Table 2.

Test accuracy studies of noninvasive imaging modalities

Outcome	No. of Studies (No. of Patients)	Study Design	Factors that may Decrease Certainty of Evidence					Effect per 1000 Patients Tested		Test Accuracy CoE
			Risk of bias	Indirectness	Inconsistency	Imprecision	Publication bias	Pretest probability of 40%	Pretest probability of 50%
True‐positives	8 (180)	Cohort & case‐control type studies	Serious a	Serious b	Not serious c	Serious d	None	286 (218‐337)	358 (272‐421)	⨁◯◯◯ VERY LOW
False‐negatives								114 (63‐182)	142 (79‐228)
True‐negatives	8 (193)	Cohort & case‐control type studies	Serious a	Serious b	Not serious c	Serious d	None	416 (320‐490)	347 (267‐408)	⨁◯◯◯ VERY LOW
False‐positives								184 (110‐280)	153 (92‐233)

Sensitivity was 0.72 (95% CI: 0.54‐0.84) and specificity was 0.69 (95% CI: 0.53‐0.82). Prevalences were 40% and 50%. True‐positives are patients with disease activity; false‐negatives are patients incorrectly classified as not having disease activity; true‐negatives are patients without disease activity; and false‐negatives are patients incorrectly classified as having disease activity. See Webb et al, Grayson et al, Eshet et al, Quinn et al, John et al, Santhosh et al, Karapolat et al, Nguyen et al, for reference (25, 26, 27, 28, 29, 36, 37, 38).

Abbreviations: CI, confidence interval; COE, certainty of evidence.

^aDue to patient selection (some studies did not avoid inappropriate exclusions), the results of the index test were interpreted with knowledge of the results of the standard reference, and not all patients received a reference test.

^bIndirectly compares the interventions in which we are interested (invasive vs noninvasive) when applied to the populations in which we are interested.

^cThe similarity of point estimates and overlap of CIs make inconsistency not serious.

^dClinical action would differ if the upper versus the lower boundary of the CI represented the truth.