The subgroups were defined based on: length of treatment prior to recruitment, follow-up method (telephone/face-to-face/patient completed), length of follow-up (≤ 7 months/ > 7 months) (the intention was to follow-up patients between 6 and 7 months but some took longer), and number of training sessions pharmacists attended

(less than four sessions, http://www.selleckchem.com/products/VX-765.html and four sessions). Data analysis was done on an intention-to-treat (ITT) basis, with a secondary per-protocol analyses based on actual intervention delivery. For the ITT analysis, imputation was used to estimate missing treatment satisfaction, physical and psychological health scores. All other missing outcomes were excluded from analyses. The denominators in each group vary depending on whether ITT or per-protocol analysis PI3K inhibitor was used. In total, 542 patients were randomised (295 intervention, 247 control). Baseline interviews were completed for all patients recruited and follow-up interviews

were completed for 335 (62%) (182 intervention, 153 control) patients (Figure 1). No differences in the baseline demographics between groups were seen (Table 1). In total, 121 patients had left their original pharmacy during the study and the retention status of 65 patients was determined. Intervention n = 295 Control n = 247 A total of 87 pharmacies (95 pharmacists) was contacted, of which 76 (84 pharmacists) recruited patients into the study. Four intervention pharmacies moved to the control group as they were not able to attend the training sessions and three control pharmacists moved to the intervention group to attend the training sessions. Pharmacist attendance rate for all four training sessions ranged from 60–80% across the six areas. Scores on the BECCI range from 10–41, with a mean of 30.3. The maximum

possible score is 44. The median score overall was 32 (interquartile range 24, 38), indicating good use of the technique. There Florfenicol was a reduction in both groups in the proportion of patients using illicit heroin in the last 30 days (16% decrease from 88 patients (48.4%) at baseline to 59 (32.4%) at follow-up for intervention patients and 19% decrease from 77 patients (50.3%) to 48 (31.4%) in controls), but this was not significant between groups (P = 0.83, Table 2). Within both groups, there was a significant reduction in the median number of days of illicit heroin use between baseline and follow-up (both P < 0.001). Sub-group analysis of illicit heroin use by length of time in treatment, length of follow-up and method of follow-up revealed no significant between-group differences (Table 2). Intervention n = 182 Control n = 153 Intervention n = 182 Control n = 153 Table 3 shows a reduction in the proportion who used other illicit drugs in both groups, between baseline and follow-up, although this did not reach statistical significance (P = 0.13 and P = 0.06 in the intervention and control group respectively).