Skip to main content
. 2020 Dec 15;81(6):790–797. doi: 10.15288/jsad.2020.81.790

Table 2.

Cross-lagged models

graphic file with name jsad.2020.81.790tbl2.jpg

Model Effect Estimate p [95% CI]
Alcohol (n = 1,239)
 Autoregressions
Depressiont → Depressiont+1 .27 <.001 [.20, .33]
Q×F Alcoholt → Q×F Alcoholt+1 .29 <.001 [.24, .34]
Cessationt → Cessationt+1 .34 .002 [.12, .52]
 Concurrent associations
Depressiont+1 → Q×F Alcoholt+1 .02 .049 [.001, .03]
Depressiont+1 → Cessationt+1 -.19 <.001 [-.28, -.09]
 Prospective associations
Q×F Alcoholt → Depressiont+1 .01 .64 [-.05, .08]
Cessationt → Depressiont+1 -.01 .76 [-.09, .08]
Depressiont → Q×F Alcoholt+1 .01 .68 [-.02, .03]
Depressiont → Cessationt+1 .07 .14 [-.03, .16]
Cannabis (n = 857)
 Autoregressions
Depressiont → Depressiont+1 .35 <.001 [.28, .43]
Freq. cannabist → Freq. cannabist+1 .35 <.001 [.32, .40]
Cessationt → Cessationt+1 .68 <.001 [.55, .89]
 Concurrent associations
Depressiont+1 → Freq. cannabist+1 .003 .78 [-.02, .03]
Depressiont+1 → Cessationt+1 -.13 .004 [-.23, -.04]
 Prospective associations
Freq. cannabist → Depressiont+1 .01 .74 [-.04, .06]
Cessationt → Depressiont+1 -.07 .10 [-.16, .02]
Depressiont → Freq. cannabist+1 .07 .06 [-.01, .16]
Depressiont → Cessationt+1 .01 .66 [-.03, .05]
Stimulants (n = 198)
 Autoregressions
Depressiont → Depressiont+1 .29 <.001 [.16, .45]
Freq. stimulantst → Freq. stimulantst+1 .32 <.001 [.11, .51]
Cessationt → Cessationt+1 .91 <.001 [0.60, 1.27]
 Concurrent associations
Depressiont+1 → Freq. stimulantst+1 .01 .88 [-.12, .14]
Depressiont+1 → Cessationt+1 -.28 .01 [-.49, -.07]
 Prospective associations
Freq. stimulantst → Depressiont+1 .11 .01 [.03, .19]
Cessationt → Depressiont+1 -.11 .20 [-.31, .07]
Depressiont → Freq. stimulantst+1 .05 .66 [-.17, .29]
Depressiont → Cessationt+1 -.01 .92 [-.21, .20]

Notes: CI = confidence interval; Q×F = quantity × frequency; freq. = frequency. All estimates adjusted for use of other substances (alcohol, cannabis, stimulants, other illicit substances); linear change in depressive symptoms, likelihood of cessation, and frequency of use; age at baseline; and cohort. Number of individuals in each analysis depended on the number of participants who used each substance at baseline.