Abstract
Test evaluations should focus on the likelihood that tests detect clinical events of interest and the effect that tests can have on these events by the way in which test results affect subsequent management decisions. Randomized controlled trails of medical tests are feasible and several designs are possible. Randomized controlled trails of medical tests can be made more efficient by randomizing only patients with the test result of interest. A randomized controlled trail of medical tests should incorporate a prespecified link between test and treatment options to ascertain validity and generalizability. Sample-size calculations need special attention and have to include an estimation of the discordance rate.
| Original language | English |
|---|---|
| Title of host publication | The Evidence Base of Clinical Diagnosis |
| Subtitle of host publication | Theory and methods of diagnostic research: Second Edition |
| Publisher | Wiley-Blackwell |
| Pages | 63-82 |
| Number of pages | 20 |
| ISBN (Print) | 9781405157872 |
| DOIs | |
| Publication status | Published - 2 Feb 2009 |
Keywords
- Alternative randomized designs
- Basic RCT of single diagnostic test
- Comparing test strategies
- Diagnostic testing and prognosis
- Flexible endoscopic valuation of swallowing with sensory testing (FEESST)
- Intrauterine growth retardation (IUGR)
- Prognostic impact of tests evaluation
- RCT of two tests randomizing only discordant results
- Randomizing abnormal test results
- treatment effect in test normals and treatment effect in test abnormals