Table 1.
Assessment scale. Adapted from Dufault and Klar [29]
| Item | Description | |
|---|---|---|
| Study design and focus (max = 4) | Sample size (max = 2) | Number of ecologic units included in the analysis as a proportion of the total number of units (3 levels: < 11% = 0 points; 11-79% = 1 point; > 79% = 2 points) |
| Level of inference (max = 1) | The results of the analysis are not used to draw inferences for individuals | |
| Pre-specification of ecological units (max = 1) | Ecological units are selected to suit the hypothesis (as opposed to seemingly motivated by convenience or necessity such as the use of districts, towns or counties) | |
| Statistical methodology (max = 5) | Validity of statistical inferences (max = 2) | Number of ecological units per covariate (3 levels: 0-10 = 0 points; 10-20 = 1 point; > 20 = 2 points) |
| Use of covariates (max = 1) | Analysis adjusted for covariates (e.g. sociodemographic; environmental risk factors) | |
| Proper adjustment for covariates (max = 1) | Covariates are properly adjusted when regressed upon adjusted outcomes as recommended for ecological studies [30] | |
| Spatial effects (max = 1) | Inclusion of spatial analysis | |
| Quality of reporting (max = 3) | Statement of study design (max = 1) | Key elements of the study design are presented in the report |
| Justification of study design (max = 1) | Justification of the ecological analysis, the rationale and the objectives are presented in the report | |
| Discussion of cross-level bias and limitations (max = 1) | Readers are cautioned about the limitations of the ecological design and/or the ecological fallacy |