Television-viewing time and bodily pain in Australian adults with and without type 2 diabetes: 12-year prospective relationships

Table 3 Unconditional and conditional linear growth curve models for bodily pain

	Unconditional model	Conditional models
	Model 1: (Function of age)	Model 2: Model 1 + TV time	Model 3: Fully adjusted linear-additive model	Model 4: Model 2 + TV time # T2D status	Model 5: Fully adjusted with TV time#T2D status
	Coefficient (S.E)	Coefficient (S.E)	Coefficient (S.E)	Coefficient (S.E)	Coefficient (S.E)
Fixed effect
Intercept	75.55 (0.45)***	77.53 (0.52)***	76.92 (2.20)***	77.83 (0.54)***	76.71 (2.20)***
Slopes
Age (Centred at 50 years)	- 0.28 (0.02)***	- 0.24 (0.02)***	- 0.30 (0.03)***	- 0.21 (0.03)***	- 0.30 (0.03)***
TV time		- 1.15 (0.17)***	- 0.69 (0.17)***	- 1.03 (0.19)***	- 0.56 (0.19)**
T2D status
NGM (Reference)				0	0
Prediabetes				- 0.97 (0.96)	0.91 (0.96)
T2D				- 4.07 (1.50)**	0.53 (1.48)
TV time#T2D status
NGM (Reference)				0	0
Pre-diabetes				- 0.10 (0.40)	- 0.22 (0.40)
T2D				- 0.63 (0.56)	- 0.97 (0.55)^$
Random effect
Intercept variance
Cluster^a	4.85 (1.92)**	4.09 (1.75)**	1.05 (0.87)	3.49 (1.59)**	1.03 (0.87)
Participants	192.89 (7.48)***	191.05 (7.46)***	145.05 (6.05)***	188.80 (7.42)***	145.05 (6.05)***
Slope variance	0.014 (0.006)**	0.015 (0.006)**	0.007 (0.001)**	0.015 (0.006)**	0.008 (0.001)**
Intercept-Slope covariance	1.65 (0.34)**	1.69 (0.34)**	1.04 (0.06)***	1.66 (0.34)**	1.04 (0.06)***
Within-individual variance	266.03 (4.58)***	265.65 (4.61)***	263.06 (4.75)***	265.56 (4.61)***	262.94 (4.74)***
Goodness-of-fit
AIC	95971.21	95117.23	90691.72	95024.25	90692.51
BIC	96022.25	95175.50	90858.33	95111.64	90873.61
Log-likelihood	- 47978.61	- 47550.62	- 45322.86	- 47500.12	- 45321.26
No of parameters	7	8	23	12	25

Statistically significant: *** p < 0.001, ** p < 0.01, * p < 0.05, ^$ p = 0.076
TV time#T2D Status Interaction between TV time and T2D status, TV Television-viewing, S.E Standard error, NGM Normal glucose metabolism, T2D Type 2 diabetes (included newly diagnosed and known T2D)
The fully adjusted linear additive model 3 included model 2 + sex, education level, household income, smoking status, leisure-time physical activity, waist circumference, energy intake, T2D status, SF36 mental component score, presence of chronic kidney disease, history of cardiovascular disease (CVD), and cancer.
The fully adjusted model 5 with TV time#T2D status included model 4 + sex, education level, household income, smoking status, leisure-time physical activity, waist circumference, energy intake, SF36 mental component score, presence of chronic kidney disease, history of cardiovascular disease (CVD), and cancer.
^a This represents the intercept variance that is attributable to the level 3 clustering of individuals (individuals nested in clusters); thus, describes the variance component of cluster-to-cluster variability.

ISSN: 1471-2458