Table 1 Detailed description of inclusion and exclusion criteria

Methods

Systematic review of interventions/exposures including RCTs and cohort studies; equivalent search for RCTs and cohort studies; performing quantitative meta-analysis for at least one BoE.

Umbrella reviews, narrative reviews, systematic reviews of diagnostic test accuracy, Individual patient data meta-analysis; no quantitative meta-analysis

BoE-pairs

BoE-pair with a BoE from RCTs and a BoE from cohort studies evaluating the same medical research question (e.g. association of Exenatide with pancreatitis; effect of Vitamin D on hypertension; comparing total- with unicompartimental knee arthroplasty for range of movement of the knee)

Single small study (n < 1.000 participants) for one BoE (RCT or cohort studies); BoE-pair with one BoE using a continuous outcome and the other BoE using a binary outcome (e.g. risk of hypertension vs. mean difference of systolic blood pressure)

Population

All populations (e.g. primary prevention, secondary prevention, general population, adults, children)

-

Intervention/Exposure

All types of medical interventions and exposures (e.g. drugs, invasive procedures, nutrients, vaccines)

-

Comparator

All types of comparators (e.g. placebo, drugs, invasive procedures, nutrients, vaccines)

-

Outcomes

Patient-relevant outcomes (e.g. mortality, cancer outcomes, cardiovascular outcomes, obstetrical outcomes) and of intermediate disease markers (e.g. LDL-cholesterol)

-

Study design

Randomized controlled trials (e.g. parallel, cluster, factorial, cross-over); cohort studies (e.g. prospective cohort, retrospective cohort, observational cohort analysis of RCT)

Quasi-RCTs, non-randomized controlled trials, case-control studies, cross-sectional studies, ecological studies