SARANATHAN COLLEGE OF
ENGINEERING
DEPARTMENT OF COMPUTER SCIENCE AND BUSINESS
SYSTEMS
CBM356-MEDICAL INFORMATICS
PREPARED BY
S.SENTHIL ME.,(PhD).
UNIT IV- COMPUTER ASSISTED MEDICAL
DECISION-MAKING
TOPICS TO BE COVERED:
• Neuro computers and Artificial Neural Networks application,
• Expert system- General model of CMD,
• Computer–assisted decision support system-
• production rule system cognitive model,
• semantic networks,
• decisions analysis in clinical medicine
• computers in the care of critically ill patients,
• Computer aids for the handicapped.
Neurocomputers
• Neurocomputers are specialized computing systems designed to
emulate the way the human brain processes information.
• Artificial Neural Networks (ANNs) are computing models inspired
by the structure of biological neural networks.
Applications in Medical
Decision-Making:
• Diagnostic Assistance: ANNs can assist in diagnosing diseases by
analyzing medical images (e.g., X-rays, MRIs) or interpreting clinical
data. They are trained to recognize patterns associated with specific
conditions.
• Predictive Analytics: Neural networks can be used to predict patient
outcomes, such as the likelihood of developing certain conditions, the
success of treatment plans, or the progression of a disease.
• Treatment Recommendations: They can help recommend
personalized treatment options based on patient data, historical
cases, and existing medical knowledge.
• Medical Image Analysis: ANNs are commonly applied to detect
abnormalities in radiological images, such as tumors or fractures,
with high accuracy.
Advantages of Using ANNs in
Medicine
• Improved Accuracy
• Efficiency
• Continuous Learning
General model of CMD
Expert system
• ES is an AI-based software that mimics human expertise in medical
diagnosis and treatment.
• ES utilize knowledge from medical experts, encoded in a knowledge
base
• ES uses algorithms to assist healthcare professionals in making
decisions.
Components
• Knowledge Base- Contains medical knowledge, including symptoms,
diseases, diagnostic criteria, and treatment protocols.
• This knowledge can be acquired from medical literature, expert input,
and clinical guidelines.
• Inference Engine- Applies reasoning techniques to the knowledge
base to draw conclusions or make recommendations.
• It uses rules to analyze patient data, suggest diagnoses, or
recommend treatments.
Contd..
• User Interface- Allows healthcare providers to input patient data
(symptoms, lab results, history).
• It also presents the system's recommendations, explanations, or
confidence levels.
Types of Expert System
• Rule-Based Systems:
• Use a set of "if-then" rules for making decisions.
• if a patient presents with a specific set of symptoms, the system can suggest
possible diagnoses.
• Case-Based Reasoning Systems:
• It is for find solutions for new cases based previously solved cases.
• It compares the current patient case to past cases in the database to suggest
a diagnosis.
Contd..
• Probabilistic Systems:
• Use Bayesian networks to handle uncertainty
• infer the probability of certain conditions given patient data.
• Machine Learning-Based Systems:
• Its an data-driven algorithms to learn from medical data
• improve decision-making accuracy over time.
• it can identify patterns not explicitly encoded in the rules.
Applications:
• Diagnostic Support
• Treatment Recommendations
• Monitoring and Management of Chronic Diseases
• Radiology and Imaging Analysis
Benefits:
• Enhanced Accuracy: Provides consistent recommendations based on
evidence and reduces human error.
• Time Efficiency: Allows for quicker decision-making by providing
diagnostic and treatment suggestions.
• Knowledge Accessibility: Makes expert knowledge accessible to
general practitioners, especially in remote areas.
production rule system
• production rule system is a cognitive model widely applied in
computer-assisted medical decision-making
• knowledge is represented in the form of "if-then" rules.
• Rule 1: If a patient has a high fever and a sore throat, then consider a
possible diagnosis of a throat infection.
• Rule 2: If a throat infection is diagnosed, then prescribe antibiotics.
Contd.
• Forward Chaining: Starts with the available data (facts) and applies
rules to infer conclusions or actions. For instance, based on the
patient's symptoms, the system works through rules to identify a
possible diagnosis.
• Backward Chaining: Begins with potential conclusions and works
backward to determine whether the facts support them. For example,
if a specific diagnosis is considered, the system checks if the patient’s
symptoms align with that condition.
Contd.
• Working Memory: Stores temporary data or facts about the patient's
condition as the inference engine processes rules. It gets updated
dynamically during the decision-making process.
• Conflict Resolution: When multiple rules are applicable at the same
time, conflict resolution strategies prioritize which rule to apply first.
Semantic networks
• semantic networks play a crucial role by modeling medical knowledge
and the relationships between various medical concepts
Contd.
• Components:
1.Nodes (Concepts):
1. Represent entities such as symptoms, diseases, treatments, laboratory tests, and
anatomical parts.
2. nodes might include "Fever," "Bacterial Infection," "Antibiotics," and "Cough."
2.Edges (Relationships):
1. Represent the connections between the nodes. These connections can be labeled to
indicate the type of relationship, such as:
1. "is a symptom of" (e.g., "Fever" is a symptom of "Bacterial Infection").
2. "treats" (e.g., "Antibiotics" treat "Bacterial Infection").
3. "is associated with" (e.g., "Cough" is associated with "Respiratory Infection").
Contd.
• Semantic network encodes medical knowledge in a structured, graph-
based format,to understand complex medical information.
• represent various relationships in medicine, such as symptom-
disease links, disease-treatment connections, and drug interactions
• network can be used to infer potential diagnoses or suggest
treatments
• if a patient presents with "Fever" and "Cough," the network can help
identify likely conditions that are associated with both symptoms,
such as "Influenza" or "Pneumonia."
Contd.
• Semantic networks can assist healthcare providers in making
decisions by presenting relevant medical information
• to narrow down the list of possible conditions based on patient
symptoms and history
Advantages of Using Semantic Networks in
Medicine:
1.Clear Visualization of Medical Relationships
2.Scalable and Extendable Knowledge Base
3.Integration with Other Systems
Future Enhancements
• Combining with Ontologies
• Dynamic and Adaptive Networks
• Use of Natural Language Processing (NLP)
Real-World Examples
• UMLS (Unified Medical Language System)
• SNOMED CT
Treatment plan
