TB Manual

• 1 or more published systematic review (SR) of randomized controlled trials (RCTs) that concluded the evidence was of high quality; or
• 2 or more RCTs of high quality with consistent results; or
• 1 large RCT of high quality (low risk of bias) and clear findings (of benefits and/or harms).

Diagnostic methods:

• 1 or more RCT of high quality comparing alternative diagnostic strategies for effects on patient-relevant outcomes; or
• 1 or more published SR of two or more observational studies of high quality comparing a diagnostic test to an appropriate reference standard; or
• 2 or more observational studies comparing a diagnostic test to an appropriate reference standard of high quality with consistent findings.

Etiologic research:

• 1 or more published SR concluded that the association was clear and based on good quality observational studies; or
•  2 or more observational studies with strong designs demonstrating the risk factor was strongly associated, and that association was consistent across the studies

• 1 or more published SR that concluded that the evidence from intervention studies was not of high quality; or
• Any/all available RCT(s), were not of high quality, had inadequate numbers of participants, did not directly address the patients of interest; and/or demonstrated inconsistent results (if 2 or more RT Cs were available); or
• Only evidence form observational studies was available that had weak designs, weak effect estimates or inconsistent results, or generalization from a randomized trial that involved one type of patients to a different group of patients; or
• Opinion of the authors and other experts only.

Diagnostic:

• 1 or more published SR that concluded that the evidence was not of high quality; or
• 2 or more observational studies comparing a diagnostic test to an appropriate reference standard available but not high quality and/or had inconsistent results; or
• Only 1 observational study comparing a diagnostic test to a reference standard available.

Etiologic:

• 1 or more published SR concluded that the association was not clear, and/or based on poor quality observational
  studies; or
• 1 or more observational study of poor quality, and/or with inconsistent results.