.
Monday, October 21, 2019
Clinical Decision Support Systems Essay Example
Clinical Decision Support Systems Essay Example Clinical Decision Support Systems Paper Clinical Decision Support Systems Paper Abstract     This paper explores the world of clinical decision support systems(CDSS) and the effects they have in the work place. Also how many different types of CDSS such as Bayesian Network (BN), Neural Network(NN), Genetic Algorithms(GA),and MYCIN. This paper will describe key reasons why a decision support system might fail to meet a users expectations and suggest measures that could be taken to avoid such an outcome. Finally this paper will explain what decision tree induction(DTI) means and how it used for classification problems.     While alarms and alerts are the most common visible CDSS interventions, there are many more. I will describe many different types of CDSS and their uses in relation to their health care usage. CDSS have been being developed for years for use in the health care society. for the purpose of assistance of diagnosing and treating various illnesses. There is no defined outline for CDSS as they adapt to the needs of physician and patient. Throughout this paper I will introduce various CDSS and their uses.     Most CDSS consist of the knowledge base, inference engine, and a communication mechanism. The knowledge base contains the rules and guidelines of gathered data . With this physicians can program a CDSS to think in an IF-THEN sense. It also allows to user to input additional or new information regarding the diagnoses or treatment of diseases. For example the if the machine is reading a diabetics blood sugar level and it is low then it alerts the physicians. However, some CDSS use a form of artificial intelligence to compile data, read the data, compare it to that of the patient and produce results. In other words it is basically a learning computer in which it takes information from previous account and uses them in the present.     Artificial Neural Networks (ANN) is an adaptive CDSS that uses a form of artificial intelligence, that allows the systems to learn from past patients traits and symptoms and apply an educated guess on the symptoms of a current patient. It consists of nodes called neurodes and send signals between the neurodes in a sort of straight line approach. An ANN consists of 3 main components: Input (receiving data ), Output (informing of possible diseases) and Hidden (data processing). The system becomes more and more effective as it collects data from many patients.      ANN has many advantages such as the virtual elimination of needing to program the systems and providing input various information. The ANN CDSS can process incomplete data by making educated guesses about all other data collected from past experiences. Additionally, ANN systems do not require large data banks to store all of its information in. However ANN does have disadvantages as well such as the training process to new users may be very time consuming leading users to not make full use of the systems. Another down side being disease harboring similar symptoms may be mistaken for one another. Thus the ANN is a double-edged sword.     Another less commonly used CDSS is a Genetic Algorithm (GA) a method developed in the 1940s at the Massachusetts Institute of Technology based on Darwin’s evolutionary theories. These algorithms form different combinations that are better than the previous solutions. Much like neural networks, the genetic algorithms receive their information from patient data.     Advantages of genetic algorithms are the fact that these systems go through an iterative which means to solve a problem through many assumptions of an initial guess. This process is used to produce an the best solution. The wellness/fitness function determines calculations from the ones that can be discarded. A disadvantage is the lack of leniency in the reasoning involved making it a less than first choice for physicians and clinicians. The obvious challenge in using genetic algorithms is in defining the wellness/fitness criteria. For genetic algorithms to work correctly there must be many factors such as multiple drug use, or multiple therapies. The Bayesian network(BN) is a graphical representation that shows a set of variables and their probable relationships between illnesses and symptoms. They are based on conditional probabilities, the probability of an event in relation to the probability of another event. Bayes’ rule helps physicians compute the probability of an event with the help of readily information and it processes options as new information is presented.     Some of the advantages of BN include the knowledge and conclusions of experts in the form of hypothesis, assistance in decision making as new evidence is available and are applicable to many models     MYCIN(MY) is a CDSS designed to diagnose and recommend treatment for certain blood infections such as meningitis. It has been extended to handle other infectious diseases. MY operated off what physician call if-then statements which means, if certain factors are present then a certain illness is present. It was a goal-directed system, using a sort of reverse thinking.     In the ordering phase of a clinic or hospital it is important to take all factors into account. The clinicians and physicians must look at what it is they specialize in what CDSS is best for that particular area. Also have CDSS can accommodate for other fields as well. As stated in Ten Commandments for Effective CDSS [Bates et al.2003], speed is everything. This means how quickly the CDSS gathers information and formulates a course of treatment. Also stated is the CDSS must fit into the users work flow, no time can be wasted trying to get a machine to de exactly what you want it to.     Reasons a CDSS may not live up to clinicians or physicians expectations are abundant. First, a machine is just that a machine it can only do what it is programmed to do. Second, It takes time to learn to use and machine and even after learning operation mistakes do happen. Finally, the thing most machines need and most clinicians and physicians dont have is time, the time to program and operate the machine exactly how they want to it to work. This can conflict with work schedules as well as treatment schedules.     There are ways to prevent using a seemingly useless CDSS. The most important thing is results, make sure the CDSS is capable to producing quick accurate results. Other things that should be kept in mind is that staff must be trained to use the machine make sure you have time to do so. Have the knowledge and ability to use the CDSS to its fullest potential while at the same time produce speedy results. It is stated in Ten Commandments for Effective CDSS that â€Å"All health professionals in the United States face increasing time pressure and can ill afford to spend even more time seeking bits of information†again meaning peed and time are everything.     In CDSS decision analysis, a decision tree is a support tool that uses a graph of decisions and their possible outcomes, including chance possibilities, various treatments, and diseases. A decision tree is used to identify the strategy to reach a course of action suitable for the patient. Another use of trees is to calculate rare cases of a certain disease.      a decision tree is a predictive model mapping from observations about an item to conclusions about its optimal outcome. Here is an example of a decision tree: In conclusion CDSS can be seen as a very useful tool. They are considered one of the best course of actions in the medical field. Though they may have some setbacks these are easily avoidable with the right information. Thus making CDSS a suitable tool of the health care industry. As long as the clinicians or physicians knows what they need the machine for and its capabilities based of their priorities they are a must have advancement in medical technology. [Bates et al.2003]Bates, D. W., Kuperman, G. J., Wang, S., Gandhi, T., Kittler, A., Volk, L., Spurr, C., Khorasani, R., Tanasijevic, M., and Middleton, B. (2003). Ten commandments for effective clinical decision support: making the practice of evidence-based medicine a reality. J Am Med Inform Assoc, 10(6):523–530. Clinical Decision Support Systems; Wikipedia: The free encyclopedia. (2004, July 22). FL: Wikimedia Foundation, Inc. Retrieved 02/20/09 from wikipedia.org
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment