Clinical Decision Support Systems




1. d, because the inference engine does not have to use deductive logic, but may proceed by induction and pragmatic rules. Although it is possible to have tools which require a specific core set of data to be provided completely, in most medical situations it is impractical to require completely uniform data collection before a decision can be made, so choice c is not correct because flexibility is "generally required." Note that the Problem Oriented Medical Record is not a "decision support tool for general consultation," but rather is a structured way for the practitioner to organize his or her thinking and data collection about a patient's problem(s).

2. c

3. e

4. No. There's no point in doing the test unless it might change treatment. Only a negative test could conceivably justify a revised decision not to operate. Does it? The pre-test odds of appendicitis are 3:7, the likelihood-ratio of a negative test is 1/3, the post-test odds of appendicitis are thus 1:7, and the NPV is 12.5%. The utility of not operating when there is a 12.5% chance of appendicitis is .125x.99+.875x1=99.88%. The utility of operating, however, is 99.90%. The test result, whatever it might be, would not change therapy. Whatever time and money go into the test are, under these conditions, wasted. Why do it?