Table 5.

Mixed effects models testing the incremental influence of individual indices of active involvement (over and above 12-step attendance) on PDA over time

Model	Concurrent (n=118)			Lagged (n=113)
	est	se	t	est	se	t
Contact with sponsor	0.61	0.30	2.01*	−0.45	0.49	−0.93
Contact with other members	0.14	0.23	0.63	0.41	0.26	1.58
Read 12-step literature	−0.07	0.40	−0.18	0.03	0.47	0.07
Talked or shared in meetings	0.11	0.06	1.93†	−0.12	0.09	−1.41
Helped set up or run meetings	0.11	0.12	0.87	0.37	0.15	2.38*
Steps completed	−2.75	1.74	−1.58	−2.20	2.35	−0.93

^†p<.1;

^*p<.05; Abbreviations: PDA = percent days abstinent

¹Models were run separately for each index of 12-step involvement. Estimates adjusted for age, race, baseline self-efficacy and abstinence goal, and time-varying inpatient/outpatient SUD treatment and 12-step meeting attendance (coefficients not shown due to space). In the concurrent models, 12-step attendance and involvement indices, and time-varying covariates are concurrent with PDA. In the lagged models, 12-step attendance and involvement indices, and time-varying covariates are lagged one time period behind PDA. 12-step meeting attendance was significantly associated with PDA in all concurrent models (ps<.001), but not in the lagged models.