Skip to main content

Advertisement

Comparing diagnostic tests and biomarkers: trials in people with discordant test results

Objectives

Diagnostic tests are traditionally compared for accuracy against a gold standard, but there is growing interest in tests (or biomarkers) used to guide treatment choices rather than specifically to diagnose. The question of whether health outcomes are better using one test or another in the same population can be answered with a randomised trial. It has been suggested that an efficient approach to trialling is to give both tests to all participants, and randomise and follow up those with discordant results. We describe how to plan and analyse such a trial, and consider its efficiency compared with a conventional trial design.

Methods

We investigated two estimates of risk difference for a binary outcome: one based on analysing outcomes as if from a conventional trial (the trial estimate), and one which combined estimates of different parameters in the manner of a decision analysis (the decision analysis estimate). From theory we derived the bias and standard error of each. We also considered the impact of randomising participants to a testing strategy before the tests are administered rather than after.

Results

The trial estimate and decision analysis estimate are both unbiased. Using the decision analysis estimate (but not the trial estimate) the same precision is achieved by randomising before testing as by randomising after. Giving both tests to all participants means fewer need to be recruited: in one example from the literature the proposed design was nearly four times more efficient in this sense than a conventional trial design.

Conclusions

We suggest the term “randomised discordance trial” for the design we have described. A discordance trial offers an efficient way to compare two diagnostic tests or biomarkers. Randomising before testing avoids selection bias. We have derived formulae for calculating sample size and for estimating risk difference and standard error using this design.

Author information

Correspondence to Richard Hooper.

Rights and permissions

This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Reprints and Permissions

About this article

Keywords

  • Standard Error
  • Diagnostic Test
  • Randomise Trial
  • Decision Analysis
  • Treatment Choice

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate. Please note that comments may be removed without notice if they are flagged by another user or do not comply with our community guidelines.