Vol.3, No.8, 461-469 (2013) Open Journal of Preventiv e Me dic ine
http://dx.doi.org/10.4236/ojpm.2013.38062
Healthcare intelligence risk detection systems*
Reza Safdari, Jebraeil Farzi#, Marjan Ghazisaeidi, Mahboobeh Mirzaee, Azadeh Goodini
Department of Health Information Management, School of Allied Medical Sciences, Tehran University of Medical Sciences, Tehran,
Iran; #Corresponding author:
Received 21 August 2013; revised 28 September 2013; accepted 8 October 2013
Copyright © 2013 Reza Safdari et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
ABSTRACT
Background: Today, in healthcare field that is
changing rapidly, decision-makers encounter
with ever-increasing inquiries on clinical and ad-
ministrative information to realize customers’ le-
gal and clinical requirements. Therefore, making
decisions on healthcare has changed into a vital,
complex and unstructured issue. The present
paper mainly focuses on describing decision-
making advantages, possible risk to improve ef-
ficiency of decision-making on healthcare, and
especially medical procedures. Methods: The
present research is a review study, which has
been carried out by searching through the au-
thentic scientific sources, including Pubmed,
Google scholar, Iranmedex, and other informa-
tion sources. While defining care intelligence,
here, we introduce Knowledge Discovery Data-
base, the Clinical Support Systems, and Intelli-
gence Risk Detection Model and provide the
conceptual model. Other issues studied in this
paper include the Risk Possibility Assessment
T echnique, Risk Possibility Detection using know-
ledge management techniques, and expert sys-
tems. Result s & Conclusion: Modeling the Intelli-
gence Support System is necessary for design-
ing Real-Time Risk Detection Information Sys-
tems in clinical measures. As taking medical
procedures involves complex decision-makings
and possibility of high risk, operational applica-
tion of the techniques derived from knowledge
and data mining models under study will play a
crucial role in increasing possibility of success
of the measure and pro moti ng safe ty of patients .
Keywords: Risk Detect ion Sy stems; Decisi on
Support; Healthcare Intelligence
1. INTRODUCTION
An effective decision is vital in all healthcare activities.
However, decision-making in this field is complex and
unstructured; therefore, it is necessary for a decision
maker to collect multi-spectral data and information to
make an effective decision at the time he/she confronts
with multiple options. Unstructured decision-making in
complex and dynamic atmospheres is a challenging issue
and a decision maker encounters with low-level informa-
tion almost in every situation [1]. Needs for the pertinent
knowledge, useful data and information are crucial and
in fact, knowledge controlling values, tools acceptance,
and knowledge management methods, technologies and
tactics are essential to guarantee the efficiency and ade-
quacy of decision-making process [2]. By recognizing
this issue, Wickremesinghe and Schaffer [1] developed
Intelligence Continuum. It is a regular method to enable
application of principles and tools of KM, which is nec-
essary for improving decision-making on healthcare,
being confident of the favorability of the results obtained
from the decision-making, and maximizing its advantage
for patients. Therefore, this idea was developed in the
present paper and specially focused on medical proce-
dures; this is a field which is not only remarkable, but
also requires several risks and vital decision-making
processes and in fact it is a suitable atmosphere to de-
scribe advantages of this method. This paper focuses on
the way the Intelligence Risk Detection Model affects in
real-time and appropriately on healthcare support systems.
2. METHODS
The present research is a review study, which has been
carried out by searching the Health, Intelligence Risk
Detection and Systems keywords through the authentic
scientific sources, including Pubmed (13 Item without 11
unrelated works), Google scholar (4 Item without 3 un-
*Competing interests: The author(s) declare that they have no compet-
ing interests.
Authors’ contributions: RS, JF, MG, MM and AG contributed to
conception and design, analysis and interpretation of data, and drafting
and revising the manuscript. JF, MG, MM and AG acquisition of data.
All authors read and approved the final manuscript.

R. Safdari et al. / Open Journal of Preventive Medicine 3 (201 3) 461-469
462
related works), Iranmedex (3 Item), and other informa-
tion sources. While defining care intelligence, here, we
introduce Knowledge Discovery Database, the Clinical
Support Systems, and Intelligence Risk Detection Model
and provide the conceptual model. Other issues studied
in this paper include the Risk Possibility Assessment
Technique, Risk Possibility Detection using knowledge
management techniques, and expert systems.
3. BACKGROUND AND
RELATED WORKS
Before the conceptualize the model we must explain
the aspects of model and related works as HI, intelli-
gence tools, risk assessments and applied systems of
model.
3.1. HI
The concept of intelligence is very extensive. As this
concept entered the business fields, extensive concepts
have formed in this concern, including competitive intel-
ligence, care intelligence, business intelligence, and mar-
keting intelligence. A major part of the sources related to
intelligence is concentrated on industrial organizations
and they have not taken into consideration sufficiently in
the field of health; therefore, to have a better under-
standing, we discuss the differences and similarities be-
tween it and other fields [2,3] (Table 1).
Therefore, healthcare intelligence helps administrative
and clinical management to know the capabilities avail-
able in health network. It facilitates administrative and
clinical decision-making through combining all kinds of
measurement principles on different kinds of internal and
Table 1. Comparing characteristics of healthcare and other
fields.
Differences Similarities
Management is concentrated in
most fields; however, it has
various clinical and
administrative fields on
health department.
All departments require
improvement in charge, quality
and delay through integrated
processes.
Most departments have a certain group of customers with limited
varieties of products. Healthcare requires multiplicity of actors with
distinguished needs.
Following the success of “Focus on Customer Method” in other
fields, the focus in healthcare is mainly on patient; however, the
variety of customers of other departments is also mentioned.
Most industrial systems have
principles of precise
measurement. Feelings and
choices of individuals are also
taken into consideration
in healthcare.
Like other fields, healthcare takes
advantage of system
combination.
external actors, which is caused by a wide variety of
processes [1,3-5] (Figures 1 and 2).
3.2. Intelligence Tools
The intelligence tools are very wide and varied. These
tools, also called “computational intelligence”, include
the following techniques:
1) Expert Systems: “Expert Systems” were formed
since the mid-sixties and they have extensive applica-
tions in different fields of business today. Expert Sys-
tems are made of the following two main elements:
a) Theoretical Understanding of Issue
Collecting creative terms of problem solving are ob-
tained from experience. Deduction ability is the strong
point of an expert system and it is exactly the very factor,
which make an expert system intelligent. Deduction in-
cludes obtaining results of theorems.
Therefore, in order to make a deduction, there should
be a theorem and a result; the latter was achieved from
that theorem [5].
2) Artificial Neural Networks: Artificial Neural Net-
works were introduced since 1940s by Mcculloch and
Pitts. As these models were inspired by biology, artificial
neurons in the neural networks receive sum of data from
the rest of neurons or external stimuli. Here, artificial
neurons’ action is similar to the one of brain neurons.
Then, they transfer the modified data to the other neu-
rons or external outputs; this process is similar to human
brain function. Transfer of information from one neuron
to another neuron is a method for activation and/or reac-
tion to a specific neuronal response based on the data or
stimuli it receives [4].
3) Fuzzy Logic, Fuzzy Sets and Fuzzy Systems:
Fuzzy Sets are defined based on a membership function,
which is the image of a universal set in the range of [zero
and one]. Each element has a grade of membership. A
fuzzy set is made by generalization of classical sets the-
ory. In classical sets theory the membership of elements
in a set is assessed in binary terms. An element either
belongs or does not belong to a set. By contrast, fuzzy set
theory permits the gradual assessment of the membership
of elements in a set [6].
4) Rough Sets: This concept was introduced by Paw-
lak in 1982. Rough sets theory is based on estimation of
concepts (sets) using the available knowledge in an in-
formation system. Its application on lost values is con-
sidered as one of the strong points of this theory. Rough
sets theory calculates two sets of upper and lower ap-
proximations for processing incomplete data [5].
5) Evolutionary Algorithms: Evolutionary algorithm
is the algorithm, which is a subset of the evolutionary
calculation and is placed in the subdivision of artificial
intelligence. This algorithm uses different mechanisms,

R. Safdari et al. / Open Journal of Preventive Medicine 3 (201 3) 461-469 463
Figure 1. Processes of healthcare field [3,4].
OPEN A CCESS
Figure 2. Framework of technology in healthcare intelligence.
including production, leap, composition, and selection.
Selected solutions for optimization problems play the
role of components and, among these components and
people, cost function decides which solutions are re-
mained. These algorithms are frequently used in different
parts of control systems, design, scheduling, optimization,
data restoring, and management [5,7].
a) Combined Techniques: Most of the above tech-
niques can be combined with respect to terms of use to
achieve a new method; for instance, both fuzzy systems
and neural networks can be used simultaneously in
health business.
b) Data Mining, Web Mining: “Data Mining” is a term,
which is used to describe knowledge discovery from
databases. Data Mining is a process that uses statistical,
mathematical, artificial intelligence and machine learning
techniques to extract and identify useful information and
knowledge on large databases. According to the new
definition, “It is a process to find mathematical patterns
among large sets of data”. These patterns may include rules,
dependencies, procedures or models of anticipation [7].
Web Mining can also be defined as exploration and
analysis of the information of interest and useful infor-
mation on a web and about web, which is usually defined
by the web-based tools. The term “Web Mining” was
first used by Etzioni (1996) [6].
6) MAS: AOP is an approach in programming, the
importance of which is ever increasing and some know it
as a more developed form of the Object-Oriented Pro-
gramming. The agent is a software entity, which is
placed on a dynamic environment and is able to recog-
nize characteristics and conditions of the environment
and act autonomously (independently) to achieve a goal
[5].
3.3. Applied Systems of Model
In the following section, general topics about key
scopes of decision support systems, risk detection system,
expert system and their importance for designing risk
detection design is discussed to improve healthcare deci-
sion-making process.
3.3.1. Expert Systems
Expert systems (knowledge-based systems) are also

R. Safdari et al. / Open Journal of Preventive Medicine 3 (201 3) 461-469
464
known as Rule-based Systems. Analyses and programs of
these systems are based on If-Then. The algorithm de-
signed for expert systems is programmable and/or is de-
signed in the form of shells in which the rules and
knowledge should be created and executed. Expert sys-
tems protect knowledge, increase optimization, improve
quality of services and products, do activities in danger-
ous working environments, provide high reliability, dis-
tribute knowledge through its rapid transfer, and reduce
costs in an organization [5].
One of the knowledge-based systems is KDD which is
defined as an important recognition process for authentic,
new, and probably useful and understandable patterns in
data [8,9]. One of the most common applications of
KDD in the field of healthcare is IC. Intelligence Con-
tinuum includes application of data extraction methods,
BI and KM, so that offering favorable healthcare would
be easier in IC model [1].
3.3.2. DSS
Decision Support Systems is a subdivision of Opera-
tion Research, which were created due to the quantifica-
tion techniques of complex issues of management. DSS
includes OR application in supporting management deci-
sion-making processes [5].
Although some studies have been carried out about
DSS in the field of healthcare, DSS applications in clini-
cal and diagnosis activities became 10 times bigger dur-
ing recent decades [10]. Principally, the present research
encompasses the clinical and medical aspects and fo-
cuses on how to apply IT and improve decision-making
efficiency for physicians. Specifically, the computer-
based decision support systems merely focus on the
software designed for this purpose to help clinical deci-
sion-making, and features of patients are in accordance
with computer knowledge basis to carry out patient’s
specific assessment and/or recommendation to physi-
cians [11]. In addition, computer registration of patient’s
information, internet, common decision-making proc-
esses, and new rules facilitate medical decision-making
supporting systems [12]. Consequently, to conduct our
present research, sources containing CDSS and MDSS
are equally related and computer systems are used for
supporting decision-making of healthcare specialists.
Decision making about medical procedures for pa-
tients is considered as a complex and multilateral meas-
ure in most cases. Patients may have various symptoms;
however, this measure is mainly operational. Here, com-
plete anatomic restoration of an organ can be considered
as an example. If it is decided to carry out restoration
later, the risks and advantages of medical procedures in
proportion to the possible risks of taking no measure
should be weighed up. In addition, making a decision on
an alternative treatment using drugs or taking a measure
or a combination of both depend on several factors [1,2].
Decision-making process on medical procedure can be
divided into 3 steps. In the first step or the step prior to
taking measure, a physician acquires sufficient informa-
tion about a patient and his/her medical conditions and
decides whether such a measure is the best option as far
as his profession is concerned or not. When the decision
is made and prior to taking a measure, the patient should
decide on the acceptance and rejection of the physician’s
decision with respect to the anticipated results. When the
patients and physicians are in accord, in step two, the
decision-makings on the unique situations are considered
during a measure. Finally, in the step after medical
measure or step 3, decision are mainly made at two lev-
els: the first level, the strategies for guaranteeing suc-
cessful result for a patient during post-care period and
step b, the lessons that should be applied for future cases.
In order to overcome this complexity, we define two
steps of decision-making within three different and key
processes of decision-making on performing the clinical
measures. The first type of decision-making is called
medical decision-making, which is mainly related to
physician. The second type of decision-making is called
personal decision-making, which is mainly related to
patients because some of the results of a medical proce-
dure are related to the quality of life of patients in which
the patient’s decision-making is vital.
Although, DSS is usually discussed well in healthcare
field, unfortunately, acceptance of these solutions is at a
low level due to unwillingness of physicians to use
computer systems for these purposes [13]. Special atten-
tion should be paid to the guarantee of clinical applica-
tion of the pertinent approach. Real-time risk detection
system is probably useful and suitable and, in turn, its
application and acceptance are reinforced.
3.3.3. Intelligence Risk Detection Framework
in Healthcare
A physician’s performance is usually assessed indi-
rectly, using care results in a hospital, relevant tools, and
probability of risk. Although determination of risk prob-
ability is important for assessing performance and com-
paring results for people or organizations, today, the
present methods pay attention to more extended field of
care. This seems to be rather misleading and physician’s
performance is assessed after a procedure and usually 30
days after the procedure. The weak results may be due to
technical errors, nurses’ errors, drug’s errors and/or care
and the supervision lower than the standard level [14].
With respect to the study of resources, it can be real-
ized that observing these criteria using risk probability
indicators has been under development since 1980 and
several conditions and health plans have been applied
according to the plans for selecting type of medical

R. Safdari et al. / Open Journal of Preventive Medicine 3 (201 3) 461-469 465
measure [15]. Such systems are based on several factors,
including diagnosis, earlier application, demographic
conditions, chronic diseases and personal evaluation
about health and/or executive condition. For instance,
GRAM is a hierarchical and clinical model of the
healthcare application [16]. This model was studied on
100,000 people who were selected randomly from sev-
eral HMO. This model uses the data related to population,
diagnostic qualification, and cost. This system classifies
diseases according to clinical features and the relevant
reactions toward disease. There are 350 diagnostic
grades, which are classified into 9 groups. This model
has not been used yet [2]. Analytical study of the similar
systems shows that there are some constraints, the most
important of which are as follows:
1) Concentration on risk factors related to cost man-
agement and financial issues instead of issues of medical
procedures;
2) Application of risk regulatory framework instead of
risk disclosure framework;
3) Without performance for application of new risk
factors;
4) Lack of attention to some approaches to assess sys-
tem.
In conclusion, with respect to these constraints using
the current systems and by careful consideration of di-
agnosis of possible risk in the field of healthcare, espe-
cially measures, it can be realized that application of
some of IT-based methods, such as expert systems/
knowledge discovery is a data mining function and per-
formance of the current methods for risk assessment are
improved significantly.
To find the relationships between risk factors and rela-
tionships between these factors and results of medical
procedures, an intelligent model would be very efficient
and effective. This means that optimization of a physi-
cian’s performance by applying the effect of possible
risk factors on clinical procedures is considered as a sig-
nificant advantage in intelligence risk detection model.
Due to the diversity of population of patients based on
diagnosis, symptoms, type of the implemented proce-
dures, patient’s age at the time of procedure, and other
necessary factors, assessment of risk probability for
medical procedures is a challengeable issue [17].
With respect to this issue, analysis of the results
achieved for risk probability of the medical procedure
leads to improvement of the performance of physicians
and care centers. Although assessment of risk probability
is an essential issue in decision-making of healthcare
processes, most researchers studying in the field of
healthcare believe that there are a few intelligent systems
with real-time risk assessment factor.
3.4. Risk Assessment
In order to assess the improvement of clinical proce-
dures, it should be noted that identification of possible
risk factors is a useful method [14]. To identify possible
risk factors, it is necessary to develop a multidimensional
model to make evaluation of possible risk accurately. In
the studies carried out, different methods have been de-
vised to study the aspects of the issue and evaluate pos-
sible risk factors; each of them needs different levels of
participation of specialists [10,13].
In the first method, some of the possible factors of
special risks are presented with the presence of some
experts in a concentrated team. Then, the specialists
identify the principle aspects of the risk, which are used
for decision-making in medical procedures. The layers of
these factors provide the risk factors for risk assessment
checklist. In another method, physicians are requested to
fill out risk assessment checklist by assessing risk factors
and define the relationships between these factors or the
relationships between these factors and some actual and
anticipated results and ranges of risks to determine key
performance indicators (KPI). In addition, the methods
recommended by clinical specialists are also presented
and they are requested to assess reaction to these issues
[1,2] (Figure 1).
Risk Detection Using Expert Systems
To apply the intelligent technology in the process of
risk assessment, we recommended using data mining
techniques after discovering knowledge. In studies, dif-
ferent kinds of health data also have a considerable effect
on the technique used for data mining [7]. In fact, after
the end of the step of data collection, some appropriate
procedures should be defined, including sharing and
neural networks. To apply necessary methods of data
mining, development and application of the model is
created after the process of risk assessment of a small
database, which encompasses patients’ data. Some parts
of data show risk factors. Then, following steps (1 to 6)
should be applied [7,9,18].
Step 1. Business Understanding: It includes problem
objective, evaluation of current condition, definition of
data mining objectives, structure of sets of data, and
creation of a project scheduling.
Step 2. Data Understanding: This step includes prepa-
ration of preliminary data sets, data description, explora-
tion of data, and assessment of data quality. Exploration
of data, including statistical parameters, can also occur at
the end of this step. Models like clustering can also take
place during this step to recognize patterns in data.
Step 3. Data Preparation: This step includes selecting
and changing the relevant features, cleaning data, com-
piling data and discussing findings with experts during
data preparation. A deeper search in data can be per-
formed in this step. In addition, other models may be
used to extract the patterns based on the business under-

R. Safdari et al. / Open Journal of Preventive Medicine 3 (201 3) 461-469
466
antee continuous improvement of the anticipated capa-
bilities [1,2,13].
standing.
Step 4. Modeling: Web mining software tools like
(Visualization) and (Clustering Analysis) are useful for
preliminary analysis. Tools like recognizing public rules
are able to extract preliminary correlation rules. When a
more understanding of the data is achieved by recogniz-
ing the pattern, which is obtained with respect to the
outputs of the preliminary models, more specialized
models can be used in terms of the type of data.
As risk factors are effective in reducing people’s qual-
ity of life, the importance of a correct design of deci-
sion-making processes in medical procedure has become
significantly important. The way to follow procedures
always involves innovation in medical technologies;
therefore, by indentifying these risk factors, healthcare
axes can be classified into four key components [1,2,4,
13]:
Step 5. Evaluation: The results obtained from the
models, which were used at the earlier steps are evalu-
ated at the early stages using the performance evaluation
set and within the defined issue and the defined objec-
tives. This will lead to identifying the following needs.
1) Physiological issues with respect to the importance
of quality of life;
2) Technological issues based on new technologies and
their performance;
Step 6. Deployment: Deployment includes arranging a
discussion with experts about data extraction to study the
accuracy of the discussed theorems and to discover the
possible or knowledge patterns from the results of data
extraction. Therefore, if new risk factors or patterns are
discovered, they will discuss the findings with deci-
sion-makers to determine appropriate procedures.
3) Biomechanical issues with respect to health condi-
tion of patients to perform remedial action;
4) Financial issues on costs of these technologies with
quality and type of remedial action;
Consequently, we discuss IRD Model to support deci-
sion-making for a better treatment for this population of
patients during medical procedures and after it. This can
provide better results for the patients and their families.
Of course, a different attitude was adopted toward the
models proposed in literatures [1,2,13] in terms of gen-
erality of issue and scope of discussion, which needs a
detailed account [4].
4. IRD PROPOSED MODEL
Figure 3 shows the steps of risk assessment. The out-
put of risk assessment process helps to determine the
important risk factors of healthcare and to anticipated
results based on the relevant factors. The anticipated re-
sults enable physicians to decide on the care/treatment/
medical plan. To make the final decision, measures are
taken to provide patients with necessary information on
the care plan. Any conflict in making decision on proce-
dure and patients indicates a high level of risk and is
considered as a negative consequence for the procedure.
Such a conflict is used to evaluate possible risk in the
future for the same patient or the other patients in the
system. After medical procedures, the actual results are
also compared with the anticipated results. This com-
parison is an evaluation approach for the model to guar-
To develop such a model in terms of its theatrical and
operational aspects, it is necessary to combine three key
scopes of expert system, risk detection system and deci-
sion support system (Figure 4).
According to the decision-making process framework,
two types of decision-makings (personal decision-mak-
ing and decision-making based on medical procedures)
are defined. This model (Figure 3) leads to detecting
real-time possible risk factors and anticipation of results
of medical procedures. The results in personal and
medical decision-makings are of paramount importance
[19].
Figure 3. Conceptual model for risk assessment [2].
OPEN A CCESS

R. Safdari et al. / Open Journal of Preventive Medicine 3 (201 3) 461-469 467
Figure 4. Intelligence risk detection model.
In the relevant conceptual model, the actual results are
compared with the anticipated results to evaluate risk
detection process. Therefore, actual results sometimes
present new risk factors or new measurement. Healthcare
intelligence reporting tools are the best solutions for cre-
ating final report. They show the most important items,
apply them in the process of risk possibility evaluation,
and these evaluations are repeated subsequently.
5. DISCUSSION
This study discusses the general framework to com-
bine real-time and intelligent risk detection by supporting
decision-making in the field of healthcare. Medical pro-
cedures were also selected due to the complex entity of
decision-making in this field and several risk factors of
this type in decision-makings.
Lack of interaction between college researchers and
those who are active in healthcare industry make detec-
tion of possible risks difficult. In addition, it limits op-
portunities to apply data mining techniques, reduces
value of knowledge discovery, and weakens the way data
are extracted in the field of detecting healthcare risks.
Detection of possible risks of measures has many di-
mensions and aspects whose major focus is mainly on
pathological process, physiological variables, some de-
ductions related to public health, social paradigm, and
quality of life [20]. Consequently, detection of risk fac-
tors is not easy in all aspects; however, evaluation of risk
probability, principally performed by participation of
experts, would be possible based on the new techniques.
Here, we try to discuss some of its main dimensions.
Some limited studies have been carried out to study
the direct advantages of real-time risk detection and an-
ticipation of result to promote decision-making process
in the healthcare field [2,3]. The advantage of this model
is its continuation, as actual and anticipated results
should be compared regularly to correct risk factors,
which leads to improving anticipations in the future. The
important characteristics of the IRD model in combining
3 solutions of IT to remove a clinical problem is related
to the definition and evaluation of patients’ results which
are combined with some scales of evaluation. It should
be noted that the theoretical framework of a problem,
which were developed in the present study using the
thoughts of other researchers [1,2,4,8,13], can be evalu-
ated in the following studies; however, evaluation of
operability of the model will bring about the following
challenges:
1) Applications of IRD Model;
2) Long time developments in healthcare field to
develop capabilities for intelligence models in risk
detection [2];
3) Its major application in case studies;
4) Hospitals’ incentives to execute a project.
6. CONCLUSION
The present paper mainly focuses on describing risk
possibility decision-making to improve efficiency of de-
cision-making in the field of healthcare, and specifically
medical procedures. Today, in the field of healthcare that
is changing rapidly, decision-makers encounter with
ever-increasing inquiries on clinical and administrative
information to realize customers’ legal and clinical re-
quirements [21]. Therefore, making decisions on health-
care has changed into a vital, complex and unstructured
issue [4]. Intelligence risk detection is a challenging field
within health profession [3]. This is not only due to the
fact that it is difficult to experiment with a project and
create samples from educational contents, but also due to
the fact that the field of healthcare has a variety of ser-
vices, scopes, reasons, and irregular relations [1]. We
discussed the application of expert systems and knowl-
edge management in anticipating care results and detect-
ing possible risks of medical procedures at high levels.
The discussed model is based on two steps of (medical
and personal) decision-making processes [22] and in-
cludes a type of decision support systems, which is suit-
able for anticipating the high quality results of healthcare
process [5]. The results update the system to detect pos-
sibility of risk more accurately and adaptively and abili-
ties of anticipating the results are compared with the
fixed model. The present study emphasizes on the im-
portance of risk detection, anticipation, knowledge dis-
covery, and decision on the process of decision-making
to perform clinical procedures on patients. During fol-
lowing steps, it can be applied in appropriate clinical
environments. Therefore, modeling an intelligent support
system is vital for the majority fields of healthcare, espe-
cially for high-risk and complex decision-making cases.
Therefore, it is necessary to design real-time risk detec-
tion information systems in clinical procedures and con-
duct further studies in this field. As taking medical pro-
cedures involves complex decision-makings and possi-
bility of high risk, operational application of the tech-
niques derived from knowledge and data mining models
under study will play a crucial role in increasing possi-
bility of success of the measure and promoting immunity

R. Safdari et al. / Open Journal of Preventive Medicine 3 (201 3) 461-469
468
of patients.
7. ACKNOWLEDGEMENTS
We would like to thank Ahmad Soltani, expert of Health Information
Management at research center for Health Information Management,
for his contributions to the research presented here. We would also like
to thank the reviewers for their insightful comments.
REFERENCES
[1] Wickramasinghe, N. and Schaffer, J. (2006) Creating
knowledge-driven healthcare process with the intelligen-
ce continuum. International Journal of Electronic Health-
care, 2, 164-174.
[2] Moghimi, H., Zadeh, H., Schaffer, J. and Wickrama-
singhe, N. (2012) Incorporating intelligent risk detection
to enable superior decision support: The example of or-
thopaedic surgeries. Health and Technology, 2, 33-41.
http://dx.doi.org/10.1007/s12553-011-0014-z
[3] Mettler, T. and Vimarlund, V. (2008) Understanding busi-
ness intelligence in the context of health care. Proceed-
ings of the 13th International Symposium for Health In-
formation Management Research (ISHIMR), Massey Uni-
versity, 61-69.
[4] Safdari, R., Torabi, M., Farzi, J., Mirzaee, M. and Goo-
dini, A. (2012) Intelligence risk detection models: Tools
to promote patients safety level. Proceedings of the 3th
Symposium for E-hospital & Telemedicine, Tehran Uni-
versity of Medical Sciences, Tehran, 56.
[5] Mahler, E.M., Schmidt, R.M. and Kvitash, V.I. (1993) An
artificial intelligence system to predict progression of
immune dysfunction in healthy older patients. Journal of
Medical Systems, 17, 173-181.
http://dx.doi.org/10.1007/BF00996942
[6] Turban, E., Mclean, E. and Wetherbe. J. (2002) Informa-
tion technology for management. 3rd Edition, John Wiely
& Sons INS., Hoboken, 520-545.
[7] Wu, J., Roy, J. and Stewart, W.F. (2010) Prediction mod-
eling using EHR data: Challenges, strategies, and a com-
parison of machine learning approaches. Medical Care,
48, S106-S113.
http://dx.doi.org/10.1097/MLR.0b013e3181de9e17
[8] Wickramasinghe, N., Bali, R., Gibbons, C., Choi, C. and
Schaffer, J. (2009) Optimization of health care operations
with knowledge management. JHIMS, 3, 44-50.
[9] Aral, K.D., Güvenir, H.A., Sabuncuoğlu, I. and Akar, A.R.
(2012) A prescription fraud detection model. Computer
Methods and Programs in Biomedicine, 106, 37-46.
http://dx.doi.org/10.1016/j.cmpb.2011.09.003
[10] Miller, R.A. (1994) Medical diagnostic decision support
systems—Past, present, and future: A threaded bibliogra-
phy and brief commentary. Journal of the American
Medical Informatics Association, 1, 8-27.
http://dx.doi.org/10.1136/jamia.1994.95236141
[11] Hunt, D.L., Haynes, R.B., Hanna, S.E. and Smith, K.
(1998) Effects of computer-based clinical decision sup-
port outcomes: A systematic review systems on physician
performance and patient. JAMA, 280, 1339-1346.
http://dx.doi.org/10.1001/jama.280.15.1339
[12] Fieschi, M., Dufour, J.C., Staccini, P., Gouvernet, J. and
Bouhaddou, O. (2003) Medical decision support systems:
Old dilemmas and new paradigms? Tracks for successful
integration and adoption. Methods of Information in
Medicine, 42, 191-198.
[13] Moghimi, F., Seif Zadeh, H., Cheung M. and Wickrama-
singhe, N. (2011) An intelligent risk detection framework
using business intelligence tools to improve decision effi-
ciency in healthcare contexts. Proceedings of the 17th
Americas Conference on Information Systems (AMCIS),
United States, 4-7 August 2011, 1-8.
http://aisel.aisnet.org/amcis2011_submissions/173/
[14] Larrazabal, L.A., Del Nido, P.J., Jenkins, K.J. and Gau-
vreau, K. (2007) Measurement of technical performance
in congenital heart surgery: A pilot study. The Annals of
Thoracic Surgery, 83, 179-184.
http://dx.doi.org/10.1016/j.athoracsur.2006.07.031
[15] Keenan, P., Buntin Beeuwkes, M., McGuire, T. and New-
house, J.P. (2001) The prevalence of formal risk adjust-
ment in health plan purchasing. Inquiry, 38, 245-259.
http://dx.doi.org/10.5034/inquiryjrnl_38.3.245
[16] Hornbrook, M.C. and Goodman, M.J. (1996) Chronic
disease, functional health status, and demographics: A
multi-dimensional approach to risk adjustment. Health
Services Research, 31, 283-307.
[17] Kang, N., Cole, T., Tsang, V., Elliott, M. and de Leval, M.
(2004) Risk stratification in paediatric open-heart surgery.
European Journal Cardio-Thoracic Surgery, 26, 3-11.
http://dx.doi.org/10.1016/j.ejcts.2004.03.038
[18] Kumar, A. and Gosain, A. (2009) Analysis of health care
data using different data mining techniques. Interna-
tional Conference on Intelligent Agent & Multi-Agent
Systems (IAMA), Chennai.
[19] Safdari, R., Farzi, J., Mirzaee, M. and Goodini, A. (2013)
Intelligence Risk Detection systems for medical proce-
dures. Proceedings of the 1st conference for Telemedicine,
Amirkabir University, Tehran, 46.
[20] Wilson, I.B. and Cleary, P.D. (1995) Linking clinical
variables with health-related quality of life: A conceptual
model of patient outcomes. JAMA, 273, 59-65.
http://dx.doi.org/10.1001/jama.1995.03520250075037
[21] Farzi, J., Salem Safi, P., Zohoor, A.R. and Ebadi Fardazar,
F. (2008) The study of National Diabetes Registry System:
Model suggestion for Iran. Journal of Ardabil University
of Medical Sciences & Health Services, 3, 288-293.
[22] Garg, A.X., Adhikari, N.K., McDonald, H., Rosas-Are-
llano, M.P., Devereaux, P.J., Beyene, J., Sam, J. and Hay-
nes, R.B. (2005) Effects of computerized clinical decision
support systems on practitioner performance and patient
outcomes: A systematic review. JAMA, 293, 1223-1238.

R. Safdari et al. / Open Journal of Preventive Medicine 3 (201 3) 461-469 469
ABBREVIATIONS
KM: Knowledge Management;
HI: Healthcare Intelligence;
MAS: Multi-Agent Systems;
AOP: Agent-Oriented Programming;
DSS: Decision Support Systems;
CDSS: Clinical Decision Support Systems;
MDSS: Medical Decision Support Systems;
GRAM: Global Risk Assessment Model;
IRD: Intelligence Risk Detection;
KDD: knowledge discovery in databases;
IC: Intelligence Continuum;
BI: Business Intelligence.