proportion of physician visits was considered acceptable in light of the negligible costs associated with enrolling and following up with participants who do not visit a physician. The following cut-off points were also set in order to establish the logistic feasibility of the distribution method prior to proceeding with a larger study: (1) ability to recruit a sample of approximately 500 participants within 6 months without using paid advertisements, based approximately on the recruitment rate from a study we conducted distributing free NRT that required at least one visit to a pharmacy [29], assuming one quarter as many individuals would be interested in prescription cessation medications compared to NRT [30] and a lower eligibility rate due to a greater number of contraindications associated with bupropion and varenicline; (2) eligibility rate of at least 70 %, similar to comparable randomized trials comparing bupropion and varenicline [31, 32]; (3) at least 45 % of eligible participants would receive medication, based on the assumption that at least 50 % are able to visit a physician, and of these, almost all would receive a prescription due to the exclusion of those with contraindications at the initial assessment; and (4) no more than 7 days would be typically required to reach the participant via telephone for medication counselling and to confirm shipping address, operationalized as the median number of days between filling the prescription and sending the medication out for delivery, permitting most participants to receive medication within 30 days. Additional feasibility data collected included reasons for not scheduling a visit with a physician, frequency of medication delivery issues, and the average number of days between study enrollment and prescription dispensing and delivery. Logistic feasibility outcome data were collected from participants at each follow-up survey (to maximize the probability of having feasibility data for each participant, given loss to follow-up) and also extracted from pharmacy records. Where a participant provided conflicting responses at different follow-up time points with respect to feasibility outcomes, the earliest response is reported based on the assumption that recall would be less accurate as time elapsed. Acceptability of the motivational email component of the intervention was assessed by asking participants at each follow-up throughout the 12-week treatment period to rate the helpfulness of the emails on a fivepoint scale from 1 (not helpful at all) to 5 (very helpful). Responses provided later in treatment replaced ratings provided earlier in treatment, where applicable. Participants were also asked throughout and at the end-oftreatment follow-up how much of the 12-week supply of medication they had used. This was done because completion of treatment is a predictor of successful quitting [33]. Participants were asked to select from several
response options provided, which were collapsed into a smaller number of categories for analysis (none, less than 1 week, 1 to less than 4 weeks, 4 to less than 8 weeks, 8 to less than 12 weeks, all 12 weeks). As above, responses provided further along in treatment replaced earlier responses, where applicable. Treatment outcome was assessed in order to obtain an estimate of treatment effect to inform a sample size calculation for a future randomized trial. The outcome variable related to the effectiveness of the intervention was 7-day point prevalence of abstinence from smoking at 41 weeks after enrollment, approximately 6 months after the individual completed their 12-week course of medication. Seven-day point prevalence abstinence was defined as not having smoked even a puff over the previous 7 days.