Nowadays, technology is growing rapidly. With such tremendously expansion of technology, many field of industry is taking the opportunity in adopting these technologies to transform their business move to match with the environment. Medical is one of the industries that changing their services to provide better treatment and better treatment to patients. Many scientific center, clinics or medical organization is investing on Clinical Decision Support System to enhance the quality of decision making from the improvement of prognosis.
What is Clinical Decision Support System?
Clinical Decision Support Systems are "productive knowledge systems designed to use several items of patient data to generate case-specific advice" from Wyatt J, Spiegelhalter D, 1991 (OpenClinical 2001-2009)
It designed to integrate with a medical knowledge databases as well as patient data to create circumstance specific advises to users. In another words, it was created to healthcare professional to make medical decision.
Instead of taking the place of diagnosis as a job of computer program, it somewhat designed to support the professional medical experts because computer struggles to perform as a human being and it could cause error which may harm and risking others people survivability.
In some area, computer systems can help the clinician in retrieving details needed in the progress of examination such as patient's health background, all kind of assessment and lab test. Furthermore, the result of drug and allergies toward the individual will be taken into account to help a occupied clinician to handle over hundred patients in a day. (Clinical Decision Support System, Citizendium, 2006)
What is the goal of Clinical Decision Support System?
CDSS generally is employed to assist clinician by using the point of medical to provide some expert opinion or tips. A clinician may connect to CDSS in doing determination of diagnosis, analysis and etc by relating to provided patient data.
Previous theories of CDSS were to use the CDSS to virtually make decisions for the clinician. (Clinical Decision Support System, iScanMyFood, 2010). Right now, clinician is able to type information to the machine and await CDSS to result a good choice to advice them the correct action.
By been through the computer research, clinician isn't only making decision through own knowledge which might not exactly be the most suitable result from a identification but also getting advices from computer to increase the quality of decision making. In another words, it dished up as a peripheral brain.
Functions of Clinical Decision Support System
There are 4 basic functions contain in Clinical Decision Support System that happen to be Administrative, Managing clinical complexity and details, Cost control, Decision support by based on Perreault & Metzger.
Administrative means system must be administrable which means that it must have the ability to support medical coding and documents, procedures and referrals of the medical center. In order to achieve that, CDSS is obviously created through multiple programs and it recognizes perfectly on every medical's standard technique.
Other than that, it must be able to manage specialized medical complexity and details. It continues patients on research and chemotherapy protocols as medical experts always does. It tracks patient orders, recommendations follow-up the position of patient and preventive treatment after prescription.
Cost controllable by avoiding any duplication of process, report or any pointless lab test and also to monitor medication purchases to confirm any incorrect places that will be a direct injury to particular medical center's financial
Decision Support is mean to aid clinical analysis and treatment solution functions and promoting use of best practices, condition-specific rules, and population-based management. (OpenClinical 2001-2009)
Characteristics and Types of Clinical Decision Support System
Characteristics of CDSS
There are 4 basic component usually required by CDSS that are Inference Engine, Knowledge Base, Justification Module and Working Memory.
Inference Engine
Inference Engine unit is the main part of CDSS. It used knowledge from data source integrated with the machine as well as the knowledge about the individual to create an productivity or a finish predicated on certain condition. Inference engine motor control the activities of the machine and guide system with the best actions. For an example, it will start to detect the condition to bring about the alert or summary to be displayed in a diagnostic improvement.
Knowledge Base
Knowledge Base purchased the knowledge Inference Engine used to provide to the users. In Knowledge bottom part, it contains every risk factor to handle in new lesions and risk results. It'll be constructed with the involvement of clinical domains experts with also every activity of create, edit and maintenance. In another way, some knowledge platform is established through computerized process. Automated process knowledge is received from external resources such as catalogs, magazine, journal articles and repository by the computer application. The procedure of creating an understanding base is complicated and complicated. To make it easier, there are tools specially intended to accomplish the acquisition and elicitation of knowledge basic. There can be an example tool called Protege, a knowledge- based development environment.
Working memory
Working recollection is a collection of patient data or form of a note which is stored inside databases. These data may include patient's age, name, data of beginning, gender and etc or allergies, history medical information or problems and other information.
Explanation Module
Explanation Module dependable in composing justification for the conclusions attracted by the Inference Engine by applied Knowledge basic and patient data. This component is not shown in every CDSSs.
In another way, CDSS can work on synchronous method and asynchronous function. In synchronous setting, users can converse directly with request to wait for the productivity from system. Users will have to wait for the output to be able to keep their works. For instance, CDSS inspections for drugs connection or any possible drugs that patient allergies to then clinician is only going to able to continue to diagnose patient by predicated on the result produced by CDSS. When there may be in asynchronous setting, CDSS is undertaking independently while does not required user to hold back for. For a good example make a checkup reminder for patients.
CDSS can be grouped as open-loop or closed-loop systems. Open-loop CDSS will generate a summary but it requires no action immediately by its own. Usually users will need the actions on the final decision. For a good example, CDSS creates alert or reminder to users to take the activities. A Closed-loop CDSS is the opposite of open-loop CDSS. It will take actions by its own without any intervention from users. For a good example, system will automatic save up all details of diagnosis process.
CDSS can be an event keep an eye on, a consultation system or a medical guideline. Even keep an eye on is a software application that changes every available data into electric format and uses its designed knowledge basic to send reminder to clinicians appropriately. Consultation system allows customer enters the details of a case and in other ways, the system will provide user a set of issues that may explain the case and suggestion the best action to be taken.
Clinical Guideline essentially developed by a group of medical experts and disseminated by the federal government or by professional company and it apply in almost all of the CDSS. This clinical guide has been presented with every declaration of guidelines regarding to a specific health condition. Apart from providing recommendation from various methods, it could be taken as instances in medical education.
Type of Clinical Decision Support System
Knowledge-based Clinical Decision Support System (Expert System)
Knowledge-based expert systems are manufactured insurance agencies experts use the biomedical books to identify romantic relationships between independent factors (such as signs and symptoms) and centered factors (such as likely underlying diseases).
It is made up of related established such as local clinic information, patient data and other compiled data and put it on with IF-ELSE-THEN predefined rules to guide through the whole improvement of decision making. However, rules may also be acquired from numerous kinds of decision trees and shrubs.
These rules-based CDSS is the most usually found among all the clinical application. It will alert individual when there's a possible drug doses or allergies which might harm or associated risk patient life by based on patient details such as age group, sex, weight, level and etc.
Example: if the system rules used to find out drug relationship, the solution will started to run also to identify every possible high-risk drug interaction, the rules might be IF medication A is taken AND medication B is used THEN alert consumer. By going through these predefined rules, provided information must be always up-to-dated to avoid any wrong end result which can lead to misdiagnosis.
To construct a rule-based system for medical decision support, an expert with domain knowledge always must be recruited to create and handle the knowledge base and coach the system. To train an expert system is very time-consuming and it the effect that produced is merely useful in a small scope task. Therefore, a rule-based CDSS is not popular to deliver the critical communication to clinician. (Clinical Decision Support System, Citizendium, 2006)
Non Knowledge-Based Clinical Decision Support System
Non Knowledge - Structured CDSS does not apply any data from knowledge foundation however they used another kind of artificial brilliant called Machine Learning. From the word of Machine Learning, it means a machine will learn from the past experience and past lesson that given by experts. This kind of idea has applied in this kind of CDSS. Computer will learn everything in past medical progress and discover pattern in specialized medical data.
Non Knowledge - based mostly CDSS is trained from the relationship between symptoms and signs (also called independent factors) and diseases (also called dependent parameters). Machine Learning is using case-based to carry on every lessons because the machine has been trained from prior cases.
There are 2 type of non knowledge - established systems are manufactured neural sites and hereditary algorithms. It includes some mathematical models that can watch and emulate the properties of something plus some kind of adaptively learns the simulated properties of that. (Clinical Decision Support System, Citizendium, 2006). Manufactured neural networks kind of CDSS can examine the features or patterns from patient data to derive the associations between your symptoms and a analysis. (Wikipedia, 2010). It can perform supervised or unsupervised machine learning depending on way of providing the available information.
Genetic Algorithm is dependant on a several procedures of looking and simplifying and use the directed selection achieve ideal CDSS effect. The algorithm will first determine properties of packages of answers to a difficulty. Every solution that generated will be recombined, mutated and replicate the process again. The rotation of finding solution will not stop until an effective solution is found. The knowledge found in finding solution is derived from patient data. It usually focus on those disease that induced by narrow list of symptoms. (Wikipedia, 2010)
Architecture of Clinical Decision Support System
3. 1. Basic Idea of Decision Support System Architecture
Since Clinical decision support system is some sort of decision support system that is design to assist clinician in decision making tasks. The architecture design of decision support system always consists of two major sub-systems which is individuals decision maker and personal computers. Construct a conclusion support system with only computer hardware and computer software is not a correct theory because there could be some unstructured or semi set up decision (those decisions cannot be decide by using a collection of mathematical model or method) is not able to be programmed by system because it's accurately nature thinking from a real human which is elusive and complex. There is no such independent aspect in a decision support system. It always requires a human decision machine as another component of decision support system to combine with personal computers. The function of individuals decision maker is never to build a data source for decision support system. Instead of build a repository, it functions as a "decision maker" that delivers judgment, share their experience and exercises intuition throughout the complete procedure for decision making.
The very first step of decision making is get started with the creation of your decision support model (decision support model is the solution or the way that helps consumer to filter or decide the precise effect) by using some involved DSS program such as Microsoft Excel. System will connect to database through Repository Management Systems (DBMS) and deal the info from databases with your choice support model through Model-Based Management System (MBMS). DBMS is an application that used to create, take care of as well as control the access to the database. MBMS can be an application that inserted inside a DSS program that allow individual to produce, edit and delete your choice support model. By going through DBMS and MBMS, model can associate with the data from database to produce a specific decision.
DSS diagram. png
Figure 1. 0 Decision Support System diagram
The diagram above shows DBMS and MBMS is integrated with the DSS to communicate with the models and data source to provide lead to users.
3. 2. Four-Phase Style of Clinical Decision Support Architecture
Four-Phase Style of specialized medical decision support structures is discussing 4 type of architecture that is used in scientific decision support system development. These architectures also representing the evolutionary of professional medical decision support system. This 4 type of structures is standalone decision support system (1959), involved system (1967), standards-based system (1989), service models (2005). The stages is happen sequentially, every phase is discovered and inspired from previous phases.
Standalone Decision Support System
The first phase is Standalone decision support system which took place in 12 months 1959. They were systems that operate separately from professional medical system. The clinician got to purposely seek the system out and enter information of his medical conditions and then wait for the machine to interpret the effect. This sort of system is straightforward to build up because user that comes with medical knowledge and computer skills can make one of it. It is simple to talk about as well because the machine is easy to build up, it can be categorized as a straightforward system, consumer can just make a copy of this program and then mail to another who wishes to work with the system. You will find limitations such as they required end user to enter everything needed by the machine to make it inference. Another disadvantage is user surely got to seek out the way the system works and stream. Customer that is lack of medical knowledge may have problem in system utilization and might causes a whole lot of medical problem. Thus, they can not be proactive. In addition, it very frustrating, it may takes half to one hour to enter a case because the model's feature is very small and it required a whole lot of information to generate an output.
Integrated System
Due to the significant problems from standalone CDSS, developers started to require the architecture into another which is integrated system. The invented of Integrated system have resolved a whole lot of problems. To begin them is termination of multiple individual input. The information is stored electronically following the first input by the user. Another significant solution is system can be proactive. They can alert end user when it identify dangerous between drugs connections or the dosing mistake automatically. The major disadvantage of built in system is difficult to talk about. This technique is very complex because it directly constructed with large scientific system. Therefore, it can't straight share to others who are not using the same professional medical system. Unlike standalone system which built only predicated on self applied knowledge and computer skills. It could be send to anyone who wanted to utilize it. Another major problem is knowledge management problem. When there is an update for knowledge or medical guide, it maybe must find the foundation code to know where is guide used.
Standard-Based System
In order to make content sharable, several research and work had been performed to standardize scientific decision support content. The standardization of content has beat many disadvantage of designed system. It stocks the specialized medical decision support content by split the code that explaining this content from source code. However, it still has some restrictions. First, there exists way too much standard format to choose. You can find over hundred of standard to stand for a straightforward notification. Standardized encoded may constrain a user's standard. The "standard" that end user designed to write has the difficulty to appropriate for the "standardized standard".
Service Models
Service Models, the newest CDSS structures. It recombined medical information system and specialized medical decision support system components by utilizing a standard application development software (API). This models standardizing both professional medical decision support system and clinical system into one software. Both systems is only going to look of them costing only one specialized medical system and one CDSS at the same time although the knowledge about patient and remedies are across many places.
Clinical Decision Support's Algorithm
4. 1 Artificial Neural Network
Artificial Neural Network is a way that utilized by non knowledge-based CDSS. It required training from experts in a form of artificial intelligence. It will base on days gone by experiences or regarded examples to create a set of means to fix a medical problem. They have the "Human-Brain-Like" action rather than "Computer-Like". Because of the capability of knowing the "behavior" of problem through its experience, they are commonly used in identification problems. From the result, this methodology is very well in determining slim and well-defined scientific problem.
Three general kind of algorithm used by machine learning which is unsupervised, reinforcement and supervised.
Unsupervised Learning
Unsupervised learning means the computer identify some natural grouping inside a database by founded about how "similar" the things are and what makes a "Good" group without having to be provided examples of feature values of items. Therefore, the way of machine learning also known as "clustering". Alas, unsupervised learning is not being used in many studies of varied type of examination.
Reinforcement Learning
In reinforcement learning, it isn't provided any samples of feature principles of items. Rather than giving the examples, it is given a specific main point or feedbacks which have the ability to determine whether the system is on the right course.
Supervised Learning
In supervised learning, the computer is given the examples of feature value of items. The reason why of doing supervised learning is to build up a "classifier" that can predicts all the probability from given predetermined classes or samples based on a couple of qualities and features to spell it out the items.
4. 2 Bayesian Network
Bayesian Network shows a couple of variables and dependencies of conditional among the list of parameters via Directed Acyclic Graph (DAG). Each node in the graph signifies a changing and particular node will link to its neighbor to show the dependencies among the corresponding variables. This algorithm provides a simple understanding and classification between any two nodes. It can help anticipate and compute every probability event might occur in a particular condition. In the stand of medical view, it can compute every probability diseases by based on the symptoms given. For example, fever, cough, sore throat and chilling might trigger symptoms of Dengue disease.
There are two important aspect consists in this algorithm which are structure and a set of parameters. Structure of the Bayesian Network is constructed from DAG. Every node in DAG may be given value by the mother or father node. Guidelines are describing the partnership and the possibilities of the node to its parent or guardian. These components can support Bayesian Network computation by using the chain guideline. Therefore, parameter and structure learning must be undertaking to fully signify probability syndication. Parameter learning is to designate each node in DAG is about distributed based on varies conditional. Composition learning is to recognize just how of distribution throughout the complete network by predicated on the local data.
When learning Bayesian Network, the quantity of training data is very important and it straight influenced the correctness of the network. Therefore, training data must be provided enough through occupation of experts to provide various form of knowledge to improve the reliability of the models. Professionals might provide some knowledge that specifying a condition among the variables in Bayesian Network.
Bayesian Network Example. png
Figure 2. 0 Example of Bayesian Network
The example shows that fever and chilling maybe the symptoms of Dengue Disease. In another way, chilling maybe the medial side aftereffect of fever.
4. 3 Reasonable Condition
