Table 3.
Multilevel Models of Erectile Function and Sexual Satisfaction Predicting Cognitive Performance at Baseline and Across Time
| Variable | Episodic memory | Executive function | Processing speed | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| β (95% CI) | p | σ 2 | SE | β (95% CI) | p | σ 2 | SE | β (95% CI) | p | σ 2 | SE | |
| Fixed effects | ||||||||||||
| Time-varying Age (centered at 56) | −0.19 (−0.17, −0.55) | <.001 | −0.35 (−0.39, −0.32) | <.001 | −0.42 (−0.44, −0.39) | <.001 | ||||||
| Model 3: erectile function | ||||||||||||
| 3a Baseline erectile function (γ01) | 0.11 (0.05, 0.16) | <.001 | 0.08 (0.03, 0.14) | .003 | 0.11 (0.06, 0.17) | <.001 | ||||||
| 3a Change in erectile function (β2) | 0.08 (0.04, 0.13) | <.001 | 0.03 (−0.01, 0.08) | .164 | 0.04 (−0.01, 0.08) | .088 | ||||||
| 3b Baseline erectile function (γ01) | 0.11 (0.05, 0.16) | <.001 | 0.07 (0.02, 0.13) | .014 | 0.10 (0.04, 0.15) | .003 | ||||||
| 3b Change in erectile function (β2) | 0.09 (0.04, 0.13) | <.001 | 0.02 (−0.03, 0.06) | .435 | 0.03 (−0.02, 0.07) | .248 | ||||||
| Model 4: sexual satisfaction | ||||||||||||
| 4a Baseline sexual satisfaction (γ01) | 0.06 (−0.0003, 0.11) | .051 | 0.03 (−0.02, 0.09) | .224 | 0.04 (−0.01, 0.09) | .185 | ||||||
| 4a Change in sexual satisfaction (β2) | 0.07 (0.02, 0.11) | .005 | 0.03 (−0.02, 0.07) | .228 | 0.04 (.001, 0.08) | .049 | ||||||
| 4b Baseline sexual satisfaction (γ01) | 0.04 (−0.01, 0.10) | .127 | 0.01 (−0.04, 0.07) | .640 | 0.03 (−0.03, 0.08) | .360 | ||||||
| 4b Change in sexual satisfaction (β2) | 0.06 (0.02, 0.11) | .004 | 0.01 (−0.04, 0.05) | .717 | 0.03 (−0.01, 0.07) | .182 | ||||||
| Random effects | ||||||||||||
| Residual | 0.19 | 0.01 | 0.21 | 0.01 | 0.16 | 0.01 | ||||||
| Intercept | 0.45 | 0.03 | 0.44 | 0.03 | 0.48 | 0.03 | ||||||
| Time-varying age | 0.03 | 0.009 | 0.01 | 0.01 | 0.004 | 0.007 |
Notes: Each model represents an age-based multilevel model predicting the outcome shown in each column header. Model 3a/Model 4a was adjusted for age centered at age 56, race, income, education in years, committed status, and sexual activity. Model 3b/Model 4b was adjusted for depressive symptoms, physical health, number of cardiovascular diseases (CVDs), number of non-CVD chronic health conditions, alcohol consumption, current smoking status, and erectile dysfunction medication use. Model 3 only included erectile function predictors and Model 4 only included sexual satisfaction predictors. Changes in erectile function and sexual satisfaction were calculated by subtracting follow-up values from an individual’s baseline values. Baseline erectile function/sexual satisfaction predict baseline intercept (β0it). Change in those measures predicted change in cognitive performance (Yijt). CI = confidence interval; SE = standard error.