Table 4.
Post-imputation WECare Medication intervention treatment effects by ignorability/uncertainty scenario. One-hundred models with 2 imputations per modelwere used to generate 200 imputations. Multipliers were generated by drawing from a Normal distribution. MAR, Weak NMAR, Strong NMAR, and Misspecified NMAR correspond to Normal distributions with means of 1, 1.3, 1.7, and 0.8, respectively. Amounts of uncertainty None, Mild, Moderate, Ample correspond to Normal distributions with standard deviations of 0, 0.1, 0.3, and 0.5, respectively.
Ignore Assump. | Uncertainty | Est. | SE | LCI | UCI | p-val | γ̂ | γ̂w | γ̂b |
|
|
---|---|---|---|---|---|---|---|---|---|---|---|
MAR | None | −0.69 | 0.25 | −1.18 | −0.19 | <0.01 | 0.34 | 0.34 | 0.00 | 0.00 | |
Mild | −0.69 | 0.27 | −1.22 | −0.17 | 0.01 | 0.42 | 0.35 | 0.06 | 0.15 | ||
Moderate | −0.70 | 0.38 | −1.46 | 0.05 | 0.07 | 0.69 | 0.34 | 0.35 | 0.51 | ||
Ample | −0.71 | 0.52 | −1.73 | 0.31 | 0.17 | 0.81 | 0.31 | 0.49 | 0.61 | ||
Weak | None | −0.70 | 0.30 | −1.29 | −0.11 | 0.02 | 0.37 | 0.37 | 0.00 | 0.00 | |
NMAR | Mild | −0.71 | 0.31 | −1.31 | −0.10 | 0.02 | 0.41 | 0.38 | 0.02 | 0.05 | |
Moderate | −0.71 | 0.39 | −1.48 | 0.05 | 0.07 | 0.62 | 0.37 | 0.25 | 0.40 | ||
Ample | −0.72 | 0.51 | −1.72 | 0.29 | 0.16 | 0.77 | 0.35 | 0.41 | 0.54 | ||
Strong | None | −0.70 | 0.35 | −1.39 | −0.00 | 0.05 | 0.36 | 0.36 | 0.00 | 0.00 | |
NMAR | Mild | −0.70 | 0.35 | −1.39 | −0.01 | 0.05 | 0.37 | 0.37 | 0.00 | 0.00 | |
Moderate | −0.71 | 0.40 | −1.49 | 0.07 | 0.07 | 0.50 | 0.38 | 0.12 | 0.24 | ||
Ample | −0.71 | 0.48 | −1.66 | 0.23 | 0.14 | 0.66 | 0.37 | 0.29 | 0.44 | ||
Misspec. | None | −0.67 | 0.22 | −1.12 | −0.23 | <0.01 | 0.27 | 0.27 | 0.00 | 0.00 | |
NMAR | Mild | −0.68 | 0.25 | −1.16 | −0.20 | <0.01 | 0.38 | 0.29 | 0.10 | 0.26 | |
Moderate | −0.69 | 0.38 | −1.43 | 0.05 | 0.07 | 0.71 | 0.29 | 0.42 | 0.59 | ||
Ample | −0.70 | 0.52 | −1.72 | 0.32 | 0.18 | 0.82 | 0.27 | 0.55 | 0.67 |
SE: standard error; LCI: lower 95% confidence interval; UCI: upper 95% confidence interval