Randomised control trial of the effectiveness of an integrated psychosocial health promotion intervention aimed at improving health and reducing substance use in established psychosis (IMPaCT)

Table 2 Results of the mixed effect linear model for main outcomes (physical and mental health score) with treatment arm at 12 months or 15 months

	SF-36 Physical Health Score				SF-36 Mental Health Score
Variable	Coefficient	(95% CI)	z	P	Coefficient	(95% C.I.)	z	p
Arm (HPI = 1)	−1.77	(−3.88 to 0.34)	−1.65	0.10	0.37	(−2.41 to 3.15)	0.26	0.79
Time (15 mo = 1)	−0.15	(−1.73 to 1.44)	−0.18	0.86	0.22	(−1.75 to 2.18)	0.22	0.83
Arm x Time	0.77	(−1.49 to 3.03)	0.67	0.51	−1.05	(−4.15 to 2.04)	−0.67	0.51
Borough	Chi2(9) = 8.37, p = 0.50				chi2(9) = 8.82, p = 0.45
Pairwise comparison	Coefficient	[95% CI]	z	P			z	P
12 mo: HPI vs TAU	−1.77	(−3.88 to 0.34)	−1.65	0.10	0.37	(−2.41 to 3.15)	0.26	0.79
15 mo: HPI vs TAU	−1.00	(−3.15 to 1.14)	−0.92	0.36	−0.68	(−3.56 to 2.19)	−0.47	0.64
Group HPI: 15 mo vs 12 mo	−0.15	(−1.73 to 1.44)	−0.18	0.86	0.22	(−1.75 to 2.18)	0.22	0.83
Group TAU: 15 mo vs 12 mo	0.62	(−1 to 2.25)	0.75	0.45	−0.84	(−3.24 to 1.56)	−0.68	0.49
Random effect care coordinator	1.06	(0.07 to 15.29)			2.17	(0.72 to 6.56)

Results of the mixed effect linear model for main outcomes (physical and mental health score) with treatment arm, time (12 months or 15 months), the interaction between treatment arm and time, borough and baseline values of outcome. Pairwise comparisons show treatment effect estimates at 12 and 15 months and changes from 12 to 15 months within each treatment arm. Care coordinator was included as a random factor to account for the dependency of observations with care coordinator. The dependency of repeated observations within individuals was modelled by estimating the variance–covariance structure of the residuals
TAU treatment as usual, HPI IMPACT therapy (Health Promotion Intervention)

ISSN: 1471-244X