Exam: Final 1 (Introduction to Medical Informatics) (http://www.cpmc.columbia.edu/edu/textbook) 1. What is the purpose of the HL7 data standard? 2. Why are sensitivity and specificity considered a better measure of a test's performance than accuracy (TP+TN/total)? 3. A 50 year old woman is considering elective cholecystectomy (remove the gall bladder). If she has elective surgery, then there is a 4% chance that she will die due to surgery. If she survives surgery, then she will lead a healthy life and her gall bladder cannot become infected. If she does not have elective surgery, then there is a 20% chance that her gall bladder will become infected, requiring emergency surgery. With emergency surgery, there is a 30% chance of dying. Assume that if she survives, then she will lead a healthy life. Draw the decision tree for this situation, and calculate the utility. Should she have the elective surgery? (Hint: since it is assumed that when she survives she is healthy, the utility for surviving can be set to 1.) 4. Compare a rule-based expert system (eg, MYCIN) to a probability-based system (including quasi-probability like QMR) for doing diagnosis in medicine. What are the relative advantages and disadvantages of each? 5. What are the obstacles to sharing electronic medical knowledge bases? 6. The separation of domain knowledge from problem-solving knowledge achieves what goal(s)? 7. What are the advantages of computer-aided instruction over traditional lecture-style instruction in medical education? 8. Why do clinical information systems in modern hospitals and office practices look so backward compared to what is available in a computer science department? 9. What is a "gold standard" in evaluation? What can be done if one does not exist? 10. What are the major functions performed by health- assessment information systems (what are they used for)?