Mini Expert System for Emergency Radiation Accident Patient Management C.J.Liao 1, P.C.Chen 1 and M.H.Ou 2 1 Undergraduate School of Medicine 2 Graduate.

Mini Expert System for Emergency Radiation Accident

Patient Management

C.J.Liao1, P.C.Chen1 and M.H.Ou2

1Undergraduate School of Medicine

2Graduate School of Pharmacy

Kaohsiung Medical University, Kaohsiung, Taiwan

TAIWAN

• Population: 22,370,461• Land Area: 32,260 sq km

Motivation

• Nuclear power project has long been a controversial issue in Taiwan

• March 18, 2001: A site emergency event Level 3A was triggered of the Third Nuclear Power Plant.

Motivation (cont.)

• Drill of emergency medical management of multiple-casualty radiation accident

• Doctors looked up the books and journals to make decision when patients of various conditions came in

• They even worked on the patient while asked others how to do for next step

Motivation (cont.)

• Victims of radiation accident are very rare.

• Not so proficiently

• Can we help them?

Radiation Accident

• Accident: An occurrence that results in personal injury, disease or death or property damage

• Radiation Accident: An accident correlated to ionizing radiation

Types of Radiation Accident

• Industrial and medical uses of radiation

• Transportation accidents

• Radiography over-exposures

• Released from fixed nuclear facility (e.g., reactor)

Types of Radiation Accident (cont.)

• Malicious/terrorist events: – Radiological dispersal devices – Intentional irradiation – Nuclear weapon

Chernobyl Nuclear Power Plant Disaster

• 01:23, 1986.04.26• Chernobyl, former USSR• Killed more than 30 peop

le immediately• High radiation levels in t

he surrounding 20-mile radius, 135,00 people evacuated

• Over 2,500 deaths

Radiation Bioeffect

• Deterministic Effects

(Non-stochastic effect)

• Stochastic Effects

• Radiation Hormesis

Deterministic Effects

• They have a threshold

• They are dose related

• They are predictable

• There is a gradation in response

• If the dose is received over a longer period the effects will be less

Stochastic Effects

• The severity of the response is independent of the dose

• The total dose given is more important than the time period over which is was given

• Effects may present many years later

• Cancer is the commonest stochastic effects

Radiation Protection Guidelines

• Time: the shorter, the safer

• Distance: the farther, the safer

• Shielding: the more, the safer

• Quantity: the less, the safer

Exposure

• Occurs when entering an area there is a radiation source

• The patient does not become radioactive.

• Example: Receiving a chest X-ray

Contamination

• External contamination: Radioactive material on the surface of a patient

• Internal contamination: Some radioactive material entered into the body

• The patient may have radiation hazard to others

Hospital Emergency Care of Radiation Accident Patient

• REAC/TS. Guidance for Radiation Accident Management. <http://www.orau.gov/reacts/guidance.htm>

• Mansfield G. Planning for Management of Radiation Incidents. <http://www.llnl.gov/seaborginstitute/training/planning_radiation.pdf>

• Wagner RH, Henkin RE, Halama JR. The Medical Management of Radiation Accidents. <http://www.meddean.luc.edu/lumen/MedEd/radiation/Radaccident/HOMEPAGE.html>

DSS vs. ES

• From UMDNJ Dr. Mital’s teaching materials……

• (about this portion, I’ll rearrange it later on…… And in practical, we do not make a clear distinguish between DSS and ES, and in this paper we use ES or CDDSS for this kind of systems….)

FORMAL DEFINITION OF A DECISION SUPPORT SYSTEM

A DSS is an interactive, flexible and adaptive system, specifically developed for supporting solutions of management decision problems for improved decision making or diagnosis of some problems.

It supports all the phases of the decision making and includes a knowledge-base.

In order for this to be successful, the system must be:

• Simple • Adaptive

• Robust • Complete , on important issues

•Easy to control/use • Easy to communicate

COMPONENTS OF DSS

1. Data Management: Includes a database and is managed by the DBMS software.

Expert systems (Decision Support Systems):

When an organization has a complex decision to make or a problem to solve, it has to turn to experts, which could be difficult to find and at times very expensive.

Expert systems are attempts to mimic human experts

Expert systems are branch of “applied artificial intelligence”. They are used frequently in medical diagnostic systems, mineralexploration, managing assets and liabilities, corporate planning, organizational management and administration, etc.

Differences Between DSS and ESDSS ES

Objective Assist human decision maker

Replicate (mimic) human advisers and replace them

Who makes the recommendations (decisions)

The human and the system The system

Major orientation Decision making Transfer of expertise and rendering of advice

Major query direction Human query the machine Machine queries the humans

Nature of support Personal, groups, and institutional

Personal (mainly), and groups

Characteristics of problem area Complex, integrated, wide Narrow domain

Types of problems Ad-hoc, unique Repetitive

Content of database Factual knowledge Factual and procedural knowledge

Reasoning capability None Yes, limited

Manipulation method Numerical Symbolic

Computerized Decision Support Systems

• DSS: Decision Support Systems

• ES: Expert Systems

• GDSS: Group DSS

• EIS: Executive Information Systems

• ANNs: Artificial Neural Networks

DSS-ES connections

These two seems to be completely different. These differences are technological as well as managerial.

Because of different capabilities, they can complement each other, creating powerful and integrated computer based systems that can improve managerial decision making capabilities ( or process).

Clinical Diagnostic Decision Support System (CDDSS)

• A computer-based algorithm that assists a clinician with one or more component steps of the diagnostic process

• For most of the CDDSS to work, they must consist of two basic parts: – Knowledge Base– Inference Engine

Clinical Application of CDDSS

• Reported by many institutes• Most of them began as funded academic

research projects • Not all of them showed a favourable result

• Wolf FM, Friedman CP, Elstein AS, Millerd JG, Murphy GC, Heckerling P, Fine P, Miller T, Sisson J, Barlas S, Capitanod A, Ngc M, Franzc T. Changes in Diagnostic Decision-making After a Computerized Decision Support Consultation Based on Perceptions of Need and Helpfulness: A Preliminary Report. <http://www.amia.org/pubs/symposia/D004169.PDF>

CDDSS in Emergency Medicine

• Usually designed for a specific purpose • Acute myocardial infarction and/or ischemia

– Outer eye: Computers in Nursing. 19(3):114-7, 2001

– HAZMATs: Journal of Hazardous Materials. 71(1-3):503-21, 2000

– Chest Pain: • Cardiology. 92(2):128-34, 1999

• Artificial Intelligence in Medicine. 10(1):75-92, 1997

• Computer Methods & Programs in Biomedicine. 52(2):93-103, 1997

• Academic Emergency Medicine. 3(7):689-93, 1996

• J Cardiovasc Nurs. 16(3):1-8, 2002

Stryer DB. The development and role of predictive instruments in acute coronary events: improving

diagnosis and management.

J Cardiovasc Nurs. 16(3):1-8, 2002

• “Approximately 6 million patients with chest pain are seen annually in Eds. Together, AMI, angina, and nonspecific chest pain accounted for more than 2.5 million admissions and $47 billion in hospital charges in 1997. In 1998, AMI resulted in nearly 460,000 deaths, the leading cause of mortality in US……”

Evaluation of the Goal

• Frequency -> very rare – Not proficiently– Can not afford an expensive system

• Doctors looked up the books -> Need experts’ help, possible contamination

• Asked others -> Need experts’ help• Exposure -> Shortest time• Contamination Control -> Mouse operation only• Knowledge -> Rule-based: “IF…THEN…”

Evaluation of the Goal (cont.)

• To build a simple, low-cost and easy-to-use mini expert system for emergency radiation accident patient management

Inference Engine Selection

• Free 30-day Exsys CORVID ver. 2.0 evaluation program

• Up to 150 nodes

• Web-based

Knowledge Base Building

• Extract knowledge from the documents

• Translate into “IF…THEN…”

• Cascade the rules to form a “decision tree”

• Define variables and values

• Divide “decision tree” into “logic blocks”

• Create “single variable” node to add important notices

Implementation

• Copy the HTML file and “ExsysCorvid.jar” to a floppy disk and rename the HTML file to be “index.html” -> Run from disk

• Transfer “index.html” and “ExsysCorvid.jar” to certain directory under httpd server root directory of a web server -> Run from web

Discussion

• CDDSS applications seldom last long– Funded research projects – No knowledge base update

• An exception: AMI/angina/chest pain– Markets (Population)– Mortality

Discussion (cont.)

• Should CDDSS be a sophisticated and expensive monster and can only serve the purpose of million patients a year.

• In this project, we may provide a different point of view to build a simple mini expert system for a different kind of utility.

Conclusion

• A simple, low-cost and easy-to-use mini expert system for emergency radiation accident patient management is built using the free 30-day Exsys CORVID evaluation program

• http://coolylab.kmu.edu.tw/wc2003test/Radptmgm.html

• Further clinical testing is necessary