Skip to main content

Advertisement

Springer Nature is making Coronavirus research free. View research | View latest news | Sign up for updates

Table 2 Study design and the potential level of evidence

From: Randomised and non-randomised studies to estimate the effect of community-level public health interventions: definitions and methodological considerations

Study type Designs with the potential to provide strong evidence Designs with the potential to provide reasonable evidence Designs likely to provide weak or misleading evidence
CRT CRT that is blinded to participants and investigators OR uses an objective outcome measure CRT that is not blinded to the participants AND uses a subjective outcome measure AND study participation is not obvious to participants (see text) CRT that is not blinded to the participants AND uses a subjective outcome measure AND study participation is obvious to participants (see text)
NCT NCT with very good balance of the outcome of interest at baseline across arms
NCT conducted as controlled interrupted time series analysis
NCT with imbalance of the outcome of interest at baseline across arms AND no major trend in the control arm NCT with imbalance of the outcome of interest at baseline across arms AND a major trend in the control arm
NCT without baseline measure of the outcome of interest
CBA   CBA with reasonable balance of the outcome of interest at baseline across arms AND no major trend in the control arm
CBA with only one cluster per arm AND multiple outcome measures before and after the intervention
CBA with imbalance of the outcome of interest at baseline across arms OR major trend in the control arm
CBA with only one cluster per arm AND single outcome measure before and after the intervention
BA   BA with multiple outcome measures before and after the intervention
BA with adopter/non-adopter comparison before and after the intervention
BA conducted as interrupted time series analysis
BA with a single outcome measure before and after the intervention
BA using post-intervention adopter/non-adopter comparison without baseline