TABLE 4.
E-values for effect measures and for CI limits for the associations between social capital and subsequent mental health (Well-Being Survey 2018–2019 and health insurance claims data 2017–2019, United States).
| Social connectedness | Depression | Anxiety | ||
|---|---|---|---|---|
| Effect estimate a , b | CI limit c | Effect estimate b | CI limit c | |
| Social connectedness | 2.08 | 1.68 | 1.72 | 1.31 |
| Items of social connectedness | ||||
| My relationships are as satisfying as I would want them to be | 1.69 | 1.27 | 1.51 | 1.00 |
| There are people who really understand me | 1.71 | 1.36 | 1.53 | 1.03 |
| How often do you feel lonely? | 1.96 | 1.60 | 1.71 | 1.37 |
| I am content with my friendships and relationships | 1.83 | 1.45 | 1.64 | 1.24 |
| I have enough people I feel comfortable asking for help at any time | 1.74 | 1.41 | 1.67 | 1.29 |
| I feel connected to the broader community around me | 1.78 | 1.40 | 1.35 | 1.00 |
| People in my broader community trust and respect one another | 1.91 | 1.52 | 1.45 | 1.00 |
See VanderWeele and Ding [42] for the formula for calculating E-values.
The E-values for effect estimates are the minimum strength of association on the risk ratio scale that an unmeasured confounder would need to have with both the exposure and the outcome to fully explain away the observed association between the exposure and outcome, conditional on the measured covariates. For example, in the studied population an unmeasured confounder would need to be associated with both feeling lonely and depression by risk ratios of 1.96 each, above and beyond the measured covariates, to fully explain away the observed association between the two variables.
The E-values for the limit of the 95% confidence interval (CI) closest to the null denote the minimum strength of association on the risk ratio scale that an unmeasured confounder would need to have with both the exposure and the outcome to shift the confidence interval to include the null value, conditional on the measured covariates. For example, in the studied population an unmeasured confounder would need to be associated with both feeling lonely and depression by 1.60-fold each, above and beyond the measured covariates, to shift the upper limit of the confidence interval to include the null value for the association between the two variables.