Table 2. Descriptions of design options to study the impact of TB diagnostics on patient-important outcomes with references on methodological issues.

Design category (and methodological references)	Design sub-type (and methodological references)	Description	References of studies included in the review
Randomized Controlled Trials (RCTs) [25,64]	Individual RCT [65]	Individuals are randomized to either receive or not receive the intervention	[55,66]
	Cluster RCT [25,67–70]	Parallel group: Clusters are randomized to either receive or not receive the intervention throughout the entire study period	[39,40]
		Stepped-wedge: The sequence in which clusters receive the intervention is randomized such that all clusters receive it by the end of the study (also called phased-implementation trial)	[60]
Quasi-Experimental studies [23,27,71–73]	Pre/post implementation [74–76]	Uncontrolled: Two cohorts are compared between two different time periods: the pre cohort receives the usual care during the baseline period and the post cohort receives the intervention during a subsequent and distinct time period.	[41–45,51,56,57,59]
		Controlled: Like ‘Uncontrolled pre/post implementation’ but including a contemporary control cohort that receives the usual care during both the pre and the post period, also called ‘difference in differences design’.	none
	Interrupted time-series [77–80]	Multiple measurements over time before and after implementation of the intervention analyzed using segmented regression or ARIMA models.	none
	Instrumental variable [81–85]	The effect of the intervention on the outcome is captured through another variable (the “instrument”) that affects the outcome only by affecting the intervention and does not share any causes with the outcome.	none
	Regression discontinuity [86–89]	The effect of the intervention on the outcome can be estimated if individuals receive the intervention based on whether they are above or below some threshold value on a continuous variable.	none
Single-cohort designs	Hypothetical studies [28]	A single cohort receives both baseline tests and the index test but results from the index tests are not used for patient management. “Hypothetical” changes in patient-important outcomes—had results been available to doctors—are estimated using a combination of study data, assumptions and potentially data from other studies.	[29,53,58,90]
	Observational studies	Inferences about the effect of the index test on patient-important outcomes are attempted based on a single cohort receiving both baseline tests and the index test with both being used for patient management.	[46–50,52,54]
	Diagnostic before/after studies [61,91–94]	Inferences about the effect of the index test on patient-important outcomes are attempted based on comparisons of the pre-test management plan (i.e. planned patient management before availability of index test results) and post-test management plan.	none