information management.
The current database system for analysis, data design and obtaining source data changes is used to ensure the data
warehouse has parallel source data systems. The main consideration in the design of the fact table is the selection of
theme metrics, such as treatment costs and patient care activities, thus determining the key links with the dimension
table.
 Analysis services of data warehouse
The data warehouse as a platform for online analytical processing is a comprehensive technology that involves
using Structured Query Language (SQL) integrated functions, multi-dimensional analysis techniques applied to design
comprehensive analysis, and procedures for the preparation of a series of SQL search issues and technology. On the
basis of the above data warehouse, the analysis services of the data warehouse are created as follows.
(1) Establish a connection:
Create Open Database Connectivity (ODBC) using a Microsoft SQL Analysis server data source connection; i.e., the
Microsoft SQL Server connection.
(2) Design the data model:
The hospital information system can be divided into the hospital management information system and clinical
information system according to theme. The hospital information system includes the hospital information
management subsystem, personnel management information subsystem, and pharmacy management information
system.
Identify the themes and domain, thus creating a dimension table. Most dimensions of the data warehouse of the
analysis server are built into a shared dimension; e.g., the time, doctor, and patient cube, other dimensions as
non-shared dimensions.
(3) Create a cube:
A cube is an online analytical process in which a data warehouse quickly accesses technical data. Once a cube is
created, a dataset that matches the object of sustainable access is created through the analysis manager of the browse
data option to view a multi-dimensional set or query the database to obtain large quantities of data.
(4) Collation of datasets based on the above model:
The basic information includes the patient, the payment schedule, outpatient form a cube, and the cube-patient
pose. Online analytical processing (OLAP) and a data warehouse are inextricably linked although they are based on
different concepts. The data warehouse for a hospital is a large-scale database of historical data, which are mainly used
for business analysis and to support decision making. These data cannot be used in the transaction processing system
directly. The design mode of the OLAP model based on the data warehouse is shown in Figure 7.
Client A
Data source 1
Data source 2
Data
warehouse
Visualization
Multidimensional
treatment
Data source …
OLAP server
Establish a
connection
Figure 7.
Design the
data model
Create a cube
Client B
Client …
Collation of
data sets
The design mode of the OLAP model based on the data warehouse.
RESULTS
The establishment of a hospital information system provides the information bus for summarizing medical data,
storage, analysis, management, and information sharing to provide data exchange. Information on medical and health
services must be kept in the data center, as a basis for future decision making. A number of distributed hospitals or
different departments in a single hospital may provide medical services to patients and share resources, such as in the
case of data sharing among hospitals, resource coordination, and service sharing; the hospital would have a central
database, or each department would have its internal local database. Data warehouse technology has become a
potential solution to the problem of integrating data from different sources. This paper proposed a design method for a
data warehouse to build a hospital information system and share the regional coordination of medical services.
DISCUSSION
A hospital information system based on data warehouse effectively uses a management information system with
statistical analysis and data mining technology to handle massive quantities of historical data, to summarize new
knowledge of clinical and hospital management, and to provide support for scientific decision making. Data warehouse
technology is continually evolving. We need a data warehouse based on medical information for further in-depth study
that will allow hospitals to provide increasingly valuable decision-support information, including information on disease
control and diagnosis and treatment, information on hospital management, and information to promote hospital
services, such as regional medical services. A consortium that shares medical resources could thus be realized.
References
1. Goble CA, Stevens RNG, Bechhofer S, Paton NW, Baker PG, Peim M, Brass A. Transparent access to multiple
bioinformatics information sources. IBM Systems Journal, 2001, 40(2):532-551.
2. Guo QY, Hao FD, Duan XL, Xie XQ, Liao W. Multi personal computer storage system: solution of sea capacity PACS
storage. Chin Med J (Engl). 2003, 116(5):650-653. PMID: 12875670.
3. Philip AB, Laura MH. Information integration in the enterprise. Communications of the ACM archive, 2008,
51(9):72-79.
4. Paul R, Hoque ASML. A storage and search efficient representation of medical data. 2010 International Conference
on Bioinformatics and Biomedical Technology, 418-422.
5. Srinivasan V, Hillol B, Viswanath V, Jack B. Six strategies for electronic medical records systems. Communications of
the ACM archive, 2008, 51(11):140-144.
6. Eckman B A; Bennett C A, Kaufman J H, Tenner J W. Varieties of interoperability in the transformation of the
health-care information infrastructure. IBM Systems Journal, 2007, 46(1):19-41.
7. J. Anhøj. Generic design of Web-based clinical databases. J Med Internet Res, 2003, 5(4):27. PMID: 14713655
8. Tao Y H, Miao J T. Workstation scheme and implementation for a medical imaging information system. Chin Med J
(Engl) 2003,116 ( 5) :654-657. PMID: 12875671
9. Hoque A S M L. Storage and Querying of High Dimensional Sparsely Populated Data in Compressed Representation,
Lecture Notes on Computer Science, 2002, 2510:418-425.
10. Nadkarni PM, Brandt CM, Marenco L. WebEAV: automatic metadata-driven generation of web interfaces to
entity-attribute-value databases. J Am Med Inform Assoc 2000,7(4):343-356.
11. Wisniewski MF, Kieszkowski P, Zagorski BM, Trick WE, Sommers M, Weinstein RA. Development of a clinical data
warehouse for hospital infection control. J Am Med Inform Assoc 2003, 10(5):455-62. PMID: 12807807
12. Xuezhong Z, Shibo C, Baoyan L, Runsun Z, Yinghui W, Ping L, Yufeng G, Hua Z, Zhuye G and Xiufeng Y. Development
of traditional Chinese medicine clinical data warehouse for medical knowledge discovery and decision support .
Artif Intell Med 2010, 48(2-3):139-152. PMID: 20122820
13. Mary F W, Piotr K B S, Brandon M Z M S, William E T M D, Michael S B A, Robert A W M D. Development of a clinical
data warehouse for hospital infection control. J Am Med Inform Assoc 2003, 10(5):454-462. PMID: 12807807
14. Jonathan SE, Kenneth WS, Robert DP, Jane RS, Robert ER. Case study: a data warehouse for an academic medical
center. J Healthc Inf Manag 2001, 15(2):165-175. PMID: 11452578
15. Grant A, Moshyk A, Diab H, Caron P, de Lorenzi F, Bisson G, et al. Integrating feedback from a clinical data
warehouse into practice organization. Int J Med Inform 2006,75:232–239. PMID: 16153886
16. Li J, Li M, Deng H, Duffy P, Deng H. A web database application for phenotype data management. Oxford
Bioinformatics, 2005, 21(16):3443-3444.
17. William H I. Building the data warehouse. Hoboken:Wiley technology publishing, 2005: 101-122.
18. http://en.wikipedia.org/wiki/Data_warehouse.
19. Torben PB，Christian JS. Research Issues in Clinical Data Warehousing, 1998 10th International Conference on
Scientific and Statistical Database Management, Capri, 43-52.
20. Johnson SB. Generic data modeling for clinical repositories. J Am Med Inform Assoc 1996,3(5):328-339. PMID:
8880680
21. Razan P, Abu S M, Latiful H. Optimized column-oriented model: a storage and search efficient representation of
medical data. Lecture Notes in Computer Science, 2010,6266: 118-127.