Table 5.
Primary and sensitivity analyses: correlation between MVPA and total direct costs
| Type of analysis | Total direct costs (yes) (f.s.) | Total direct costs (Amount) (s.s.) |
|---|---|---|
| Odds [95% CI] | Exp(Estimate) [95% CI] | |
| Primary analysis model, Ref: < 7 h/week MVPA | 1.00 | 1.00 |
| ≥7 h/week MVPA1 | 0.95 [0.76-1.19] | 0.99 [0.88-1.12] |
| SAM 1: Exclusion of high utilization, Ref: < 7 h/week MVPA | 1.00 | 1.00 |
| ≥7 h/week MVPA1 | 0.95 [0.76-1.19] | 1.05 [0.93-1.18] |
| SAM 2: four-level MVPA variable, Ref: < 3.5 h/week MVPA | 1.00 | 1.00 |
| ≥3.5 and < 7 h/week MVPA | 1.04 [0.68-1.59] | 1.12 [0.89-1.41] |
| ≥7 and < 10.5 h/week MVPA | 0.92 [0.61-1.40] | 0.90 [0.72-1.13] |
| ≥10.5 h/week MVPA | 1.01 [0.68-1.51] | 1.18 [0.95-1.47] |
| SAM 3: plus Interaction MVPA & gender, Ref: < 7 h/week MVPA | 1.00 | 1.00 |
| Effect of ≥ 7 h/week MVPA if gender female | 0.94 [0.69-1.28] | 1.31 [1.10-1.55]** |
| Effect of ≥ 7 h/week MVPA if gender male | 0.96 [0.69-1.32] | 0.77 [0.64-0.91]** |
| SAM 4: plus Interaction MVPA & BMI, Ref: < 7 h/week MVPA | 1.00 | 1.00 |
| Effect of ≥ 7 h/week MVPA if normal weight | 0.96 [0.76-1.23] | 0.99 [0.86-1.13] |
| Effect of ≥ 7 h/week MVPA if severely underweight | 1.21 [0.22-6.76] | 2.08 [0.98-4.45] |
| Effect of ≥ 7 h/week MVPA if underweight | 0.91 [0.36-2.29] | 2.09 [1.29-3.38]** |
| Effect of ≥ 7 h/week MVPA if overweight, not obese | 0.89 [0.31-2.54] | 1.09 [0.64-1.86] |
| Effect of ≥ 7 h/week MVPA if obese | 0.65 [0.06-6.78] | 1.65 [0.71-3.83] |
| Effect of ≥ 7 h/week MVPA if BMI information missing | 0.77 [0.26-2.29] | 0.19 [0.10-0.37]*** |
Abbreviations: BMI: body mass index CI: confidence interval; Exp(Estimate): exponential function of l’beta estimate for costs; f.s.: first step of the model (logistic regression); h: hours; MVPA: moderate to vigorous physical activity; Odds: odds ratio estimates for costs; PA: physical activity; %: percentage; Ref: reference category; SAM: sensitivity analysis model; s.s.: second step of the model (overall generalized linear mixed regression).
***/**/*, values nominally significant at the 0.1%/1%/5% level (without adjustment for multiple testing).
Model information:
1effect of respectively higher MVPA (≥7 h/week) for both sexes compared with reference category lower MVPA (<7 h/week).
all SAMs were calculated using covariates of the primary analysis; number of observations vary in data sets and between steps of the regression model:
f.s.: primary analysis, SAM2, SAM3, SAM4 (N = 3356), SAM1 (N = 3355).
s.s.: primary analysis, SAM2, SAM3, SAM4 (N = 2913), SAM1 (N = 2912).