Table 3.
Assessments of Missing Data Patterns: Logistic Regression with Outcome being Attrition, with and without Including Demographic Variables.
| Attrition in Immediate Recall Scores as Outcome | |||||||
|---|---|---|---|---|---|---|---|
| Model 1 | Model 2 | ||||||
| Estimation Odds Ratio | 95% CI | p-value | Estimation Odds Ratio | 95% CI | p-value | ||
| Subject Specific Slopes (random effects of age-associated slope obtained from mixed effects model) | 0.03 | 0.02 – 0.05 | <.0001 | 0.36 | 0.12 – 1.06 | 0.06 | |
| Subject Specific Baseline (random effects of intercept obtained from mixed effects model) | -- | -- | 0.94 | 0.90 – 0.98 | 0.01 | ||
| baseline age | -- | -- | 1.10 | 1.07 – 1.14 | <.0001 | ||
| female | -- | -- | 0.64 | 0.55 – 0.75 | <.0001 | ||
| Education: high school (vs. < high school) | -- | -- | 0.91 | 0.75 – 1.11 | 0.64 | ||
| Education : > high school (vs. < high school) | -- | -- | 0.76 | 0.62 – 0.94 | 0.01 | ||
| Cohort 1912–1921 | -- | -- | 0.57 | 0.40 – 0.82 | 0.04 | ||
| Cohort 1922–1931 | -- | -- | 0.48 | 0.28 – 0.82 | <.0001 | ||
| Cohort 1932–1943 | -- | -- | 0.79 | 0.35 – 1.80 | 0.50 | ||
| Study (MoVIES) | -- | -- | 7.59 | 4.52 – 12.77 | <.0001 | ||
| Attrition in Delayed Recall Scores as Outcomes | |||||||
| Model 1 | Model 2 | ||||||
| Estimation Odds Ratio | 95% CI | p-value | Estimation Odds Ratio | 95% CI | p-value | ||
| Individual Slopes (random effects of age-associated slope obtained from mixed effects model) | 0.01 | 0.01 – 0.03 | <.0001 | 0.50 | 0.11 – 2.26 | 0.37 | |
| Individual Baseline (random effects of intercept obtained from mixed effects model) | -- | -- | 0.85 | 0.80 – 0.91 | <.0001 | ||
| baseline age | -- | -- | 1.11 | 1.08 – 1.14 | <.0001 | ||
| female | -- | -- | 0.63 | 0.53 – 0.73 | <.0001 | ||
| Education: high school (vs. < high school) | -- | -- | 0.89 | 0.73 – 1.08 | 0.71 | ||
| Education : > high school (vs. < high school) | -- | -- | 0.75 | 0.60 – 0.92 | 0.008 | ||
| Cohort 1912–1921 | -- | -- | 0.57 | 0.40 – 0.82 | 0.06 | ||
| Cohort 1922–1931 | -- | -- | 0.47 | 0.27 – 0.79 | <.0001 | ||
| Cohort 1932–1943 | -- | -- | 0.79 | 0.35 – 1.78 | 0.49 | ||
| Study (MoVIES) | -- | -- | 9.70 | 5.74 – 16.40 | <.0001 | ||
In Model 1, less age associated declines (i.e., larger random effects in slopes obtained from the mixed effects models) are significantly protective against attrition in both tests. However, individual specific age associated trajectory was not associated with attrition once we controlled for demographic variables. These results confirm that there is no evidence to reject our MAR assumption and mixed effects models can be applied to assess age associated cognitive trajectories. (Please see Appendix for further explanation on this issue).