Warning: Undefined variable $zfal in C:\Inetpub\vhosts\kidney.de\httpdocs\mlpefetch.php on line 525
Deprecated: str_replace(): Passing null to parameter #3 ($subject) of type array|string is deprecated in C:\Inetpub\vhosts\kidney.de\httpdocs\mlpefetch.php on line 525
Warning: Undefined variable $sterm in C:\Inetpub\vhosts\kidney.de\httpdocs\mlpefetch.php on line 530
Warning: Undefined variable $sterm in C:\Inetpub\vhosts\kidney.de\httpdocs\mlpefetch.php on line 531
English Wikipedia
Nephropedia Template TP (
Twit Text
DeepDyve Pubget Overpricing |
lüll Statistical techniques for medical decision making: applications to diagnostic radiology Doubilet PMAJR Am J Roentgenol 1988[Apr]; 150 (4): 745-50The average radiologist will never compute a probability by using Bayes theorem or carry out a logistic regression, ROC, or cost-effectiveness analysis. Why, then, do we think that radiologists should be familiar with these techniques? Radiologists base many decisions on information gleaned from the published literature. In the past decade, the techniques discussed here have begun to appear in medical (including radiologic) publications. A radiologist with no comprehension of ROC analysis, for example, would be unable to critically assess a study that used this technique to compare MR imaging and CT in the detection of liver metastases and would have to accept or reject its conclusions blindly. A knowledgeable radiologist, on the other hand, could judge whether the study employed proper methodology, and could accept or reject its conclusions on that basis. The four cases presented here are specific examples of generic problems facing the radiologist: predicting the likelihood of disease on the basis of a test result, using several pieces of information provided by a single test to arrive at a diagnosis, comparing the efficacy of radiologic tests or interpretive techniques, and choosing among available tests or procedures on the basis of their relative cost-effectiveness. It is likely, therefore, that the techniques of medical decision making discussed here will appear with increasing frequency in the radiology literature. The list of potential applications is long. Some of the questions that can be addressed by the techniques presented here are: How predictive of IUGR are the various proposed Doppler criteria? How can they be used in conjunction with conventional sonographic criteria to diagnose IUGR? Which technique is best for detecting prostate cancer or for staging known cancer--sonography or MR imaging? Which MR pulse sequence is best for a variety of organ systems and clinical indications? Is routine screening obstetric sonography cost-effective? Applications of statistical decision-making techniques to these and related questions will improve the quality of health care provided by radiologists, if the statistical techniques are done properly and understood by the intended audience.|*Decision Support Techniques[MESH]|*Radiology[MESH]|Bayes Theorem[MESH]|Cost-Benefit Analysis[MESH]|ROC Curve[MESH]|Regression Analysis[MESH] |