Table 3.
Physical, emotional, and cognitive factors as mediators of the associations between diabetes and physical activity levels and evolution across aging (Age and Age2)
| Variables | b (95% CI) | p | % explained | ||
|---|---|---|---|---|---|
| Model without mediators | |||||
| Diabetes | Level | −.13 (−.15; −.11) | <.001 | ||
| Age (10-year follow-up) | −.04 (−.06; −.02) | <.001 | |||
| Age2 (10-year follow-up) | −.01 (−.02; .01) | .222 | |||
| Model including all the mediators | |||||
| Diabetes | Level | −.06 (−.08; −.05) | <.001 | 53.1% | |
| Age (10-year follow-up) | −.01 (−.02; .02) | .797 | 93.9% | ||
| Age2 (10-year follow-up) | .01 (−.01; .02) | .429 | All | ||
| Models including one mediator | |||||
| Depressive symptoms | Diabetes | Level | −.11 (−.12; −.09) | <.001 | 19.4% |
| Age (10-year follow-up) | −.03 (−.05; −.01) | .009 | 35.0% | ||
| Age2 (10-year follow-up) | −.01 (−.02; .01) | .571 | 54.2% | ||
| Depressive symptoms | Level | −.18 (−.18; −.17) | <.001 | ||
| Age (10-year follow-up) | −.06 (−.07; −.06) | <.001 | |||
| Age2 (10-year follow-up) | −.01 (−.01; −.01) | <.001 | |||
| Subjective energy | Diabetes | Level | −.12 (−.14; −.10) | <.001 | 9.6% |
| Age (10-year follow-up) | −.04 (−.06; −.02) | .001 | 13.2% | ||
| Age2 (10-year follow-up) | −.01 (−.02; .01) | .322 | 19.3% | ||
| Subjective energy | Level | −.22 (−.23; −.21) | <.001 | ||
| Age (10-year follow-up) | −.07 (−.08; −.05) | <.001 | |||
| Age2 (10-year follow-up) | −.01 (−.02; −.00) | .025 | |||
| Grip strength | Diabetes | Level | −.12 (−.14; −.10) | <.001 | 10.5% |
| Age (10-year follow-up) | −.04 (−.06; −.01) | .001 | 17.2% | ||
| Age2 (10-year follow-up) | −.01 (−.02; .01) | .458 | 39.6% | ||
| Grip strength | Level | −.32 (−.33; −.30) | <.001 | ||
| Age (10-year follow-up) | −.03 (−.04; −.01) | <.001 | |||
| Age2 (10-year follow-up) | −.03 (−.04; −.02) | <.001 | |||
| ADL | Diabetes | Level | −.10 (−.13; −.09) | <.001 | 16.9% |
| Age (10-year follow-up) | −.02 (−.04; −.00) | .035 | 48.4% | ||
| Age2 (10-year follow-up) | .00 (−.01; .01) | .952 | 95.2% | ||
| ADL | Level | −.58 (−.59; −.56) | <.001 | ||
| Age (10-year follow-up) | −.08 (−.10; −.06) | <.001 | |||
| Age2 (10-year follow-up) | .00 (−.01; .01) | .913 | |||
| IADL | Diabetes | Level | −.10 (−.11; −.08) | <.001 | 27.3% |
| Age (10-year follow-up) | −.02 (−.04; .00) | .082 | 58.1% | ||
| Age2 (10-year follow-up) | .00 (−.01; .01) | .941 | All | ||
| IADL | Level | −.58 (−.60; −.57) | <.001 | ||
| Age (10-year follow-up) | −.12 (−.13; −.10) | <.001 | |||
| Age2 (10-year follow-up) | −.02 (−.03; −.01) | .001 | |||
| Sleep | Diabetes | Level | −.13 (−.15; −.12) | <.001 | 2.6% |
| Age (10-year follow-up) | −.04 (−.06; −.02) | <.001 | 5.8% | ||
| Age2 (10-year follow-up) | −.01 (−.02; .01) | .294 | 14.2% | ||
| Sleep | Level | −.10 (−.11; −.09) | <.001 | ||
| Age (10-year follow-up) | −.04 (−.06; −.02) | <.001 | |||
| Age2 (10-year follow-up) | −.01 (−.02; −.00) | .020 | |||
| Verbal fluency | Diabetes | Level | −.12 (−.14; −.11) | <.001 | 7.0% |
| Age (10-year follow-up) | −.03 (−.05; −.01) | .002 | 21.6% | ||
| Age2 (10-year follow-up) | −.01 (−.02; .01) | .537 | 49.7% | ||
| Verbal Fluency | Level | .14 (.13; .15) | <.001 | ||
| Age (10-year follow-up) | .07 (.06; .07) | <.001 | |||
| Age2 (10-year follow-up) | .02 (.01; .02) | <.001 | |||
| Delayed recall | Diabetes | Level | −.13 (−.14; −.11) | <.001 | 3.8% |
| Age (10-year follow-up) | −.04 (−.06; −.02) | <.001 | 10.8% | ||
| Age2 (10-year follow-up) | −.01 (−.02; .00) | .325 | 19.5% | ||
| Delayed recall | Level | .07 (.06; .07) | <.001 | ||
| Age (10-year follow-up) | .04 (.03; .05) | <.001 | |||
| Age2 (10-year follow-up) | .01 (.01; .02) | <.001 |
Note. All models were adjusted for gender (male, female), body mass index (BMI), birth cohort, country of residence, education, household’s ability to make ends meet, and participants’ attrition. “Age (10-year follow-up)” and “Age2 (10-year follow-up)” estimated the effect of diabetes in the linear and quadratic evolution in the engagement in physical activity over a 10-year period. The percentage of the effect of diabetes on physical activity levels and evolution across aging explained by the variables were estimated based on the exacts estimates, not on the rounded values presented in the table. The model including all the mediators was meant to test the overall percentage of the association between diabetes and physical activity that was explained by all the potential mediators. The models including only one mediator at a time were meant to test the percentage of the association between diabetes and physical activity that was explained by this mediator of interest.