Decision making about healthcare-related tests and diagnostic strategies: A framework provides guidance for when research providing test-accuracy information is sufficient

Mustafa R1, Brozek J2, Santesso N2, Akl E3, Falck-Ytter, Y4, Wiercioch W2, Bossuyt P5, Garg A6, Lelgemann M7, Bühler D8, Schünemann H2
1University of Kansas Medical Center, 2McMaster University, 3American University of Beirut , 4Case Western Reserve University and Louis Stokes VA Medical Center, 5Academic Medical Center, Amsterdam, 6Western University, 7Medizinischer Dienst des Spitzenverbandes Bund der Kranken-kassen e.V. (MDS) , 8Abteilung Medizin. GKV

Background: Decisions about tests and diagnostic strategies are more challenging to make than therapy ones primarily due to the paucity of direct evidence about patient important benefits and harms of using the test. To make up for this deficit, decision makers often have to link different pieces of evidence and make assumptions about the consequences of using tests on patients and populations.

Objective: We provide overarching principles and a conceptual framework for unique considerations when making decisions about tests and when test accuracy results may be sufficient to extrapolate about benefits and harms.

Methods: Using an iterative approach the investigators arrived at overarching principles for making decisions about tests. We applied the framework to examples to illustrate the guidance.

Results: We summarise the overarching principles that guide making decisions about a test. We also present a new framework that guide coverage decision making. In this framework we explain when test accuracy may be sufficient to extrapolate about healthcare benefits and harms which include when there is high or moderate certainty about non-inferiority or superiority of test accuracy for a 'new' test that is intended to replace a test with established link to improving patient outcomes. It also includes when there is high or moderate certainty of evidence that the test accuracy of one test is equivalent or better than the combined accuracy of two tests (one of which is the test evaluated separately).

Conclusion: The framework presented in this article should be used when making coverage decisions about tests. We believe this guidance will help determine when decision makers need to fully model the results of tests and their effect on patient important outcomes to assess benefits and harms and when they can make decisions without the need for that step. We also believe the framework guides decisions about covering the test for evidence synthesis purposes.