Behavioral Intervention to Reduce Sexual and Injection Risks Among Female Sex Workers Who Also Inject Drugs in Mexico (FSW-IDU)

• Combined HIV/STI 12-month Incidence Rates of HIV, Syphilis, Chlamydia, Gonorrhea and Trichomonas Vaginalis. [ Time Frame: 12 months, with measuring points at baseline and at 4, 8, and 12 months past baseline ] [ Designated as safety issue: No ]

  Combined incidence for HIV/STI was calculated over the 12-month study period and included only those who a) had at least one follow-up visit and b) at baseline tested negative for HIV and any of the aforementioned STIs.

  In the calculations we accounted for the time each participant spent at risk of HIV/any STI during the follow-up period, by using available information on each participant for each time point (i.e. baseline, 4-, 8-, and 12-months was used).

  The analytic method used for this outcome analysis was Poisson regression with robust variance estimation. The outcome variable was a binary variable indicating whether a participant has contracted HIV or a new STI during the 12-month follow-up period. The primary factor of interest was the intervention group. The log ("time spent at risk of HIV/any STI") was used as an offset variable in order to account for the time spent at risk of HIV/any STI by each participant.

  
  

• Change (Baseline to 4-, 8-, and 12-months) in the a)Mean Number of Unprotected Sex Acts With Clients and b)Ratio of Unprotected Sex Acts With Clients (Relative to the Number of Sex Acts With Clients. [ Time Frame: 12 months, with measuring points at baseline and at 4, 8, and 12 months past baseline ] [ Designated as safety issue: No ]
  The analytic method used for a) was negative binomial regression with the number of unprotected sex acts with clients as the outcome variable and intervention group, time point (baseline, 4-, 8-, and 12-months), and the interaction between the two as the main effects of interest. Similarly, for part b) we used negative binomial regression, except that in this case log(total number of sex acts with clients) was used as an offset variable in order to create the ratio of unprotected sex acts (relative to total number of sex acts).
  
  
• Change (Baseline to 4-, 8-, and 12-months) in the a) Proportional Odds of Higher Receptive Needle Sharing and b) Mean Score of the Injection Risk Index (IRI). [ Time Frame: 12 months, with measuring points at baseline and at 4, 8, and 12 months past baseline ] [ Designated as safety issue: No ]

  For a) we asked: "In the past month, how often have you used a needle or syringe that you knew or suspected had been used before by someone else?" (possible responses: 1=never, 2=sometimes, 3=about half the time, 4=often, 5=always), with a higher response indicating a higher risk.

  Analyzed using ordinal logistic regression with frequency of receptive needle sharing as outcome variable and intervention group, time point, and interaction between the two as main effects of interest.

  For b)IRI calculated based on: receptive needle sharing, sharing a bottlecap/spoon/cooker, sharing cotton filter for a needle, or rinse water after someone else has used it, and using a used syringe to divide drugs. Score constructed by calculating average score between responses to injection risk indicators, with higher score representing higher risk. We analyzed using gamma regression with IRI as outcome and intervention group, time point and interaction between the two as main effects of interest

  
  

Our specific aims are:

• Aim 1: To evaluate the efficacy of a behavioral intervention to decrease sharing of syringes and injection paraphernalia among FSW-IDUs. We hypothesize that FSW-IDUs in the active experimental injection risk reduction condition will report: (a) less receptive and distributive needle sharing; (b) less sharing of injection paraphernalia; (c) obtaining syringes and injection paraphernalia from safer sources.
• Aim 2: To evaluate the efficacy of a behavioral intervention to increase condom use among FSW-IDUs in the context of ongoing drug use. We hypothesize that FSW-IDUs in the active experimental sexual risk reduction condition will: (a) report less unprotected vaginal and anal sex; and (b) have fewer incident cases of specific STIs.
• Aim 3: To evaluate the joint effects of these two behavioral interventions to increase condom use and reduce sharing of needles and syringes/injection paraphernalia among FSW-IDUs. We hypothesize that the joint effect of these interventions will generate greater risk reductions compared to either intervention alone.
• Aim 4: To determine the extent to which theoretically-important components of our interventions (i.e., self-efficacy, outcome expectancies, attitudes, intentions) represent underlying mechanisms of change in primary outcomes (i.e., sexual- and injection-related risk reductions).
• Aim 5: To explore subgroup differences in the efficacy of: a) the sexual risk reduction, and b) the injection risk reduction intervention based on background characteristics, contextual factors, social factors and intrapersonal factors.