Table 1 Bias assessment tool

Confounding bias

Did the study analysis adjust potential confounders appropriately?

Confounders adjusted

1: Many confounders and unobserved confounding

2: Some confounders (in particular mobility or socioeconomic variables) or unobserved confounders

3: None or cannot tell

Control of the population

Did the study control for population?

Population, age structure of the population

1: Control and/or including population density

2: Control only by including population density

3: No control

Control of the spatial and/or temporal dependence

Did the study control the spatial and/or temporal extra variability?

Spatial and temporal dependence

1: Control

2: Partial control (control of only one dependence, for example)

3: No control

Control of non-linearities

Did the study control for non-linearities?

Non-linearities (parametric or non-parametric)

1: Control

3: No control

Measurement errors

What was the heterogeneity of indicators used in the study?

Measurement errors in the explanatory variables (exposure variables and covariates)

1: Control

2: Partial control (including lags, for example)

3: No control

Statistical model

Did the study use appropriate statistical model?

Statistical model

1: Models for count data response variables

2: Control of heteroscedasticity and rates as response variables

3: Models with normally distributed errors and count data response variable; or no control of heteroscedasticity and rates as response variables

Overall Study Rating

Strong (low risk of bias): one domain, at most, was rated as 3

Moderate: up to 2 domains were rated as 3

Weak (high risk of bias): ≥ 3 domains were rated as 3