Yur'ev Dmitriy

Faculty of Computers and Information Science

Speciality "computer systems of medical and technical diagnostics"

Theme of master's work:

"The development of specialized computer systems for the diagnosis of biochemical analysis of blood"

Scientific adviser: Ph.D., Assistant Professor in the Faculty of CIS Yaroshenko Nikolay

Introduction. Rationale for the relevance of the theme

Biochemical analysis of blood — this is a laboratory method of investigation used in medicine, which reflects the functional state of organs and systems of the human body. It allows you to evaluate the work of many internal organs including liver, kidney, active inflammatory rheumatoid process, and violation of water-salt metabolism and imbalances of trace elements. Biochemical analysis of one of the methods of laboratory diagnosis, which is very informative for the physician and has a high degree of reliability. Therefore, it can not only reveal the full picture of the functioning of an organ, but also tell whether a person is experiencing shortage in some minerals or vitamins. Biochemical analysis helps diagnosis, assign and adjust treatment as well as to determine the stage of the disease. Currently, treatment of the analysis and the diagnosis is in hand.

Pressing problem in medicine is to automate setting the preliminary diagnosis of diseases by biochemical analysis of blood and urine.

Purpose and objectives

The development objective of SCS is the statement of the preliminary diagnosis on biochemical analysis of blood and urine.

To achieve the stated goal of SCS is necessary to develop that will accomplish the following tasks:

• Creating a database containing information about the results of biochemical analysis of patient data.

• Analysis and choice of treatment outcome data.

• Development of a mathematical model of diagnosis.

• Development of software for setting the preliminary diagnosis.

• Development of technical support SCS.

Alleged scientific novelty and practical value

Biochemical analysis of blood pulled twenty five indicators by which one can assess the condition of man. Indicators of the analysis are as follows: total protein, total bilirubin, direct bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase, amylase, gamma-glutamyltransferase (gamma-GT), creatine kinase, cholesterol, triglycerides, glucose, urea, creatinine, uric acid, iron, magnesium, phosphorus, calcium, chloride, lactate dehydrogenase (LDH), potassium, sodium, and C-reactive protein, rheumatoid factor, antistreptolysin-O.

There are certain rules of the biochemical analysis of blood - that is, the number of indicators that should be present in the blood of a certain age and gender. This is statistically established indicators of healthy people is a reference. Deviation from these parameters are the symptoms of various disorders in the organism's activity, the failure of certain organs or organ systems.

Novelty of the project is to develop a mathematical model of the automation process of diagnosis. The practical value of the work lies in the fact that the developed system will automatically process the data and developed models to put a preliminary diagnosis of the patient's final decision makes the diagnosis of a doctor, that accelerates and optimizes the performance of health workers in diagnosis.

A review of existing research and development on

Analysis of existing diagnostic tools

Biochemical blood analyzer uses a mechanical, optical and computer technology to obtain the value of the concentration of either substance in the blood. Biochemical analyzers allow us to determine the levels of enzymes (amylase, ALT GGT and so on.) Substrate (bilirubin, glucose), minerals (sodium, potassium), fats (cholesterol, triglycerides). There are different analyzers, which differ in the number of tests that are conducted, performance, reliability, but the structure of all the biochemical analyzers are divided into two types: semi-automatic and automatic.

In semi-automatic biochemical analyzer doctor conducts mixing of reagents, the necessary calculations, and the analyzer to measure, heating, data processing, printing out. Number of tests performed by such analyzers are not large, high complexity, not excluded the human factor on test results.

Such shortcomings are not in an automatic biochemical analyzer, which conducts dosing, mixing rinsing, processing, calculations, printing their own. Biochemical analyzers of this class are used large laboratories that maximizes save time and get accurate results of the study.

But all these systems have common flaws - they require additional processing of the results, which takes time. To solve this problem, create a dedicated SCS. Developed by SCS will improve the quality of diagnosis at the expense of today's electronic database of patients that will automate the process of diagnosis.

Research

Development of a database containing information on
results of biochemical analysis of patient data

Requirements for information management database include:

• Creating a database should be using modern development tools;

• Automated entry and processing of information stored in the database;

• Support the integrity and secure data storage
  

  that is an essential factor in the projected database;

• Provide input of primary documents, using appropriate forms
  

  enter and complete the appropriate forms;

• Organization validation of input data;

• Implementation of data access requests in the formation;

• Ensuring high performance and data access speed, and generate reports;

• To protect against unauthorized access to the database, with user authentication.

Description of the method of decision-making process of the analysis

The basis of any decision support system (DSS) is a body of knowledge, structured in order to simplify the decision process. For experts in the field of artificial intelligence term knowledge means the information you need the program to behave "intelligently." This information takes the form of facts and rules. Facts and rules in the DSS are not always either true or false. Sometimes there is some degree of uncertainty about the reliability or accuracy of the facts rule. If this doubt is expressed explicitly, it is called the "coefficient of confidence."
The coefficient of confidence - a number that indicates the likelihood or degree of certainty with which you can consider this a fact or a rule or a reliable fair. This ratio is an estimate of the degree of confidence in the decision issued by the system. The theory of self proposes the concept of trust and distrust. These concepts are independent of each other and thus can not be combined in the same way as probabilities, but they can be combined and presented in a mathematical model of diagnosis presented below. The principles of decision support systems, knowledge-based, are illustrated in Pic 1. User transmits to the system or the facts in this case, the symptoms and gets as a result of the expertise or the diagnosis. In its structure, the expert system is divided into two main components - the knowledge base and inference machine. The knowledge base contains the knowledge (evidence) on which the machine forms the conclusion of the inference (diagnosis). These findings represent the responses decision support system to the user wishing to obtain expertise.

Picture 1 — Basic principles of operation of a decision support system

Development of mathematical models of diagnosis

List of problems solved with the help of mathematical methods:

• Ensuring the adequacy and accuracy of the results;

• Providing high-speed processing of data.

Ensuring the adequacy and accuracy of the results lies in the DSS From the above stated, that the theory of self proposes the concept of trust and distrust. These concepts are independent of each other and thus can not be combined in the same way as probabilities, but they can be combined in accordance with the following formula:
CC (B, C) = MD (B, C)-MN (B, C), where CC - ratio confidence MC - a measure of confidence, MD - a measure of distrust P - the probability, C - certificate or event. There are several ways to achieve goals, each with a different CC for a given set of facts. When we have a system based on knowledge of several interrelated rules, each of which is doing the same conclusion, but with a different coefficient of confidence, then each rule can be considered as part of the evidence that supports the joint conclusion. Result of the combination is calculated as:
CC (P1, P2) = CC (P1) + CC (P2) [1-CC (A1)].
Providing high-speed processing of data is the complexity of the diagnostic construct of Production of expert systems (ES), and the hardware capabilities of SCS.

Consider one of the approaches to the diagnostic of Production of ES. The corresponding interpretation can be different and therefore the domain of applicability is discussed structures wide enough. A given large number of facts

that consists of elements of two types. Elements a_ij determine the normal declarative knowledge of a particular subject area. Elements q_i determine the type of interaction with the environment and in this case are questions of the user as an alternative menu:

Some of q_i,have a different meaning - the resulting conclusions or diagnoses, designed in the form of appropriate messages to the user. Production in this system are

All the facts and a lot of productions are organized into a system that is represented as a graph, "OR". A fragment of a graph with vertices of any screening, diagnosis q₉, q₁₀, q₁₁, q₁₂ (terminal nodes) is shown in Pic. 2 [2].

Picture 2 — Diagnosing count

Requirements for software development,
for setting the preliminary diagnosis

Software should be a set of programs to help solve functional problems in a specialized computer system. The software includes general operating system software and special software.

Overall the system software should include:

• Operating system;

• Drivers for peripheral devices;

• Anti-virus systems;

• Archivers.




Special software:




• Should be built on a modular principle;

• Must include a software module of the planning process;

• The software should be an open system and should provide;
  

  The possibility of extending the functionality;

• Provide the ability to add new modules and changes to existing ones.

Requirements for the development of technical support SCS

The technical part of the system includes: CPU, motherboard, memory, chipset, zhostky drive, monitor and video card, keyboard, mouse, printer.

Requirements for Maintenance:

• CPU socket type must match
  

  selected system board with a network interface;

• The clock frequency of at least 2 GHz;

• Bus bandwidth of 800 MHz;

• 512 KB second-level cache;

• Number of Cores - 2 or more;

• The capacity of RAM at least 800 MHz;

• Monitor LCD 17 inches;

• HDD drive - 7200 rev / min.

• CD-DVD optical drive;

• Keyboard;

• Optical mouse;

• Printer, scanner;

Conclusion

1. The chosen method of forming the preliminary diagnosis, the calculation of their probabilities.

All calculations are based on the values ??of the characteristics of blood and urine tests (performance),

obtained by biochemical analysis of blood and urine.

2. The basic parameters of the biochemical analysis of blood and their impact on the relevant disease.

3. Developed by SCS, which includes the program "Diagnosis" for pre-order

дdiagnosis, and a database that stores personal data of the subjects, the doctors, the user database,

adjustment program and the results of surveys and actually own diagnoses.

4. Automated processing of studies, and it has an important role in medicine, including reduces

time spent on the survey and display indicators are not in their

allowable range. Automation also reduces the errors that appear

by human factors.

References

1. Данилова Л.А., "Анализ крови и мочи" - 4-е изд., исправ. СПб.: "Салит-Медкнига", 2003г. - 128с.

2. Корячкин В.А., Страшнов В.И., Чуфаров В.Н. "Клинические функциональные и лабораторные тесты в анестезиологии и интенсивной терапии" - 2-е изд., перераб., и доп. СПб.: Санкт-Петербургское медецинское издательство. 2004г. - 304с.

3. Лившиц В.М., Сидельникова В.И. "Медецинские лабораторные анализы" Справочник - М.: "Триада-Х", 2000г. - 312с.

4. Хмелевский Ю.В., Усатенко О.К. "Основные биохимические константы человека в норме и при патологии" - 2-е изд., перераб. и доп. К.: "Здоров'я", 1987г. - 160с.

5. Руанет В.В. "Теория и техника лабораторных работ" учеб. пособ.,/Под ред. проф. Хетагуровой А.К. - М.:"ВУНМЦ Росздрава", 2007г. - 176с.

6. Назаренко Г.И., Кишкун А.А. "Клиническая оценка результатов лабораторных исследований" - М.: "Медецина", 2000г. - 544с.

7. Продеус А.Н., Захрабова Е.Н. "Экспертные системы в медицине" - М.: "Мир", 1991г. - 408с.

8. Маршалл В.Д "Клиническая биохимия" - М.: "Бином" 2002г. - 383с. [Электронный ресурс] &mdash Режим доступа: http://patfiza.net/books/med/biochem/Marshall_Klinicheskaya_bioximiya_%281999%29%28ru%29%28373s%29.djvu

9. Черноруцкий И.Г. "Методы принятия решений" — СПб.: "БХВ-Петербург", 2005г. — 416с.

10. Бочков В.Н., Добровольский А.Б., Кушлинский Н.Е., Логинов В.А., Панченко Е.П., Титов В.Н., Ткачук В.А. "Клиническая биохимия" учеб. пособ., "ГЭОТАР-МЕДИА" 2008г. - 264с.

Note

The master's work has not completed yet. Completion date is December 2011. Full text can be received from the author or his scientific advisor after this date.