Bias in trials comparing paired continuous tests can cause researchers to choose the wrong screening modality

Table 1 For x <θ, observed screening test results, and observed and true disease status.

Test 1 Score Inequality	Test 1 Result	Test 2 Score Inequality	Test 2 Result	Signs and Symptoms	Observed Disease Status	True Disease Status
T ₁ <x	-	T ₂ <x	-		-	-
T ₁ <x	-	T ₂ <x	-	No signs	-	+
T ₁ <x	-	T ₂ <x	-	Signs	+	+
T ₁ <x	-	x <T ₂ <θ	+		-	-
T ₁ <x	-	x <T ₂ <θ	+	No signs	-	+
T ₁ <x	-	x <T ₂ <θ	+	Signs	+	+
T ₁ <x	-	θ <T ₂	+		-	-
T ₁ <x	-	θ <T ₂	+		+	+
x <T ₁ <θ	+	T ₂ <x	-		-	-
x <T ₁ <θ	+	T ₂ <x	-	No signs	-	+
x <T ₁ <θ	+	T ₂ <x	-	Signs	+	+
x <T ₁ <θ	+	x <T ₂ <θ	+		-	-
x <T ₁ <θ	+	x <T ₂ <θ	+	No signs	-	+
x <T ₁ <θ	+	x <T ₂ <θ	+	Signs	+	+
x <T ₁ <θ	+	θ <T ₂	+		-	-
x <T ₁ <θ	+	θ <T ₂	+		+	+
θ <T ₁	+	T ₂ <x	-		-	-
θ <T ₁	+	T ₂ <x	-		+	+
θ <T ₁	+	x <T ₂ <θ	+		-	-
θ <T ₁	+	x <T ₂ <θ	+		+	+
θ <T ₁	+	θ <T ₂	+		-	-
θ <T ₁	+	θ <T ₂	+		+	+

The situations listed in this table are those where x <θ, i.e. that the test score is less than the threshold that leads to referral to the invasive, yet sensitive and specific secondary test. The signs and symptoms column is left empty in those cases where the observed disease status does not depend on the development of signs or symptoms.

ISSN: 1471-2288