Table 4. Bayesian network propagation: Predicting change in outcomes.

Targeted behavior	Change in targeted behavior(s)	Outcome	Average change in Outcome [Mean(SD) of outcome by targeted behavior category]
PA	< 270 to ≥ 380 counts/min/day	BMI	BMI decreases 1 kg/m2 [31.5(4.9) to 30.5(4.9)]
BMI	≥ 30 to < 30 kg/m2	Insulin	Insulin decreases 0.30 (log) pg/mL (26% decrease) [6.29(0.50) to 5.99(0.49)]
BMI	≥ 30 to < 30 kg/m2	CRP	CRP decreases 0.76 (log) mg/L (50% decrease) [15.28(1.07) to 14.56(1.05)]
BMI	≥ 30 to < 30 kg/m2	QoLp	QoLp increases 6.1 [64.6(17.9) to 70.7(17.7)]
BMI	≥ 30 to < 30 kg/m2	QoLm	QoLm increases 0.3 [74.9(17.3) to 75.2(16,9)]
Sleep2	> Q3 to < Q1	QoLp	QoLp increases 20.5 [56.6(16.7) to 77.1(16.3)]
Sleep2	> Q3 to < Q1	QoLm	QoLm increases 23.9 [62.9(15.9) to 86.8(14.7)]
Sleep2 + BMI	sleep2 > Q3 to < Q1 BMI ≥ 30 to < 30 kg/m2	QoLp	QoLp increases 26.7 [54.3(16.1) to 81.0(16.1)\
Sleep2 + BMI	sleep2 > Q3 to < Q1 BMI ≥ 30 to < 30 kg/m2	QoLm	QoLm increases 23.3 [62.8(15.8) to 86.1(14.5)]

^a Derived from 5000 simulated datasets via network propogation.

Q3: 75^th%-ile; Q1: 25^th%-ile