Men’s and women’s response to treatment and perceptions of outcomes in a randomized controlled trial of injectable opioid assisted treatment for severe opioid use disorder

Table 2 Treatment outcomes by gender at six months

Outcomes at six months	Women N = 62	Men N = 140	Women vs. Men N = 202
Street opioid use
Days illicit heroin use^a	3.69 (2.08, 5.52)	3.84 (2.54, 5.27)	−0.15 (−2.16, 1.98)
Days illicit opioid use^a	5.28 (3.18, 7.59)	4.84 (3.43, 6.26)	0.44 (−1.93, 3.17)
Proportion of urine positive for street heroin markers^c	0.27 (0.18, 0.40)	0.25 (0.16, 0.35)	OR 1.16 (0.53, 2.51)
Retention in treatment
Proportion of participants receiving treatment ≥28 days^c	0.83 (0.72, 0.91)	0.79 (0.71, 0.85)	OR 1.34 (0.62, 2.87)
MAP health symptom scores
Physical health^b	12.29 (10.43, 14.15)	11.37 (10.07, 12.67)	0.92 (−1.38, 3.23)
Psychological health*^b	6.95 (5.01, 8.90)	9.35 (8.01, 10.69)	−2.39 (−4.72, −0.07)
Other outcomes
Days of non-legal activity^a	3.61 (1.61, 5.90)	3.14 (1.86, 4.53)	0.47 (−1.82, 3.14)
Days of crack cocaine use^a	7.10 (4.70, 9.76)	5.16 (3.26, 7.62)	1.95 (−0.08, 4.28)

*Indicates significance below p < 0.05
Differences in proportions (urine positive and retention) are presented as odds ratios (OR). For all other variables mean difference in days or scores are presented. Both proportions and means are presented with 95% confidence intervals in brackets
All models were adjusted for treatment arm and the average daily dose. Models were also adjusted for baseline values using analysis of covariance (ANCOVA) to account for gender difference at baseline. This is with the exception of the retention variable, and urine positive for street heroin markers (there is no baseline value for retention, and as per inclusion criteria all patients had urine positive for street opioids at baseline)
^aContinuous outcomes with an excess of zero counts: Zero-inflated Poisson regression was used. Adjusted mean difference between the two gender groups and confidence intervals were estimated by the Bootstrap method
^bContinuous outcomes: Linear regression models were used to estimate the mean difference and 95% CI
^cBinary outcomes: Logistic regression was used to estimate odds ratios and 95% CIs to compare the proportions between groups

ISSN: 1747-597X