The aim of this study was to determine factors associated with engaging in sex work among long-term opioid injection drug users receiving OAT in the frame of a clinical trial. A higher proportion of women (53.6%) compared to men (1.3%), reported engaging in sex work in at least one of the seven research visits from baseline to 24 months. At treatment entry, age and education were associated with sex work, while during the study period, psychological symptoms and frequent heroin and cocaine use in the prior 30 days were associated with sex work.
At treatment initiation, women who were younger and had fewer years of education were more likely to be engaged in sex work, factors which have previously been associated with sex work among women using injection drugs [4, 10, 12]. Housing, ethnicity and incarcerations have also been documented in previous studies . The lack of such associations in the present study reflects the homogeneity of the NAOMI sample, possibly due to study inclusion criteria.
When examining factors associated with sex work involvement during the study period, women with poorer treatment outcomes were more likely to engage in sex work. Specifically, lower treatment retention, poorer scores in social relations and health related quality of life, more days of illicit drug use, injection drug use, and more frequent daily injection in the prior 30 days. These findings indicate that sex work was more likely among a subgroup of women who did not fully benefit from OAT, a noteworthy finding considering that OAT has shown to reduce many of the harms associated with long-term heroin use [2, 18]. Moreover, in the present study women who were retained successfully in OAT were less likely to be involved in sex work and therefore experienced a reduced vulnerability to harms caused by injection drug use. While this is not a causal association, it indicates that those involved in sex work were more likely to drop-out of treatment.
In the multivariate model, psychological symptoms and high illicit heroin and cocaine use in the prior 30 days were associated with sex work. Similarly, previous studies have found that sex workers accessing MMT  and syringe exchange programs  presented with higher psychological distress compared to women not engaged in sex work. In addition, a higher use of substances , including more frequent daily heroin and cocaine use [4, 10] has also been reported among injection drug using women who also engage in sex work. The results of the present study complement prior research in the context of a prospective design that allowed us to capture predictors of sex work involvement over a 24 month study period. After engaging these participants in OAT, women who continued engaging in sex work were more likely to continue using heroin and cocaine, independent of OAT retention. Thus, many women continued engaging in survival sex work and using illicit heroin, despite that OAT improves retention and reduces illicit heroin use. The complexity of the relationship between OAT effectiveness and its impact on sex work engagement requires further study.
Education is regarded as a strong indicator of social and health-related inequalities , and women with fewer years of education were more likely to engage in sex work at treatment initiation. These findings indicate that women with less education experience further vulnerabilities even within a population with very low socio-economic status. Therefore, those who provide addiction treatment services must consider this special circumstance, acknowledging women's financial needs and the stigma attached to sex work, so that services and policies do not further exclude these groups.
The present study focuses on long-term opioid injection drug using women with and without involvement in sex work. It is well known that opioid-dependent individuals often show poor mental and physical health as well as poor psychosocial functioning, especially after long-term use [3, 32, 33]. There is also growing research evidence among women and men accessing OAT demonstrating that women enter treatment with worse physical and psychological health [8, 9], as well as higher opioid and stimulant use [34, 35, 35]. Some evidence has also suggested that women have poorer OAT outcomes compared to men [8, 9]. Therefore, women using injection drugs represent a particularly important group in the provision of effective addiction treatment.
Unexpectedly, there was no association between victimization (e.g., physical, emotional, and sexual abuse) and sex work in the present study. Previous studies have shown high rates of physical and sexual abuse in sex workers' childhood, and later victimization by partners [12, 36] and clients in adulthood . Moreover, in a cohort study of youth using substances, childhood sexual abuse was independently associated with sex work . One possible explanation for the absence of this expected association may be related to the measure of victimization. The Addiction Severity Index (ASI) has been used to evaluate abuse in several studies , measuring victimization with a general question (e.g., "have you ever been physically or sexually abused?"). Therefore, [40–42] details regarding the nature of the event, which might account for the associations between victimization and sex work, are undetermined [43, 44].
Limitations of the NAOMI study have been discussed elsewhere [24, 25]. It should be noted that the analysis were intent-to-treat, therefore, some of the higher intensity drug use occurred in participants who were not receiving the treatments as provided in the study (oral and injectable arms). Several gender sensitive and sex work specific-related questions were not part of the study evaluation package (e.g., partner's use of illicit substances, income earned from sex work to support heroin use), that data would have provided a more detailed picture of the situation. In addition, the trial was not designed to investigate factors associated with sex work and we had a small sample size; however it provided an opportunity to obtain valuable information on this topic in the context of women receiving OAT. In order to better conceptualize the relationship between high intensity drug use and sex work, additional data regarding the reasons for sex work involvement during treatment, the proportion of earnings used from sex work to support illicit drug use, and information regarding the people who depend on an individual involved in sex work, should be captured..
The findings presented suggest that participation in NAOMI positively affected the pattern of sex work, showing a decline from enrolment to 24 months follow-up. At treatment entry, all NAOMI participants had not received any treatment for the six months prior to study enrolment (as per inclusion criteria); therefore considered un-treated despite the available options (e.g., methadone treatment). Engagement in treatment was associated with a decline in sex work over time. This particular group would have likely remained outside of addiction treatment services, and likely only initiated treatment for the opportunity to receive injectable diacetylmorphine. Certainly, approaches that improve treatment engagement (such as medically prescribed diacetylmorphine) for long-term treatment refractory heroin injectors, must be supported by current policies.