Table 2.
Summary of effects.
| Study [reference]; target substance and substance use outcome measures | Intervention | Control | Between groups statistic and significance | Effect size (d)a,b | Quality assessmentc | |||||||||
|
|
|
N | Pre, mean (SD) or n (%) | Post, mean (SD) or n (%) | n | Pre, mean (SD) or n (%) | Post, mean (SD) or n (%) |
|
|
|
||||
| Aharonovich et al, 2017 [30]; drugs and alcohold | ||||||||||||||
|
|
Number days primary drug use over 30 days | 21 | 12.8 (4.4) | 5.0 (4.7) | 21 | 15.3 (7.3) | 8.5 (6.1) | IRRe=.59 (.35-.99), P=.04 | .17 | Good | ||||
|
|
Number drinking days over previous 30 days | 21 | 14.2 (7.3) | 7.0 (7.6) | 21 | 13.7 (6.3) | 8.1 (5.7) | IRR=.67 (.41-1.07), P=.09 | .23 | Good | ||||
|
|
Standard drinks (14 g)f per day over previous 30 days | 21 | 3.2 (2.4) | .9 (1.0) | 21 | 2.7 (2.0) | 1.3 (1.1) | IRR=.63 (.36-1.11), P=.11 | .41 | Good | ||||
| Baskerville et al, 2018 [31]; tobacco | ||||||||||||||
|
|
Continuous abstinence at 6 months | 354 | 0 | 49 (13.8) | 371 | 0 | 57 (15.4) | ORg=.89 (.59-1.34), P=.56 | .06 | Good | ||||
|
|
7-day point prevalence abstinence at 6 months | 342 | 0 | 114 (33.3) | 366 | 0 | 143 (39.1) | OR=.78 (.57-1.06), P=.11 | .14 | Good | ||||
|
|
30-day point prevalence abstinence at 6 months | 344 | 0 | 84 (24.4) | 365 | 0 | 107 (29.3) | OR=.78 (.56-1.09), P=.14 | .14 | Good | ||||
| Boendermaker et al, 2015 [32]; alcoholh | ||||||||||||||
|
|
Standard drinks (10 g) over 14 days | 25 | 25.1 (21.4) | 20.0 (17.1) | 24 | 25.4 (19.1) | 18.4 (14.6) | F1,44=0.03, P>.05 | –.05 | Poor | ||||
| Bricker et al, 2014 [33]; tobacco | ||||||||||||||
|
|
Percent of sample abstinent over 30 days | 80 | (0) 0% | (10) 13% | 84 | 0% | (7) 8% | OR=2.7 (.8-10.3), P=.12 | .55 | Fair | ||||
| Crane et al, 2018 [35]; tobacco | ||||||||||||||
|
|
Self-reported abstinence at 12-weeks | ITTi: 14,228; PPj: 1213 | (0) 0%; (0) 0% | (234) 1.6%; (234) 19.3% | ITT: 13,884; PP: 901 | (0) 0%; (0) 0% | (124) 0.9%; (124) 13.8% | OR=1.86 (1.49-2.31), P<.001; OR=1.50 (1.18-1.90), P<.001 | .34; .22 | Good | ||||
| Crane et al, 2018 [34]; alcohol | ||||||||||||||
|
|
Average number of alcohol units (8 g) per week | NFk:336; CBl: 336; MFm: 336; APn: 336; ICo: 336; NFxCBp: 168 | 39.1 (25.97); 40.3 (28.23); 39.9 (27.09); 39.0 (26.46); 39.0 (26.62); 38.9 (26.92) | 24.5 (22.45); 27.2 (25.96); 26.3 (23.41); 23.8 (24.23); 27.1 (26.47); 23.2 (20.68) | 336; 336; 336; 336; 336; 168 | 40.7 (28.66); 39.6 (26.45); 39.9 (27.63); 40.9 (28.20); 40.8 (28.05); 39.9 (27.85) | 26.6 (26.60); 23.9 (22.79); 24.5 (25.56); 27.0 (24.53); 23.7 (21.82); 21.5 (21.09) | F1,178=.30, P=.59; F1,178=.39, P=.53; F1,178=.78, P=.38; F1,178=.13, P=.71; F1,178=2.16, P=.14; F4,43=4.68, P=.03 | .08; .09; .13; .05; .22; .67 | Good | ||||
| Davies et al, 2017 [36]; alcohol | ||||||||||||||
|
|
AUDIT-Cq (alcohol consumption score) | 104 | 6.6 (2.62) | 6.0 (2.33) | OTMr: 99; ICs: 97; WLt: 102 | 6.2 (2.54); 6.8 (2.49); 6.6 (2.46) | 5.7 (2.47); 6.1 (2.38); 5.9 (2.42) | All between-group comparisons P=ns | −.01; .03; .03 | Good | ||||
| Earle et al, 2018 [37]; alcohol | ||||||||||||||
|
|
Maximum drink (undefined) number on single night over current semester | PNF+u: 72; PNFv: 79 | 4.23 (4.14); 3.87 (4.07) | 2.97 (3.25); 3.53 (3.38) | 71 | 3.80 (4.09) | 3.82 (4.28) | See footnote$ | .31; .09 | Good | ||||
|
|
Drink (undefined) number over previous weekend | PNF+: 72; PNF: 79 | 3.08 (4.10); 2.65 (3.74) | 1.94 (2.67); 2.26 (3.28) | 71 | 3.32 (4.81) | 3.21 (4.70) | See footnote$ | .23; .07 | Good | ||||
|
|
Composite score (maximum single occasion and drink number on previous weekend) | PNF+: 72; PNF: 79 | 3.08 (4.10); 2.65 (3.74) | 1.94 (2.67); 2.26 (3.28) | 71 | 3.32 (4.81) | 3.21 (4.70) | Mean=−0.14, SE=0.07, P<.01w; mean=0.00, SE=0.07, P=nsw | .46; .24 | Good | ||||
| Kerst and Waters, 2014 [44]; tobacco | ||||||||||||||
|
|
Percent of sample any smoking on test day | 30 | (29) 97% | (27) 90% | 30 | 100% | (27) 90% | B=.04 (−.35 to .43), P=.85 | −.01 | Fair | ||||
|
|
Expired carbon monoxide (ppm) | 30 | 15.9 (5.35) | 15.5 (7.70) | 30 | 15.6 (4.68) | 14.2 (5.81) | B=1.07 (−4.13 to 1.99), P=.49. | −.20 | Fair | ||||
|
|
Salivary cotinine (ng/ml) | 30 | 394 (181) | 410 (211) | 30 | 420 (253) | 408 (268) | B=−23.4 (−105.4 to 58.6), P=.57 | −.13 | Fair | ||||
| Gajecki et al, 2014 [38]; alcohol |
|
|||||||||||||
|
|
Standard drinks (12 g) over 7 days | PPx: 153; PKy: 341 | 8.57 (6.12); 9.62 (6.26) | 8.32 (6.45); 9.75 (7.05) | 489 | 9.15 (6.18) | 8.62 (6.28) | Time x group LMMz P=.82; time x group LMM P=.41 | −.05; −.11 | Fair | ||||
|
|
Number drinking days over 7 days | PP: 153; PK: 341 | 2.17 (1.12); 2.24 (1.20) | 2.17 (1.23); 2.36 (1.23) | 489 | 2.29 (1.19) | 2.15 (1.19) | Time x group LMM P=.23; time x group LMM Z=3.39, P=.001 | −.12; −.22 | Fair | ||||
| Gajecki et al, 2017 [39]; alcoholaa | ||||||||||||||
|
|
Standard drinks (12 g) over 7 days | 71 | 16.58 (7.84) | 12.87 (9.73) | 124 | 17.16 (7.87) | 14.52 (7.46) | Z=−1.07, P=.29 | .15 | Fair | ||||
| Number drinking days over 7 days | 71 | 3.35 (1.20) | 2.51 (1.15) | 124 | 3.53 (1.39) | 3.02 (1.45) | Z=−2.12, P=.03 | .30 | Fair | |||||
| Gonzalez and Dulin, 2015 [40]; alcoholaa | ||||||||||||||
|
|
Standard drinks (14 g) over 7 days | 28 | 39.12 (20.37) | 22.07 (22.08) | 20 | 33.34 (21.58) | 12.26 (13.19) | PE=−4.28, P<.05 | See footnoteab | Poor | ||||
| Percent heavy drinking daysac | 28 | 54.25 (28.93) | 26.98 (29.47) | 20 | 47.74 (29.71) | 15.04 (24.03) | PE=−8.25, P<.05 | See footnoteab | Poor | |||||
| Percent abstinent days | 28 | 30.36 (22.48) | 51.32 (32.25) | 20 | 38.21 (29.85) | 60.90 (39.22) | PE=10.29, P<.05 | See footnoteab | Poor | |||||
| Gustafson et al, 2014 [17]; alcohol | ||||||||||||||
|
|
Number of risky drinking days over 30 daysad | 132 | 1.50 (0.34)ae | Post: 1.54 (0.49); FUaf: 1.13 (0.50) | 139 | 3.01 (0.48)ae | Post: 2.65 (0.48); FU: 2.60 (0.49) | Post mean difference=1.11, P=.10; FU mean difference=1.47, P=.03 | −.18; .24 | Poor | ||||
|
|
Percent of sample abstinent over 30 days | 132 | (100) 75.6%ae | Post: (103) 78.1%; FU: (104) 78.7% | 139 | (94) 67.7%ae | Post: (93) 66.9%; FU: (91) 65.5% | Post between group OR=1.70, P=.04; FU between group OR=1.94, P=.02 | .29; .37 | Poor | ||||
| Hasin et al, 2014 [41]; alcohol | ||||||||||||||
|
|
Standard drinks (14 g) per drinking day over 30 days | 39 | 9.3 (6.9) | 3.9 (2.1) | 43 | 8.1 (3.9) | 3.6 (1.6) | See footnoteag | .16 | Poor | ||||
|
|
Percent of abstinent days over 30 days | 39 | 58.1 (27.4) | 79.2 (22.5) | 43 | 61.3 (24.2) | 82.1 (17.8) | See footnoteag | .01 | Poor | ||||
| Hertzberg et al, 2013 [42]; tobacco | ||||||||||||||
|
|
Percent of sample abstinent for previous 7 days (bio-verified) at end of 4-week treatment | 11 | (0) 0% | (9) 82% | 11 | (0) 0% | (5) 45% | Χ21=1.6, P=.21 | .55 | Poor | ||||
|
|
Percent self-report (not bio-verified) abstinence at 3-months post intervention | 11 | (0) 0% | (6) 55% | 11 | 0% | (2) 18% | Χ21=2.0, P=.15 | .64 | Poor | ||||
| Hides et al, 2018 [43]; alcohol | ||||||||||||||
|
|
Risky single occasion drinking frequency over 1 monthah | 97 | 2.11 (0.91) | 2.23 (1.17) | 86 | 2.10 (1.08) | 2.25 (1.15) | F1,184=0.28, P=.60 | .03 | Fair | ||||
|
|
Typical standard drinks (10 g) over 1 month | 97 | 2.79 (1.41) | 2.56 (1.32) | 86 | 2.64 (1.40) | 2.24 (1.21) | F1,188=1.00, P=.32 | -.12 | Fair | ||||
| Krishnan et al, 2018 [45]; tobacco | ||||||||||||||
|
|
Percent self-reported and biochemically verified abstinence at 30 days | 39 | (0) 0% | (1) 3% | 50 | 0% | (1) 2% | P>.05 | .04 | Poor | ||||
|
|
Median number of cigarettes over 30 days (IQR) | 39 | 10.0 (6.0-20.0) | 5 (4.0-10.0) | 50 | 10.0 (8.0-20.0) | 6.0 (3.0-10.0) | P=.84 | See footnoteaj | Poor | ||||
|
|
Median expired carbon monoxide in ppm (IQR) | 39 | 23.0 (18.0-33.0) | 19.5 (15.0-26.0) | 50 | 22 (14.0-30.0) | 18.5 (10.0-28.0) | P=.52 | See footnoteaj | Poor | ||||
| Liang et al, 2018 [46]; drugs | ||||||||||||||
|
|
Number of days with primary drug use over 7 days (pre=end of week 1, no baseline) | 48 | 1.33 (2.48) | 0.71 (1.87) | 25 | 3.08 (3.37) | 2.20 (3.06) | OR=.29 (.06-1.44), P=.13 (regressed over 4 time points during intervention) | .09 | Poor | ||||
|
|
Drugs detected in urine (pre=end of week 1, no baseline) | 42 | (23) 56% | (11) 26% | 20 | (14) 70% | (10) 50% | OR=.57 (.11-2.84), P=.49 (regressed over 4 time points during intervention) | .11 | Poor | ||||
| Ruscio et al, 2016 [47]; tobacco | ||||||||||||||
|
|
Cigarettes per smoking day | 20 | N/A | N/A | 17 | N/A | N/A | F1,436=5.50, P=.01ak | N/A | Fair | ||||
|
|
Expired carbon monoxide (ppm) | 18 | 18.4 (9.8) | 14.3 (8.2) | 14 | 19.1 (6.7) | 15.4 (5.0) | F1,29=0.01, P=.92 | .05 | Fair | ||||
|
|
Salivary cotinine (ng/ml) | 18 | 504.4 (300.3) | 433.9 (257.1) | 14 | 452.9 (221.9) | 482.8 (250.0) | F1,29=0.04, P=.84 | .37 | Fair | ||||
| Witkiewitz et al, 2014 [48]; alcohol and tobacco | ||||||||||||||
|
|
Standard drinks (14 g) per drinking day | BMal: 30; DMam: 29 | BM: 5.57 (2.81); DM: 5.58 (2.45) | BM: 4.83 (2.59); DM: 4.56 (2.65) | 26 | 7.46 (3.46) | 6.05 (2.88) | WALDX22=0.1, P=.94 an | .08 | Poor | ||||
| Number of heavy drinking days over 7 daysao | BM: 30; DM: 29 | BM: 2.31 (1.53); DM: 2.45 (1.44) | BM: 2.07 (1.70); DM: 1.76 (1.33) | 26 | 2.86 (1.41) | 2.31 (1.35) | WALDX22=0.1, P=.91an | .07 | Poor | |||||
| Cigarettes per smoking day | BM: 30; DM: 29 | BM: 4.93 (3.43); DM: 4.78 (4.83) | BM: 3.28 (3.35); DM: 2.71 (2.86) | 26 | 3.76 (2.15) | 4.55 (4.07) | B=2.04, P=.002; B=1.59, P=.02 | .55ap; .45ap | Poor | |||||
| Number of days with drinking and smoking over 7 days | BM: 30; DM: 29 | BM: 2.81 (0.59); DM: 2.82 (0.64) | BM: 1.97 (1.09); DM: 2.07 (0.88) | 26 | 2.66 (0.48) | 1.76 (0.83) | WALDX22=2.0, P=.38 an | .31 | Poor | |||||
aAll effect sizes are Cohen d. Sign of effect size indicates agreement with hypothesized direction (positive implies app condition improved outcome to a greater degree than comparison conditions; ie, a reduction in consumption or an increase in rates of abstinence).
bWhere effect sizes not reported as Cohen’s d, effect sizes were converted from reported effect sizes where possible or derived using pooled baseline SDs from intervention and control groups, as described by Morris [49].
cGood quality=all criteria in the Cochrane Risk of Bias tool were met, fair quality=one criterion not met or two criteria unclear and the assessment that this was unlikely to have biased the outcome and there was no important limitation that could invalidate the results, poor quality=one criterion not met or two criteria unclear and the assessment that this likely biased the outcome and there were important limitations that could invalidate the results OR two or more criteria listed as high or unclear risk of bias.
dStudies in italics reported significant outcomes for intervention app at post-intervention and/or follow-up timepoints compared with control. Sample sizes reflect the number of participants included in the final analyses.
eIRR: incidence rate ratio.
fAmount in grams of pure alcohol in one standard drink varies across countries and is indicated in brackets.
gOR: odds ratio.
hSome data provided directly from authors.
iITT: intention to treat analysis (referred to in publication as “Missing Equals Smoking”).
jPP: per protocol analysis (referred to in publication as “Follow-up Only”).
kNF: personalized normative feedback.
lCB: cognitive bias retraining.
mMF: monitoring and feedback.
nAP: action planning.
oIC: identity change.
pTwo-way interaction between personalized normative feedback and cognitive bias retraining.
qFirst three questions of the 10-item Alcohol Use Disorder Identification Test, probing alcohol consumption.
rOTM: “One Too Many” (intervention website).
sIC: imagery control (sham website).
twl: waitlist control.
uIntervention using personalized normative feedback plus reflection
vIntervention using personalized normative feedback without reflection.
wMain analysis used variables derived by combining drinking measures and controlling for a range of other covariates.
xPP: PartyPlanner app.
yPK: Promillekoll app.
zLMM: linear mixed model.
aaSome data provided directly from authors.
abGroup×time interaction analysis during intervention (not pre-post and hence was not considered superior to control as per our definition).
acHeavy drinking defined here as 4+ standard drinks for females and 5+ for males.
adRisky drinking defined here as 3+ standard drinks (14 g of alcohol) for females and 4+ for males consumed within a 2-hour period.
aeNo baseline data were collected in this study as participants were inpatients who had not consumed alcohol for some time; authors use 4-month data as reference for 8-month post intervention and 12-month follow-up analyses.
afFU: follow-up.
ag “No between groups significance conducted.
ahRisky single occasion drinking in this study is defined as 5+ standard drinks (10 g of alcohol) during one occasion.
ajMedian scores, cannot compute Cohen d effect size.
akLMM group×day interaction based on daily smoking reports over 2 weeks (not pre-post).
alBM: BASICS-Mobile app.
amDM: Daily monitoring app.
anOmnibus chi-square test across all 3 conditions.
aoHeavy drinking defined here as 4+ standard drinks for females and 5+ for males.
apEffect size controlling for range of predictors.
$ Between group significance testing not conducted